rangifer’s diary: pt. cxii

A brief note on the eSpeak feature of this diary

If you’re reading this (unlikely), then you may have noticed some weird orange text in some of my diary entries:

You can even test it on this text right here.

If you’ve disabled JavaScript, or aren’t reading this on my website, or whatever, then these will just be ordinary anchors (hyperlinks) to audio files. With the fancy stuff, you can just click on it, or press ⏎ with it selected, & the audio should ✨magically✨ play (after taking some time to download it, most likely). Either way, they’re all just hyperlinks to plain ol’ audio files.

Occasionally, I’ll be using this feature for cheeky li’l gags. But the large majority of the time, these files contain audio generated by eSpeak NG, which is a free-software speech synthesiser primarily used for TTS purposes. The basic idea is to provide audible pronunciations (“The Pronunciations You Can Hear!”^®) of snippets of my diary that would otherwise just be text, whether it be because that text is in a non-English language, or because that text is actually a transcription in the IPA, or because English is just fucking confusing^[1].

Although I really believe this to be a huge improvement, there are limitations. Oh, so many limitations. And I feel the need to at least inform you, my dear reader, about the pitfalls of such linguistic hubris, lest ye naïvely listen to the audio & believe what ye hear.

The prices we pay for the sins we commit against natural language

If you’re a linguist or have linguistic training, that’s great. I probably don’t need to warn you about anything. So I’m going to assume that the reader has no linguistic training whatsoever.

The ears listen, but the brain hears

ℹ️ I had a deermentia moment. I forgor to put this section in the original version of this essay… Pls 4give me…

In listening to eSpeak’s audio, you likely aren’t hearing what it’s trying to tell you. This is due to the simple fact that it’s usually difficult (or impossible…) to hear distinctions made in other peoples’ speech that you don’t already make in your own.

For example, consider Arabic languages, which generally make a distinction between the two consonants transcribed as /⁠k⁠/ and /⁠q⁠/. The English speaker likely hears these as /⁠k⁠/ and /⁠k⁠/ respectively, but they’re actually two entirely separate consonants. Worse yet, the Arabophone can make further distinctions based on something called gemination (see the “Say it with me” section below), where the consonant can be “≈twice as long”. This yields the four-way distinction /⁠k⁠/ vs. /⁠kː⁠/ vs. /⁠q⁠/ vs. /⁠qː⁠/. This is, naturally, likely to be lost on the ears of an untrained listener.

But the Arabic languages are unrelated to English! That’s cheating — of course Anglophones don’t understand! But my dear reader, English speakers don’t even understand the phonology of other English speakers. For example, if you learnt English in North America (including the Caribbean), or in certain regions of England & Ireland (it’s complicated), then you don’t have a distinct LOT^[7] vowel! But eSpeak, much like many English speakers, has no trouble making the distinction:

• shawm /⁠ʃɔm⁠/.
• tom /⁠tɒm⁠/.
• palm /⁠pɑm⁠/^[8].

For those who make the distinction, none of these rhyme. If you learnt English in one of the aforementioned regions, then tom probably rhymes with palm^[8]. If you’re quite special, then all three of these^[8] rhyme for you! ✨Wow!✨

eSpeak doesn’t even know that this sound exists

eSpeak’s notion of what sounds there are to speak — which I’m going to call segments — is tied to the language specified by the user. For example, when the language is English, eSpeak recognises the segment [[I]]^[2] (the /⁠ɪ⁠/ in kit /⁠kɪt⁠/), but when the language is something else, it may not recognise [[I]] whatsoever, or worse, may assign to it a very different phoneme^[3]. This makes sense insofar as every language has its own unique phonemic inventory, which is usually wildly different from the inventories of other languages — especially when those languages aren’t closely related to it.

Unfortunately, eSpeak’s idea of the segmental inventory of a language isn’t always superb, & this is more evident in some languages than in others… Worse, eSpeak only knows of so many languages, & it’s not uncommon that I have text in a language that eSpeak has never heard of! What the hecc!!

Worse yet, the issue isn’t always one of segments. So-called suprasegmentals include things like stress (think of the difference between suspect & suspect), tone, & even pitch accent. eSpeak sometimes doesn’t have the necessary support for suprasegmentals (e.g. not having pitch accent for a pitch-accent language like e.g. Japanese). And sometimes, eSpeak forces the use of stress on languages that do not have this kind of syllable-level stress (e.g. French), or on particular words that don’t have this kind of stress.

Anglocentrism

The English-speaking origins of eSpeak are evident not just from the history of its development, but also from the program itself. For example, it uses Kirshenbaum for some reason.^[2] More importantly, eSpeak’s phonetics are often influenced by an English understanding, even when those phonetics are of a language not closely related to English. Consequently, for example, asking eSpeak to read text with a language of “Japanese” does not result in the sound of a Japanese robot speaking. Instead, it results in the sound of an English robot doing its best impression of a Japanese robot speaking. Ow? My ears?

And of course, the further a language is from being as widely-spoken as English or Standard Mandarin, the less well-developed its support in eSpeak will likely be.

Nice try, eSpeak

Sometimes, eSpeak has a good idea of a particular phoneme’s^[3] existence — which is great — but the way that it actually sounds is consistently goofy.

For example, Arabic languages have a phoneme written as ⟨ع⟩ & usually transcribed as /⁠ʕ⁠/ (“voiced pharyngeal fricative”). This is a common sound in Arabic, so obviously eSpeak supports it. But it sounds wild. eSpeak seems to think that glottal stops (transcribed /⁠ʔ⁠/, e.g. the middle of uh-oh!) are basically just short pauses, so their effect on surrounding segments is inaudible, & thus glottal stops are frequently less than audible. I’ve also found that pretty much any Vietnamese text sounds wonky in eSpeak, & I’m unsure why. Sad. And so on…

Phones

This brings me to the subject of phones. Phones are sorta like phonemes^[3], but less abstract. Speakers don’t always make hard distinctions between two different phones — this is called allophony. But if you want to make speech sound convincing, you sometimes have to use certain phones in certain situations rather than others. For example, in many (but not all) dialects of English, the “t” & “d” sounds can become [ɾ] in certain circumstances, e.g. latter & ladder can sound very similar (or even become homophones).

Thus, each phoneme has one or more possible phones that are like possible “forms” that the phoneme can take on. In the simplest case, a phoneme really only has just one phone. But even this simplest case is sometimes too much for poor eSpeak!

Take, for example, the name of my darksterity knight rusa. The actual origin of this name is Malay, but looking just at the spelling alone, it really could be a word in very nearly any language that uses a Latin writing system. The result is that, in my head, I like to think that it’s something like [ˈʁuzə]. The good news is that there are languages out there with phones that are reasonably close to these four [ʁ, u, z, ə], e.g. certain dialects of French. So I can just set eSpeak’s language to “French” & call it a day?

Not so fast. As it turns out, a phone that’s pretty dang close to [ʁ] just… doesn’t exist in eSpeak. Not in any of its languages, as far as I know. So this is actually not just a problem for my fake bullshit pronunciation of rusa, but also potentially a problem for any language that has a phone that’s well-described as [ʁ]! 🥲

I ended up trying to approximate [ʁ] with Arabic /⁠ɣ⁠/ or something, gave up, & posted it into my diary anyway. Great.

Dead languages

Some languages are dead, but it can still be useful to give an indication as to the pronunciation of their texts.

In some cases, like Latin, we have a pretty crisp idea of how it was pronounced in a given time period, & eSpeak supports the language. In other cases, eSpeak won’t support the language, & I just have to hope that I can cobble something together that approximates what I really want eSpeak to do.

In the latter cases, the situation is even worse if our knowledge of the pronunciation is not that good. This is especially the case the more ancient the language is, & the more that its writing system falls short of being perfectly phonetic.

Realisations

The “correct” (meaning both idiomatic & intended) realisation of a given text varies:

• A given speaker will vary in how they realise it, depending on many factors.
• Different speakers within a given “lect” will have different realisations, again for complex reasons.
• Linguists partition language families & dialect continua into groupings that are generally referred to as varieties, or lects,^[6] & these are — as you’d expect — often characterised by differences in phonology (amongst many other things).
• Writing systems can bring together lects that are only distantly related, & in some cases unrelated entirely. For example, what if you see a Chinese character with no context? What language is that? Different lects, many of which are completely & totally cladistically unrelated — e.g. Standard Mandarin & Japanese — will differ in how they assign a “reading” — & in many cases, multiple readings — to a given character or text.

Wow, this sounds like rubbish

You might have noticed how eSpeak, even on a good day, kinda sounds like a robot choking down an empty tin can. That’s just how it is: robots sound like robots.^[4] Human speech production is just immensely complicated! And eSpeak essentially attempts to synthesise that speech directly, by modelling the acoustic phenomena (viz. formants) generated by abstract human speakers.

This formant synthesis method may or may not sound like a dumb idea, & it certainly tends to produce highly robotic-sounding voices in eSpeak’s case, but it’s actually pretty neato in a lot of ways:

• eSpeak’s methods, when done well, produce audio that is highly intelligible, even at very rapid speeds. This makes eSpeak et al. a great fit for screen readers, for example.
• “Synthesise everything” is such an abstracted approach that it can be done well without needing lots of actual human speech (and/or other audio) as input. In theory, everything is synthesised “from scratch”, which is basically the polar opposite of the machine learning approach, which requires colossal amounts of input data (& training time) to train on.
• The abstraction also means that there are no particular, concrete limitations. Synthesising a sound that humans can’t actually physically produce isn’t necessarily useful, but what might be useful is synthesising a segmental inventory so large that it can cover many — or even all‽ — languages.

…But also it sounds like a robot screaming into a kazoo.

One quirk unique to eSpeak in particular is that the screaming robot is always male. Now, I wouldn’t even bother pointing this out, if it weren’t for the fact that the output of espeak-ng --voices has an Age/Gender column… where the value in that column is always --/M. 🥲

Audio is not a substitute for a transcription (& vice versa)

This is the spirit of what I want you to understand: audio is not a substitute for a transcription — & vice versa.

I started making use of the IPA in my diary entries a while back, because there were situations where pronunciations & phonologies were important to the subject matter. It then took me a very long time to adopt something like this “eSpeak feature”, because it’s simply not an obvious fit. It took lots of tinkering, many compromises, & a “better than nothing” attitude, to get to this point.

It’s perhaps intuitively obvious that a textual transcription can’t carry all the nuances of actual human speech made by messy humans with their physical fleshy-flesh.

On the other hand, understanding IPA transcriptions requires many layers of linguistic knowledge, & I don’t generally expect my readers to have basically any of those layers. However, the implication of this is that those same transcriptions have, embedded within them (both explicitly & implicitly), many layers of information & nuanced understanding. If hearing a pronunciation — whether by a native speaker, or by a robot gargling lumpy mouthwash — somehow bestowed upon the listener this kind of understanding, then everyone would be a linguist^[5].

Footnotes for “A brief note on the eSpeak feature of this diary”

1. [↑] You don’t know how much I’ve learnt about English phonology solely as a result of writing this diary, & you don’t want to know. Trust me. 😔
2. [↑] Inside the double square brackets ([[…]]), eSpeak uses Kirshenbaum instead of just using X-SAMPA or something. I don’t know why, & my brain is now irreversibly poisoned as a result of getting used to it.
3. [↑] A phoneme is an abstraction of a segment (speech sound), defined by whether it’s consistently distinguishable from other phonemes in the same lect. For example, the ⟨i⟩ in bit & the ⟨ea⟩ in beat are — in spite of their similarities — considered different vowel phonemes in English, because most English speakers agree that bit & beat aren’t homophones. This is called a minimal pair.
4. [↑] Hey — I’m not complaining. These days, we have robots that sound suspiciously like humans sometimes, & cresting yon side of the uncanny valley is pretty creepy if you ask me.
5. [↑] Or, a phonetician, rather. You know what I mean.
6. [↑] You should know that the notion of lect is very abstract & collectively includes things that we commonly refer to as “language groups”, “languages”, “dialects”, “registers”, “styles”, “vernaculars”, etc.
7. [↑] Usually transcribed as /⁠ɒ⁠/ in the Northern Hemisphere. This is the same as the CLOTH vowel (thus also typically /⁠ɒ⁠/) in most — but not all — dialects that have a more-than-two-way distinction between low-back vowels.
8. [↑] I’m ignoring the fact that some American English speakers have an /⁠l⁠/ in this word. If palm has an /⁠l⁠/ for you, then consider haram instead.

scroll_strategist finally gets lots of much-needed love

A while back (long time ago; far, far away; etc.), I put together scroll_strategist, a program capable of generating provably-mathematically-optimal MapleStory-scrolling strategies.

scroll_strategist itself is a library, & its only frontend is scroll_strategist_cli, where the ⟨cli⟩ stands for “command-line interface”. So unless you compiled the CLI yourself, you’ve probably never used it. Still, it has been usable for a while now, but I was never satisfied with it having only a single optimisation mode. So I added a new one.

It took a lot of thinking about it (a lot 😔), a lot of writing computer code that wasn’t very good & then making it better many times over, & a lot of rewriting the READMEs for both scroll_strategist & scroll_strategist_cli, but I basically did it. I’m proud to announce that with both of scroll_strategist’s optimisation modes combined, it’s now a very valuable program to help you with your scrolling! ✨

If you want to know more, then you should read the READMEs: scroll_strategist’s, & scroll_strategist_cli’s. But here, I want to do the “executive summary” version, which doesn’t assume that you know what “dynamic programming” is, nor that you know what a “mathematic”(???) is, nor that you know what a “““computer code””” is. The computers speak in strange tongues, & I don’t wish the ability to speak those tongues upon even my worst enemies.

So here’s the deal: I’m going to explain the basic intuition behind what scroll_strategist does, then I’m going to give a few tips about how to use it, & then I’m going to ask a very particular nerd question about the details of what scroll_strategist does.

What does scroll_strategist even do?: The executive summary version

scroll_strategist is a horticulturist

When we scroll an item, there are many forks in the road. At some of the forks, we make a choice: what scroll should I use next? At other forks, the choice is made by the fickle finger of Tyche [Τύχη] herself: which random outcome of the scroll-use actually occurs?

Because we’re often scrolling many slots on an item — sometimes even as many as ten — these forks that keep forking themselves over & over make quite a large & tangled tree. With just a little bit of ✨magic✨, scroll_strategist takes absolutely every possible fork in the road. As a result, it sees every possible outcome of every possible scrolling strategy, & knows exactly how to get there.

Because scroll_strategist knows how to get to each outcome, it knows precisely how likely that outcome is: if the path to get there involved a lot of unlikely things, like passing 30% scrolls, or failing 70% scrolls, or what have you, then that individual outcome isn’t all that likely. It also knows what was spent to get there: it knows exactly what scrolls were used along the way. This information can then be used to work backwards: if two outcomes look similar or identical, but one of them is at the end of a less savoury path — because that path is unlikely, or costs many mesos, or some combination of both — then the less savoury one can be eliminated, or pruned.

Importantly, these prunings can only occur at the forks where we make a choice. The fickle finger of Tyche simply isn’t within our control. Nonetheless, once the prunings are completed, we have a much smaller tree of forking paths than the one that we started with, because each fork in the road where we could make a choice is no longer a fork: we now have a unique strategy that tells us exactly what to do there.

Value judgement

Perhaps you can now see why scroll_strategist produces “provably mathematically optimal” strategies: it knows about the entire tree of possibilities, which allows it to determine the optimal strategy at each juncture, as it walks all the way “back” to the beginning (where we haven’t started scrolling yet). There’s just one problem: how do we decide what makes one path “better” than another?

As it turns out, the answer to this question is a lot more complicated than you think. scroll_strategist makes at least one crucial simplification: it assumes that you (the scroller) have a concrete goal in mind, such that the goal can either be met, or not met. There is no “in between”. This goal is expressed in terms of the item’s stats: if every one of the item’s final stats is at least as large as the corresponding goal for that stat, then the goal is met; if one or more of the stats fails to meet its goal, then the goal is not met.

With this simplification, the original version of scroll_strategist had just one “optimisation mode”: maximise the probability of reaching the goal. Then, it only cared about costs if necessary to break a tie. This mode is called SucProb, or suc-prob. By combining the probability information with the cost information, the new mode instead minimises the average cost spent to reach the goal once. This mode is called SucExpCost, or suc-exp-cost.

Using scroll_strategist

Both optimisation modes are well-defined & fairly simply defined. They also correspond to things that we really do frequently want to optimise for. However, these definitions are simple because they are simplifications for the purpose of mathematical optimisation. If you get a strategy from scroll_strategist that you don’t like, it’s probably because you’re actually — whether you explicitly acknowledge it or not — asking scroll_strategist for something other than what you really want in your heart of hearts.

Many, if not most, realistic scrolling situations have characteristics lying somewhere in between the assumptions made by SucProb & SucExpCost. It’s often a good idea to at least consult both optimisation modes, unless you’re absolutely certain that one mode is preferred over the other. The assumptions made by each mode can be couched in ordinary intuition:

SucProb

You have only one (1), or simply very few, of this equipment item. You cannot afford to fuck up this one equipment item, & are thus willing to spend a bit extra on scrolls if it means that this particular scrolling attempt doesn’t fail to meet your goal.

SucExpCost

You have a virtually unlimited number of equipment items similar to this one, so if this attempt fails, you can always (always!) just try again. You’re thus willing to accept strategies that are unlikely to succeed, as long as those strategies are inexpensive enough to make up for it.

Because SucExpCost is really all about cost, you should take note of the initial value given (by scroll_strategist_cli) for “Expected total cost per success”. If the “Expected further cost per success” ever exceeds that value, then you should probably give up & start over, if possible! scroll_strategist_cli will warn you about this by putting “(!)” next to the “Expected further cost per success” value.

Let’s take a brief look at two concrete examples that showcase the strengths & pitfalls of these modes.

Example 1: AVOID weapon

I needed to scroll a weapon for AVOID. The natural choice in MapleLegends is the Christmas Tree, a level 10 dagger with no job requirements that can come with 9, 10, or 11 AVOID clean. Being a dagger, it can also be scrolled for AVOID, because dagger scrolls (excepting 100%s) grant LUK. The input to scroll_strategist_cli looks like this:

{
  "stats": [22],
  "slots": 7,
  "scrolls": [
    { "percent": 10, "dark": false, "cost":   25000, "stats": [3] },
    { "percent": 30, "dark": true,  "cost": 2000000, "stats": [3] },
    { "percent": 60, "dark": false, "cost":   20000, "stats": [1] },
    { "percent": 70, "dark": true,  "cost":   50000, "stats": [1] }
  ],
  "goal": [31]
}

This one is 11 AVOID clean, which is why it has 7 slots, & why its stats vector is $\left[22\right]$. Each point of LUK is worth 0.5 AVOID to me, so it’s easiest to consider the stats in terms of virtual points of “LUK”, where 1 AVOID is worth 2 LUK. The costs are expressed in terms of typical market prices in mesos at the time.

I don’t specify the upfront cost, so it’s zero by default. This will turn out to be a problem, but upfront cost is difficult to estimate here: Christmas Trees are, in reality, a limited resource, but at the same time, each individual one isn’t “worth” much at all, because it’s a dinky level 10 weapon that noöne (aside from me, apparently) cares about.

If I try to minimise the expected cost to reach my goal by specifying the SucExpCost mode, then scroll_strategist_cli’s initial output looks like this:

Probability of success: 2.879%
Expected cost: 328_219.7
Expected cost per success: 11_400_722.5
Next scroll to use: 70%

Not gonna lie, 2.879% ain’t great! That means it takes 34〜35 attempts on average, which is very many Christmas Trees! And perfect clean ones, at that!!

But it gets worse: we can actually refine SucExpCost’s ability to optimise, by allowing it to use a so-called “null scroll” that has a 100% probability of success, is not dark, costs nothing, & grants no stats upon success. scroll_strategist_cli makes it easy to add this scroll by just passing the -z/--null-scroll option to it. The point of a null scroll is to allow the optimiser to stop scrolling when it has already reached the goal, thus lowering costs even further. With this newfound power, SucExpCost finds an even more optimal solution:

Probability of success: 2.871%
Expected cost: 114_703.8
Expected cost per success: 3_995_618.9
Next scroll to use: 10%

Notice how the expected cost per success (i.e. the thing that we’re optimising for) goes from ≈11.4M mesos to just ≈4.0M. That’s like ≈2.9 times better! However, the probability of success has suffered (albeit not by much in this particular case), & now scroll_strategist wants me to start with a 10% scroll for some reason? Ouf.

In the end, I just used SucProb…

Example 2: Estimating market prices

Example 1 might have made SucExpCost look not so great. But it actually is great! You just gotta know how to use it.

One good example that I ran into recently was estimating the market price of my Stolen Fence that I don’t use anymore. It has 15 WATK, 8 STR, & 0 slots. Normally, I know roughly fuck all about what anything is worth on the market. But with SucExpCost, all I had to do was input some scroll prices that I plucked from OwlRepo, & scroll_strategist was able to nigh instantly tell me how many mesos it takes, on average, to produce a Stolen Fence like mine. That’s crazy useful!!

How large is the tree?

ℹ️ Not a nerd? Go ahead & skip this section. It’s okay. I won’t tell anyone.

In the “scroll_strategist is a horticulturist” section above, I explained how the forking paths of scrolling possibilities make a large tree of many branches. But how large?

We’ll let our set of scrolls have cardinality $𝑠$, & the $𝑖$^th scroll in the set have ${𝑜}_{𝑖}$ possible outcomes. We’ll also say that, out of the $𝑠$ scrolls in the input set, $𝑑$ of them are dark, where $0\le 𝑑\le 𝑠$. Assuming that every dark scroll usage has exactly three possible outcomes (viz. pass, miss, or boom), we can construct the tree as follows.

We start with a graph that has one vertex & zero edges. The vertex represents the starting state, i.e. the equipment item before we begin to scroll it. We then branch $𝑠$-fold to represent scroll choice; the graph now has $1+𝑠$ vertices & $𝑠$ edges. We’ll see that each new generation adds exactly as many edges to the graph as it does vertices, because each vertex in the new generation connects to exactly one vertex of the previous generation (this is a tree, after all); thus, we can just keep track of how many vertices there are, knowing that the number of edges is just one fewer.

From this second generation of $𝑠$ vertices, we branch ${𝑜}_{𝑖}$-fold to represent scroll outcomes; the graph now has

vertices. To make things easier, we can define:

Now that we’ve three generations, we can see that the latter two of the three generations represents exactly one (1) scrolling event, meaning that we picked a scroll, used that scroll, & obtained whatever outcome it gave.

From this third generation of $𝑂$ vertices, we want to once again branch $𝑠$-fold to represent scroll choice, but we can’t quite do that. This is because some of the vertices in the third generation represent boomed items, which are leaf vertices that have no descendants. Because there are $𝑑$ dark scrolls in the scroll set, & each one has exactly one boom outcome, the number of vertices in generation three that aren’t booms is $𝑂-𝑑$. Thus, with the addition of generation four, we now have $1+𝑠+𝑂+\left(𝑂-𝑑\right)𝑠$ vertices. Branching ${𝑜}_{𝑖}$-fold again to add a fifth generation, we then have $1+𝑠+𝑂+\left(𝑂-𝑑\right)𝑠+\left(𝑂-𝑑\right)𝑂$ vertices. This concludes the second scrolling event.

Generation six itself (i.e. not the entire graph) then has $\left(𝑂-𝑑\right)\left(𝑂-𝑑\right)𝑠={\left(𝑂-𝑑\right)}^2𝑠$ vertices. Generation seven itself has ${\left(𝑂-𝑑\right)}^2𝑂$ vertices. This concludes the third scrolling event.

Now it’s fairly clear where this is going. Because each scrolling event is split into a pair of two generations of different types, the formula for how many vertices there are in the $𝑛$^th generation depends on the parity of $𝑛$:

The above formula works for any value of $𝑛$ except 1, so we need the special case $𝑔\left(1\right)=1$.

But we want to know the total number of vertices in the entire graph, after zero or more scrolling events. Thus, we basically end up with two partial geometric series that need to be summed, & then we can add those two sums together to get the final result. By using the closed-form formula for summing the first elements of a geometric series, for the even generations we get:

And for the odd generations…:

As it turns out, the number of edges in the graph being one less than the number of vertices actually makes the edges easier to count, so here’s a formula for the total number of edges:

Actual numbers

As expected, this is exponential in the number of scrolling events — that is, in the number of slots that we start with. That is to say: it’s a very big tree. But what does a realistic concrete example look like?

Let’s consider a “realistic worst-case” (not actually worst-case) scenario. Ten slots. The scroll set will then have one 10% scroll, one 30% scroll, one 60% scroll, one 70% scroll, & one 100% scroll. The 100% scroll might be an actual 100% scroll, or it might be the null scroll. Sound reasonable? This means that we have $𝑛≔10$, $𝑠≔5$, $𝑂≔11$, and $𝑑≔2$. The number of edges in the full unpruned tree is then:

That’s roughly seven billion, or ≈$7\times {10}^9$. That is, notably, a bit larger than $2^{32}$, meaning that we’d need more than a four-byte integer just to address the vertices alone, even if they ✨magically✨ took up no space!

So… yeah. That’s why scroll_strategist needs a little ✨magic✨ to actually work.

Future improvements

There’s theoretically the possibility of adding support for exotic scroll types like CS & WS. The problem with WS is perhaps obvious: it won’t always decrease the number of slots that the item has (or boom it), so there’s no guarantee of forward progress, & thus no guarantee of termination. But it seems to me that this can be solved by not treating them as individual scrolls, & instead treating e.g. “using a WS-&-10%” as actually spamming the WS-&-10% over & over until something besides a miss occurs (viz. boom or pass). CSes, on the other hand, are difficult just because of how many possible outcomes they have…

Apart from adding a new optimisation mode, & few features to the CLI, I also did some program-optimisation work to make scroll_strategist a bit more efficient — especially when it comes to heap memory usage. Although this is obviously beneficial in general, I did it particularly with an eye towards giving scroll_strategist a webpage frontend! That way, people can actually use it‽

Don’t get your hopes up, because that’s kinda a lot of work. Like, too much work. But who knows? Maybe there are people out there who’re actually interested in using it…

The blood cleric

ℹ️ This is a historically-focused companion essay to the “Heal” essay of pt. cviii of this diary.

MapleStory has its own set of urban legends, myths, & the like. A folklore, if you will. Believe it or not, one such legend is — although not necessarily odd-jobbed — apparently(?) odd-job-adjacent: that of the blood cleric.

I’ve had multiple people independently bring this topic up to me over the years, as it seems that everyone who played pre-Big-Bang (pre-BB) MapleStory in its official retail incarnations was at least exposed to the notion. Yet, when we ask the basic question of “what is a blood cleric?”, noöne can seem to give a straight answer. What gives?

It looks like it’s time for some investigating.

Attestations

ℹ️ The bulk of this essay focuses on establishing historical attestations of the blood cleric phenomenon. Skip ahead to the “What is the blood cleric?” section below if you’re uninterested.

Advanced Build:Blood Bishops.

Advanced Build:Blood Bishops. (henceforth “ABBB”) is a guide posted to the Asiasoft forums on 2008-02-01 by MapleSEA player StarryNights (IGN Mizmoonlight). It begins as follows:

Just want to share with my fellow friends, a not-so-new build for bishops.

This build has been around for quite some time already, priests adding hp so that they can survive the respective mobs themselves (among all other reasons)

However, most of them didnt really got pas[t level 120〜129], mostly because their [total MATK] is low, demotivated, blabla yadayada.

ABBB is very simple: it advises the reader to play as a pure INT mage up until level 135〜140, at which point the reader starts adding all AP from that point onward into MAXHP. There’s no funny business with saving AP, no washing — it’s just as simple as it sounds.

Being pure INT until such a late level, the blood bishop following ABBB will still end up reaching 30k MAXMP, & will still have the TMA to train efficiently on Skeles^[2] et al. with their Genesis. The extra MAXHP gained later on is then supposed to be used mostly for saving mesos:

• survive 2 hits on mongs^[3], at lv143〜145 W/O HB. (early!)
• survive 3 bumps at skelegons(or 1bump+1magic), at lv 155~ W/hb. (mg off )
• mgless tengu with dk’s (tengu is just like another mong....you guys can train like normal)
• mgless papulatus/pianus ( you will have 4k+ base hp, dispel is nothing!)
• perfect potless zakum (with 6.5k++hp, even bumping into the body isnt that fatal anymore)
• mgless horntail? (its very possible, at the non-dispel part. since generally, your hp will be as much as a range class. if a range class can survive, then blood bishops can go mgless)

Some of these touted benefits are perhaps more questionable than others, but the basic idea of saving lots of pots is clear. There’s also the added survivability in the presence of dispels, especially when it comes to Horntail.

The Paths from Cleric, Priest, to Bishop

The Paths from Cleric, Priest, to Bishop (henceforth “The Paths”; a Brazilian Portuguese translation is available here (archived)) is a guide posted to the PlayPark^[1] forums on 2007-04-26 (edited as late as 2008-02-19) by MapleSEA player Pearlite. As its name implies, this is a much larger, generalist guide to the bishop throughclass.

The Paths categorises AP builds into four groups: “normal” (just enough LUK to wear level-appropriate gear), pure INT, “blood priest”, & “blood bishop”.

The blood priest is mostly or entirely INT until level 70, at which point all AP starts to go into MAXHP, until an unspecified level when the player deems their character to have enough MAXHP, at which point AP starts going into INT (& possibly some into LUK) in perpetuity. The blood bishop is not explained by The Paths directly, as it instead simply cites ABBB.

The justification for the blood builds is similar to ABBB’s justifications, but The Paths makes it more clear why being able to take more hits without potting is so beneficial:

• Better attack pattern at certain mobs after achieving certain amount of HP

  Although blood priests’ damage is lower, they can have a better attack pattern than normal priests at mobs like grims and monglong^[3]. At certain stage when their HP is sufficient to survive 2 hits WITHOUT HB, they can cast shining ray twice then only heal themselves (do 2SR-1heal-2SR-1heal pattern), whereas normal priests can’t do that (only can do 1SR-2heal-1SR-1heal…, amount of heal depending on situation).

This implies that when grinding (at least at relatively hard-hitting monsters like e.g. Grims & Mongs^[3]) under the assumption of minimal potion expenditure, the blood priest can actually do as much — or more — DPM as the ordinary priest.

Unlike ABBB, however, The Paths addresses a major component of the blood cleric legend (emphasis in the original):

• Lower healing rate / HP recovery rate

  The amount of HP that you can heal depends strongly on your [MATK], NOT your HP. So, it’s not wise at all to create a blood priest with too little [MATK] as your healing rate will suffer and you might have a hard time healing your party mates.

[…]

The higher our heal level, the higher the amount of hp that we can recover. HP might affect the amount of hp healed, but the effect is very insignificant. The amount of HP healed strongly depends on your INT/[MATK]. There’s no point having so high HP but so little INT/[MATK] as the amount of HP that you can heal will suffer.

†There was someone from sleepywood forum training an abnormal cleric. The cleric only pumped 20 INT and the rest of the AP were added to HP, so he has a lot of hp but very low INT. His heal at lv 40 can only recover 100+ hp. He can’t even recover his own hp to full although his heal is maxed. Pathetic, isn’t it? [Link: http://forums.playpark.net/showpost.php?p=358496&postcount=10]

I have personally tested the amount of hp healed with a DK. I’m a normal build priest. His hp was [12 000〜12 999] after HB (I forgot to note down the exact number. Sorry). My HP was 1668 (full hp) when I did the test (1668 because I off HB) and his hp after using 2× dragon roar was 4982. I could easily heal his HB-ed hp to full even though I was not HBed. Assuming his HB-ed hp was 12,000 (more than that actually), thus I healed at least 7018 hp. According to heal skill description, it seems that I can only recover 5004 hp (3×1668). From this little ‘experiment’, we can conclude that max heal does NOT mean that the amount of hp recovered equals to 3 × our own hp.

In fact, without testing this out, I can feel that I can recover much more than 3× my own hp as I can easily heal a HB-ed full party to their full hp with ONE cast of heal. This shows something, doesn’t it? INT is very important to have decent hp recovery rate.

If you insist that 300% recovery rate means 3 × HP and INT/[MATK] does not affect the amount of hp healed, this video can prove you wrong:

[…]

If you look closely, the video shows that the lv45 cleric has MAX heal but only 27[ MATK]. Notice how pathetic his healing capability is? He can’t even recover his own hp to full. He needs many heals to fill up his hp bar!

†The referenced comment/post is dated 2007-05-21, although it does not link to the Sleepywood thread that it’s reporting from. The Paths misquotes the comment as saying “100+ hp”, but the original actually says “less than 100 hp”.

Does HP — be that HP or MAXHP, of the cleric themself or of the beneficiary — affect how much Heal heals? After all, the skill description does say “Recovery rate 300%”. According to The Paths, the answer is “technically possibly maybe very little… but no, not really”.

Apart from the cryptic skill description, the real problem here is as stated in the “As weal” section of pt. cviii of this diary: the magnitude of Heal’s healing is determined solely by the server, thus making it totally opaque in the absence of server source code leaks. Nevertheless, known implementations do not take (MAX)HP into account, & we seem to have no good reasons to believe that any implementation ever has.

Still, myths can persist. The Sleepywood forum thread referenced here was likely motivated by this exact myth.

Cleric Guide

Cleric Guide (archived) is a guide posted to Hidden Street sometime prior to 2009-12-28 by IGN SunMage of the GMS Scania server, with authorship credits given to greenstriker (who is probably SunMage…?), BladeOracle, yespotatoes, & kevin y. The post that’s accessible now (including through archives) was reposted by a dummy account named guides; all of guides’s posts are dated somewhere between 2009-12-28 & 2009-12-29.

Cleric Guide is actually from 2007 at the latest, as it antedates the release of 4^th job in GMS. The original version of the guide (before any updates to it were made) possibly antedated the release of 3^rd job in GMS, given that the “Priest Guide” section begins with the text “Yay! This is newly added.” — this would place its earliest version sometime in 2005. However, I consider this somewhat unlikely; it’s more likely that the authors simply hadn’t got that far yet when the first version of the guide was posted. In any case, this last version of Cleric Guide has “As of now, Global MapleStory is in version .38” buried in the middle of the guide, putting at least one version of the guide somewhere between 2007-05-02 & 2007-06-12.

In any case, this guide goes on to describe the cleric’s arsenal:

Heal—a unique skill for clerics in which you heal a certain percentage of your hp, and the undead monsters in the area will also take damage. […] The higher the lvl of this skill, the more hp you heal of you and your party members. (300% heal rate=300% of YOUR HP, this applies for party members being healed by you) Heal damage is based on your intellegence. That 300% thing is 300% of your hp divides among the party.

Already we have a claim that, although Heal’s damage is based on “intellegence”, its healing amount is apparently based solely on the level of the skill & the cleric’s “HP” — although it’s unclear whether this means MAXHP or HP. This would appear to be pretty good evidence that, at least at some point in its pre-BB history, GMS had a Heal healing formula based on the cleric’s (MAX)HP. However, it’s difficult to take this seriously when the same guide makes other clear errors in its understanding of the game mechanics:

• They claim that Heal damage is just based on INT, but in pre-BB GMS, it was also based on LUK & TMA (& target count & skill level, for that matter).
• After noting that “[p]arties will adore and invite you to their party because of this skill”, they claim that “Bless stacks (means that you can get it with other abilites that other party members have) with everything except magic armor”. But this is false, for client-sided reasons at the very least.
• They claim that Shining Ray “attacks up to 5 monsters at once”, mirroring their claim that Heal does the same (the latter of which actually being true, under the right assumptions), but this has never been true for any level of the skill in any version of the game.

This guide makes reference to blood priests twice, both within the same context of training at Gobies (meaning Deep Sea Gorge II) & graduating therefrom:

One very important note: if you are not a blood priest, either get a DK to hb you all the time, or use MG.

[…]

Once a (non-blood) priest gets to [level 110〜129], they can probably begin soloing places like grims if they have maxed Dragon to help them out.

HealingGlory’s All-In-One Cleric Guide! :D

HealingGlory’s All-In-One Cleric Guide! :D is a guide posted to BasilMarket under the title MapleStory Guide: Clerics sometime prior to 2010-02 by BasilMarket user GodsWarning. Its author is IGN HealingGlory of GMS, who originally titled it HealingGlory’s Ultimate Big Guide On Clerics! :D. osukun, Skrato, peter90, Oogleberry, Banditb0ix, dfz, & Mufalopadus are given editorial credits.

HealingGlory (the PC) dates back to ca. 2006-02. The guide itself at least slightly antedates the introduction of 4^th job into GMS, although one of its touted virtues is that its author continually keeps it updated.

All information WAS taken from my experience and my research during the time span of [2007-12-01] to [2008-01-01].

HealingGlory partitions cleric builds into “many” types apart from the elaborated standard build, listing “blood cleric”, “money cleric”, “damage disher”, “and many others”.

BLOOD CLERIC, A Cleric who puts points into HP for the sole purpose of successfully bossing

[…]

INSTEAD of adding on the INT and LUK, you add all AP on HP. (Of course, you start doing that AFTER you pump up 20 INT points to [1^st-job] advance).

The skillbuild is the same as any other cleric but you don’t use the attacking skills.

Pros:

• Has a lot of HP

Cons:

• Can’t do much magic attack. Warriors are a good alternative to the Blood Cleric.

This guide also dedicates a section to MAXHP gear:

Why? Simple. Because it keeps you Alive. Yes, yes, go and scream, but in reality HP equips are NEEDED for survival in priests!

“Idea for ultimate Party Cleric…”

A thread by stewie_gryffin on the GameFAQs Message Boards, posted on 2008-03-11 with the title “Idea for ultimate Party Cleric…” (archived), begins like so:

Since Heal is based on the Cleric’s health, what about getting just enough int to wear gear, and pumping the rest of the AP into Health?

I know this sounds dumb, but it would maximize the amount of health recovered, and would give the Cleric/Priest/Bishop more survivability.

The first reply, by Mikalot, immediately contradicts:

HP affects it as much as int? Since when?

To which stewie_gryffin responds:

Well, it says “Recovery Rate ***%”, and I thought that was of own HP. Therefore, if you have 1000 HP, it heals 3000 at Level 30 Heal. So I figured, if you had 2k HP, it would heal 6k. I didn’t think that Int affected Heal?

This appears to be a simple misunderstanding, although Oscarbear007 chimes in with…:

Hp only effects [sic] heal on party members. Int helps its actual attacking power if I remember right.

…And is backed up by AjaxTheGreater. basakboy copy-&-pastes a snippet of The Paths. ManchineelTree chimes in:

I’m fairly certain that HP doesn’t affect healing rate. My HP is 1750 right now, and I can heal a DK with HB for 16k in one shot. I have no clue what the percentage in the skill description means, but the amount you heal is chiefly based on INT/[MATK], and slightly on LUK.

soundecho then ends the thread by copy-&-pasting the Heal formulæ from Nekonecat’s 2007-10-14 Sleepywood thread, which indicate that HP & MAXHP are not taken into account by Heal.

Other miscellaneous attestations

In chronological order:

• Hidden Street forum signature of GMS Scania player keunwp (IGN Allintone), dated somewhere between 2006-08-01 & 2008-02-04:

  Lv.90 Pure Int Blood Priest
  Surviving Two Hits From Grim ^^v

• Reply to “MapleStory!” (archived) thread on SpongeBuddy Mania Forums by maxmyles, dated 2007-06-06:

  first of all, i had a magician long before anyone (besides dts) had one, so i know quite a lot of things that banoon probably doesnt. the first thing is, hp and mp are both ok to add. hp is the key ingredient to a blood cleric. mp is a very good thing when it comes to magicians. as for str and dex, your right, none of those.

• Reply to “New jobs” (archived) thread on MMORPG.com Forums by Korby, dated 2007-08-06:

  [⋮]

  24. Fire/Poison Mage

  25. Lukless Fire/Poison Mage

  26. HP Warrior

  27. Blood Priest (Not well-known, I only know of 1 person who has this)

  If 27 ain’t enough for you then bug off. Nuff said.

• Reply to “Thông tin mới và bàn luận về VMS” (archived) thread on Diễn đàn GVN (“GVN Forums”) by VMS player ShakeBottle, dated 2008-05-22:

  Cleric là dòng đầu tiên của mage có khái niệm pure-int vì cleric ko quá phụ thuộc vào đồ đạc. mua AP reset về cho luk = 4 luôn ^.^

  Có ai ở đây dám chơi Blood Priest ko

  ---

  Cleric is the first throughclass [dòng] of mages to have the pure INT concept, because clerics are not overly dependent on equipment. Buy AP Resets to reset your LUK = 4 straight away ^.^

  Does anyone here dare to play Blood Priest?

• Reply to “List of novelty builds.” (archived) thread on GameFAQs Message Boards by solborimm, dated 2010-11-19:

  No blood priest? 4 INT 1 HP per level?

What is the blood cleric?

Clearly, a blood cleric is a PC of the cleric throughclass who permanently adds some significant amount of their AP into MAXHP. However, we’ve seen various notions of what the actual details are. In approximately chronological order:

• By sometime(?) prior to 2007-05-21, we see a single example of a level ≥40 cleric who is pure MAXHP (20 base INT & ≈4 base LUK, etc.).
• By 2007-04-26, we see the notion of “blood priest”, who is mostly or entirely INT until level 70, at which point all AP starts to go into MAXHP, until an unspecified level when the player deems their character to have enough MAXHP, at which point AP starts going into INT (& possibly some into LUK) in perpetuity.
• By sometime(?) between 2006-08-01 & 2008-02-04, we see a single example of a level 90 “pure INT blood priest”.
• By late 2007 or earlier, we see the notion of “blood cleric”, who has 20 base INT & puts the rest of their AP into MAXHP.
• By 2008-02-01, we see the notion of “blood bishop”, who plays as a pure INT mage up until level 135〜140, at which point they start adding all AP from that point onward into MAXHP.
• By 2010-11-19, we see the notion of “blood priest” who adds +4 INT & +1 MAXHP per levelup.

Not “what”, but why is the blood cleric?

We can boil down the justifications for blood clerics as follows:

1. The cleric’s damage output is not valued in the way that other throughclasses’ damage outputs are. Many guides make distinctions between “solo” builds & “party” builds, and many guides discuss training locations in ways that centre around what kinds of physical attackers — the presumed primary damage-dealers — are in the cleric’s party.
2. Saving mesos. By turning off Magic Guard & healing oneself with Heal, any MP expenditure is easily recovered by MP Eater, & potion consumption is effectively or exactly zero. Investing AP in MAXHP often allows for the “turning off Magic Guard” part, & allows for the less frequent use of Heal (thus making room for the use of e.g. Shining Ray).
3. Survival — especially in bosses. Sometimes — e.g. when dispelled by a boss like Pianus (R) or Papulatus Clock — you don’t get a choice about your Magic Guard bring turned off. Moreover, even with the assistance of skills like Magic Guard & Invincible, there are almost always things out there that will smack you within an inch of your life — or directly into the grave. This is more true the lower level you are.
4. Hoping that (MAX)HP affects the magnitude of Heal’s healing. Heal’s skill description is impressively misleading, even by MapleStory standards. Moreover, the server-sided implementation of Heal’s healing calculations is opaque to players.

When it comes to (2.) & (3.), we crucially must bear in mind the environment that the blood cleric exists within:

• The blood cleric lacks access to HP & MP washing entirely, & probably has subpar access to MAXHP gear as well. MAXHP is just a stat that you can assign your AP to, like any other.
• The blood cleric lacks easy access to mesos & potions, & probably hopes to use this character as a meso-maker. In a server like e.g. MapleLegends with easy mesos, 4^th job (read: Genesis), & the ability to sell leech (not to mention the presence of HP washing), the idea of generating mesos this way, or needing to conserve mesos in this way, is absurd. But there is much less absurdity within the proper context.
• The blood cleric is underlevelled. Levelling up ain’t easy, but we all wanna be the cool priest who can train in Deep Lūdi, or who can take on the legendary Zaqqūm^[4], or whatever.
• The blood cleric might be dead. With no access to HP washing, no access to autopot, poor access to MAXHP gear, poor access to potions/food, poor access to EXP, & very little understanding — within the general population — of the ways of staying alive in MapleStory combat, adding AP into MAXHP is a pretty good strategy to avoid headstoneification. For players with easy access to autopot & to potions like e.g. those sold by Miki, this is difficult to appreciate.
• The blood cleric sometimes just doesn’t know what they’re doing. This is generally not the case with well-written guides that attest to the existence of the blood cleric, but certainly some cases can be attributed to ignorance, especially ignorance of how Heal’s healing formula actually works. This is understandable given that we’re usually talking about a time period when pre-BB MapleStory was much more of a mystery in general.

And when it comes to (4.), we have lots of good attestations, from GMS players & from MapleSEA players, stretching back to early 2007 & possibly even earlier, that firmly dispel the myths of Heal’s healing formulæ. I think that we can now safely declare, beyond any doubt that’s not purely baseless speculation, that the formulæ for Heal, in any version of any retail implementation of pre-BB MapleStory, have never observed HP — be that MAXHP and/or HP, of the cleric and/or of the beneficiary — in any way in the process of calculating the magnitude of Heal’s damage and/or healing. Still, if you’ve any Korean-language sources, I’d love to see them. 🙂

Footnotes for “The blood cleric”

1. [↑] Asiasoft is now known as Asphere Innovations, & was/is MapleSEA’s publisher. PlayPark™ is a trade name (or similar) of Asiasoft.

2. [↑] Meaning Skelegons & Skelosauruses at Nest of a Dead Dragon or The Dragon Nest Left Behind.

3. [↑] GMS players & MapleSEA players seem to agree upon not using the English-localised name Dreamy Ghost, although MapleSEA players go for Mong rather than the GMS choice of Hime [姫(ひめ)].

   Mong is from apocope of KMS’s name for this monster: 몽롱귀신 ⟨Mongnong-gwisin⟩ /⁠moŋ.ɾoŋ.kwi.sin⁠/ [-noŋ.gɥi-] “hazy ghost, vague ghost”. The use of 몽롱- here is somewhat uncommon; it’s Sino-Korean 朦朧 “faint, dim, obscure; vague; hazy”. 몽 is more typically a reading of 夢 “dream”. The similarly-named 몽달(dal)귀신 (there are no monsters with this name in KMS) refers to the vengeful ghost of a bachelor who died before being able to marry. 몽달 is probably Sino-Korean as well, but its actual etymology is unclear to me.

4. [↑] See the “Is Zakum the tree, or the statue?” section below.

The spoils of wintertide

In the “Holiday spoils” section of the previous entry in this diary, I briefly mentioned some of the rewards available to us as part of MapleLegends’s 2023–2024 Ḥanukkah–Maplemas–Versalmas event. As explained in that section, I won’t be doing a painstaking enumeration of every reward that I got — apologies to any statistics nerds reading — but I will do a brief overview.

From the Snowman Boss PQ (see the “The abominable snowman” section of the previous entry), I ended up with some ≈35 or so MCP1s, netting me something like 35 ⋅ 2 ≈ 70⁢ pCoins.

That’s already very substantial indeed! But the same PQ also awarded Holiday Surprise Boxes of varying sizes, which meant that I ended up with dozens of scrolls: 60%s, 70%s, & 30%s galore. Selling these off (apart from the garbage WACC scrolls, & the not-that-useful ones that I decided to just hoard) in the Free Market provided a source of additional joocy mesors.

And then there are the raffos^[1]. So, so many raffos. Many of these raffos were obtained from monster drops, or from ordinary year-round PQs. But most of them were from Snowman, of course. In total, across all of my characters, I turned in probably ≈250-ish raffles. My god, that was a lot of raffling.

The most significant raffle pulls were significant either because they gave yellow lines that were displayed to anyone lacking the taste to mute smegas, or because they were chairs. Chairs are, after all, what Maplers truly crave… The irony of course being that yellow lines are used to spam Maplers’ chatlogs even when the associated item is nearly worthless. Here are my “significant” pulls:

That’s four (4) yellow lines, of which only one (25%) was really worth anything. And even then, an 8 MATK clean PuAC is still only worth a handful of millions of mesos. On the other hand, that’s also 15 whole chairs! Let’s take a look at those chairs, shall we?

First up is the Holiday Latte Chair:

Yes, this chair does put you in a Rudolph costume, complete with antlers, a glowing red nose, hooves, & even a li’l tail. I’m already a deer, so it’s kinda redundant? But it’s still cool, I guess.

Then there’s the Cactus on Ice Chair:

That is indeed a Jr. Cactus — baby cactus diaper & all — encased in a giant block of ice. What else is there to say?

And finally, the Cocoa Sheep Chair, which is the best one. What, you don’t believe me? All three chairs are animated, but in order to demonstrate the true beauty of the Cocoa Sheep Chair, I need to show it to you in all of its fully-animated glory:

Oh, my god. The sheep even wiggles their hind legs. It truly is the chair of infinite cosiness. Thank goodness this was the chair that I got eight of!

But those weren’t the only things that I got from the raffle. I also ended up with a lot of CSS1/CSS3s, Green Mittens, various USE items, & of course, socks. Lots… & lots… of sox.

The Greatest Reward Of All Time™

In the “Decorating the tree” section of the previous entry in this diary, I joined Lvl1Crook (Level1Crook, xXCrookXx), Taima (Yunchang, Tacgnol, Boymoder, Girlmoder), & LawdHeComin (AppleBasket, pleLeg, JunkyardBat, Snel) in tree decoration festivities for MapleLegends’s 2023–2024 Maplemas–Versalmas community event.

I may or may not have been really hoping that we’d place in that top three, so that I could get The Greatest Reward Of All Time™. And guess what? We focking did‽‽‽ F5555555555555555555555​555555555555555555555

It’s a big chair? Made of a gorgeous ice sculpture? And you sit on a swing? And there’s a mfing glowing snow reindeer who chills on the ground next to you? Oh, and I guess we got, like, game cash & donor points & a special hair coupon. …But the chair, tho.

Bask in the full glory of chair…

…As I do every day now. 🩵

Footnotes for “The spoils of wintertide”

1. [↑] See footnote #1 of the “Dee thirty-four are” section of pt. cx of this diary.

d3[34]r

This episode of Viclocking With deer™ (“your only source of Vicloc news!”^®) is very simple: APQ, APQ, & — you guessed it — even more APQing.

I figure, if I’m gonna be out there doing a bunch of APQing, then I’d better at least look ✨good✨ in front of all these APQ strangers. By pilfering various items from the Cash Shop, from this past event, & from the Hallowe’en event, I put together a neue look for my Vicloc dagger spearwoman d34r!:

Red! What do you think? I like how the birbs sit on my lap when I sit down in a chair:

Mostly, I did APQs with Taima (Boymoder, Tacgnol) on our various characters. When I accidentally ran out of characters after already joining an APQ party, I logged onto my Vicloc clericlet d33r, & got the chance to APQ with my Vicloc wife VigiI!!:

We might only log in twice a year, but Vicloc luv is 4ever. 💟

El retorno de La Tomatina

In the “The Maple World is under attack!” section of pt. cx of this diary, I chronicled my first experience with a «La Tomatina» event in MapleLegends. Well, for reasons unknown to me, we got another one of these events!

This time wasn’t quite so drawn out, & I didn’t have to abruptly leave after the first ≈20% of it to go run a HT that I’d already signed up for… But I still attended on my Vicloc dagger spearwoman d34r, who was subjected to La seducción de La Tomatina:

But the Giant Tomatoes were quickly taken care of, as a new challenger arrived at the battlefield known only as Henesys: the Pink Bean. That makes d34r perhaps the only Viclocker ever graced with the privilege of dying to the Temple of Time’s most powerful legume:

Now, I might be Victoria’s most powerful spearwoman, but you gotta remember, I still have barely more than 10k MAXHP, even with my HB on.

But, you know, I do have another character with HB… A certain darksterity knight by the name of rusa.

Those G-clef attacks hit diff when you’re in Henesys…

…And Mrs. Ming Ming doesn’t give two shits, apparently.

The only thing more powerful than the Bean itself is, of course, the GM who left-sedded (not left-jump-sedded) everyone…:

Eventually, the Pink Bean had had its fun. As it parted ways with Henesys, it left behind the echoes of victory. Wow…!

Behind closed d00rs

In the “The ascensi0n of d00r” section of the previous entry in this diary, we saw my clericlet d00r advance to the honourable rank of priestlet. Now that d00r is finally starting to put those big girl pants on, it’s time for a very… special episode of d00rin’ With d00r™ — starring your host, d00r.

In this episode, we’re taking “d00ring” very literally, because I hit level 71, & now I have Mystic Door!:

Whoa!! I knew d00r was named that for a reason…

But that’s not all that level 71 has in st00r for d00r. In the clandestine underground laboratory of one Magatian alchemist by the name of Yulete, there awaits a certain MPQ. But first things first: in order to complete all quests associated with MPQ, I would need to at least complete the Keeny’s Research questline up to the beginning of “Keeny’s Research on Frankenroid”.

For “Keeny’s Research on Roid”, I slew a hundred or so lightbulb-headed Roids:

In English, roid /⁠ɹɔɪ̯d⁠/ is usually an aphæretic form of steroid /⁠ˈstɛˌɹɔɪ̯d⁠/, itself an informal abbreviation of anabolic steroid. In the case of these MapleStory critters, however, it’s probably more like an aphæretic form of droid /⁠dɹɔɪ̯d⁠/, itself from the aphæresis of android /⁠ˈænˌdɹɔɪ̯d⁠/, which is analysable as andro- “human; man” + -oid “resembling, like”. In this case, though, the resemblance to a human is a bit vague…

For “Keeny’s Research on Neo Huroid!”, I would need 160 Plugs from Neo Huroids:

I’m not gonna lie, these bots are a bit chunky for ya girl d00r. My gear is pretty strong, but 20 base INT only gets you so far when you’re Magic Clawing metal robots… In any case, what the hecc is a Huroid anyway? Do they have to wear diapers? Maybe it’s from human + (d)roid…

Speaking of variations on droid, for “Keeny’s Research on D.Roid!!”, I would need the Broken Mechanical Heart of a dismantled D. Roy. So, in Area C-3 of the Alcadno Research Institute’s laboratory, I fit, fought, & fiddled D. Roy until it coughed up its cardioid contraption:

The MPQventures of d00r

Now I’m really ready for MPQ. Fortunately for me, a certain outlaw by the name of Gock (Harlez, VigiI) was similarly ready. And thus the MPQing began…

Stage 2 was the only stage that allowed me to put my Heal to good use in the DPM department. Still, the Cytis^[1] are pretty chunky at 38k HP a piece, so they still took quite a few hits to kill… 😅

In stage 3, my Heal was used more for patching up my party members. The Homunculī don’t mind my Healing, & they don’t really mind my Magic Clawing all that much either…

But I figured, hey, at least I come in real useful during stage 4, thanks to my level 20 Teleport…?

Naturally, we’re only starved for mages when I’m not already playing one myself. I found myself rarely being the only Teleport user, & frequently not even the only priest in the party…

But… okay… in my defence, I used my knowledge of MPQ acquired over some one or two thousand runs to help out a number of folx who arrived at Magatia fresh-faced & new to MPQ. Like Waifuuu, who was just barely too slow to realise that immediately going from stage 4 to stage 5 wasn’t the right idea:

Ah, but even I learned something myself. You see, back in my day 👵🏽, we used to use party chat to communicate our combos in stage 6. Then, someone had the bright idea of using Google Sheets™ for this purpose, using its spreadsheet capabilities to line up everyone’s combos nice & neat, combined with its real-time collaboration features allowing each person to concurrently write their own combos & read those of their party members. This was a well-received improvement, although admittedly Google Sheets™ is kind of a sluggish buggy nightmare sometimes, so I was never averse to switching back to party chat.

Most recently, however, there’s now a web app dedicated to exactly this use-case!: https://maplestory-helper.web.app/. Whoa!! All you have to do is enter an arbitrary username that isn’t already taken, plus the ID of a room (which is just an arbitrary natural number that isn’t already taken), & you’re set up with a GUI designed specifically for solving MPQ’s 6^th stage!:

The combos are all neatly lined up, with your own combo at the bottom, a row of one-based indices (1, 2, 3, ⋯) just above that, & then everyone else’s combos above that. User input is specified via two rows of four buttons each, where each column corresponds to one of the four platforms in an in-game row (conventionally ⟨1, 2, 3, 4⟩ from left to right). The top row of buttons is for which platform is the correct one. If you find the correct platform, you can input it there. The second row is then its complement: if you test a platform and learn that it’s incorrect, then you can enter it in that bottom row of buttons.

By remembering not only the input that you give, but also that info sent by others in the room, the buttons can be colour-coded: platforms that are known to be incorrect are red along the top row of buttons, & grey along the bottom row; platforms known to be correct are bright glowing green in the top row, & the other possibly-correct ones are some dimmer shade of green.

It’s possible to assert that a platform “known” to be incorrect is actually correct. Obviously, this can only happen if someone(s) in your party makes a mistake, & it does require confirmation (“are you sure?”), but this program will accept inconsistency if necessary. I think I only had to make use of this feature once, though.

O! I am Fortune’s foole

class MpqStg6Row:
    def __init__(self, cols=4):
        self._platforms = [set(range(1, cols + 1)) for _ in range(cols)]

    def test(self, plat, pc, result):
        """
        Have the PC ``pc`` (given as an index 1, 2, 3, or 4, where you are 1)
        test the platform ``plat`` (given as an index 1, 2, 3, or 4, from left
        to right), with ``result`` being ``True`` if the platform was correct,
        and ``False`` otherwise.

        :returns: ``None``.
        """
        if result:
            for i, plat_set in enumerate(self._platforms, start=1):
                if i == plat:
                    plat_set.clear()
                    plat_set.add(pc)
                else:
                    self._remove_pc_from_set(i, pc)
        else:
            self._remove_pc_from_set(plat, pc)

    def _remove_pc_from_set(self, plat, pc):
        plat_set = self._platforms[plat - 1]

        plat_set_len_old = len(plat_set)
        plat_set.discard(pc)
        # Did removing this PC from ``plat_set`` simultaneously determine the
        # PC for whom this platform is correct?
        if len(plat_set) == 1 and plat_set_len_old > 1:
            pc1 = tuple(plat_set)[0]
            self.test(plat, pc1, True)

    def p(self, pc=1):
        """
        :returns: A tuple of the probabilities of each platform being correct
                  for the PC ``pc``.
        """
        scores = tuple(
            (1.0 if pc in plat_set else 0.0) / len(plat_set)
            for plat_set in self._platforms
        )
        score_sum = sum(scores)
        return tuple(s / score_sum for s in scores)

row = MpqStg6Row()
print(row.p())
print("~~~")

row.test(1, 2, False)
for k in range(1, 4 + 1):
    print(row.p(k))

(0.25, 0.25, 0.25, 0.25)
~~~
(0.3076923076923077, 0.23076923076923078, 0.23076923076923078, 0.23076923076923078)
(0.0, 0.3333333333333333, 0.3333333333333333, 0.3333333333333333)
(0.3076923076923077, 0.23076923076923078, 0.23076923076923078, 0.23076923076923078)
(0.3076923076923077, 0.23076923076923078, 0.23076923076923078, 0.23076923076923078)

All theſe vvoes ſhall ſerue for ſuueet diſcourſes in our time to come

And I kept MPQing. In search of the Eye of Horus, & to know those with whom I furthered my quest. Like aDeepAdept, who was wearing some rather peculiar-looking garb when I first met him:

Or iDorii, who was rather a bit upset by my creative spelling of service as ⟨cervice⟩…:

Sheesh, tough crowd.

I also partook in some rather daring attempts on (Angy) Fanky’s life. Like this one, alongside Gock, chief dit krowned, & woodsman IntoTheWoods (Thalnos, ZeroSTR, QuestOnly, Garnier, manni)!:

There was also this run with F/P mage Dovas & ranger Yassoppu, where we lost our main damage-dealer who was going to clean up Angy Fanky real quick with a targeted Mesosplosion, but we decided to attempt a normal (non-Angy) Fanky run anyway:

I popped one of them fancy punkin pies from the Thxgiving event so that I could actually help Dovas kill the damned beast, & we actually ended up making it! Wow! I’m not gonna lie; with just my usual Guavas (+30 MATK), I was hitting mostly or entirely “1” damage lines on Fanky, or at least on Angy Fanky…

Eventually, I did get the marbles & the Rocks of Wisdom necessary to complete a Horus’ Eye. In particular, I got lucky with one of my Rocks of Wisdom originating directly from an MPQ end-reward. So, it was time to do some chores for Yulete, starting with “Making the Reagent”:

And, after extracting the polychromatic fluids from the diaper robots & doing the alchemy myself, I did “Organizing the Lab Report”. This quest is not pictured here, because it consisted in little more than me painstakingly running my mouse pointer across a big dumb room.

Last but not least was “Eliminating Obstacles”, which I tried completing within the five-minute time limit by chowing down on another punkin pie slice…:

…But with no success. Well, in reality, it was some success, since running out of time doesn’t actually reset the quest, so I just re-entered to finish up my 100 kills.

And with that, I had my Horus’ Eye at last! And a pretty good one, at that!!:

It might seem silly at first blush to care about the pendant of a character destined for HS–door muledom, but think about it this way: if I’m staying at level 81 (the minimum level required to have enough SP for max-level HS) so that I can stand in the map with the characters actually doing the fighting, chances are I’m going to have some monster attacks lobbed my way in many situations. The Horus’ Eye is absolutely ✨perfect✨ for this use-case: it gives huge bumps to my DEF & MAX(H/M)P, and of course, a nifty bonus to my total avoidability on top of all that base LUK that I already have!

d00r on the h0me stretch

That being said, now that I’ve got my Horus’ Eye, I’ve little reason to continue MPQing. I don’t need that much more EXP, & quite frankly, it’s looking like our lovely MPQers aren’t having many Teleport issues of late.

So I went back. Perhaps you can guess where to I returned…

But of course, to the place where all things began: the Forest of Dead Trees. The trees might be dead now, a pitiful share of their timber harvested, & a pitiful share of that brought down the mountain, the corner of the volcano’s lip still mined & quarried for nothing, even as their miners are already dead, a train that derailed & never stopped rolling, or a snow-covered self-visiting abandonment, yet this is where we came from. Spewed forth, un-sniffed, born emetic & bushy-tailed, children of the mountains, where the 팥죽 carried us through the winters till there was no more carrying to do. The Winter is interminable now, the populations of the mountainsides the last living fossils of a time when النَّطْح knew the warmth of its own star.

It’s often quiet when there’s noöne left to make any sound. It was here where I shook hands with Safety (Danger) before we investigated the mysterious behaviour of Riche:

You see, in an attempt to clear out the Riches of all channels of a given FoDT map as quickly as possible, Safety had a plan: drop the Zombie’s Lost Tooth onto the headstone, immediately change to the next channel (ccnext) to drop another tooth on that headstone, & so on, until looping back around to the initial channel. This avoids having to wait for the headstone reactor^[4] to accept (read: destroy) the tooth & spawn the chest. Then the channels can be gone through a second time (the “second pass”), in each channel breaking the chest & then slaying Riche.

But Safety was hitting a rather annoying snag: although this method did allow him to remove Riche’s invincibility buff, it didn’t prevent Riche from warping him back to Ice Valley II about as soon as he got close enough to Riche to Heal it to death. Safety largely chalked up this difficulty to his being underleveled (in the low level ≈50s), & thus not having the damage output to consistently slay Riche before it could warp him.

I’ve got my own fair share of experience with Riche, though; not only has d00r had to kill a number of Riches to make room for her Coolie Zombie grinding, but I’ve also farmed for the Riche card set on a handful of my other characters. Like any anti-botting monster, Riche is naturally invincible & is naturally capable of teleporting PCs out of the map; unlike some other anti-botting monsters, however, these two superpowers are disabled individually.

I was pretty sure that I remembered two separate blue-text messages showing up in my chatlog when taking care of a Riche, & indeed, you can see in the image above that Riche has a padlock icon above its head, causing it to be unable to warp me — in spite of it still being invincible (as evidenced by the “MISS”). When the headstone accepts the tooth, Riche’s warping skill is disabled; when the chest is broken, Riche’s invincibility is removed.

The fact that the headstone reactor was accepting the tooth & spawning the chest in place of itself, but simultaneously failing to disable Riche’s warp skill, is almost certainly a bug. I suspect that the server is, for whatever reason, unable to bestow status effects ((de)buffs & the like) upon monsters that don’t have a monstercontroller^[5]. And of course, with Safety already having changed channel by the time that the headstone activates, the unlikely presence of an interloper is the only possible way for Riche to have a monstercontroller at that point in time.

In any case, with the help of a few of Danger’s buffs (thx for the MW30!), I FoDTIII’d my way from level 79 to 80, & 80 to the big 81:

LXXXI

Sooo… that’s it? I’m level 81 now?

It’s a little strange, but strictly speaking, d00r isn’t the only character that I’ve brought to “completion” in this here diary: I also have my PQ mules kinds, sorts, sets, & potpan. Still, PQ mules are not “mules” in the usual sense: they are — typically, at least — created to be played, not to be muled per se (although they can be, just like almost any character can). Playing the same PQ over & over is MapleStory gameplay if I’ve ever seen it.

As a HS–door mule, though, d00r has no particular game content to do. That’s a weird feeling to have for a character that I’ve invested not just abstract “time” into, but actually interesting, fun, & memorable gameplay into. From her humble beginnings KPQing up to level 29, to the many questing adventures that she had with GiIf & Skjal (particularly see the “@ ur cervice” section of pt. cvi of this diary, as well as the “p0rtal” section of pt. cvii & the “Hysteria in the area of Masteria” section of pt. cviii), to her grindventures across the undead Maple world (see the “Grindin’ With d00r” section of pt. cix, & particularly the “A world t00r with d00r” section of pt. cx), to her 3^rd-job journey (see the “The ascensi0n of d00r” section of pt. cxi), to MPQ & beyond, d00r has had quite the adventure.

And I am happy — proud, even — to know that I documented that journey in this here diary. I’ll never be able to forget the times that I had with GiIf & Skjal, and that kinda means a lot to me.

Despite being a HS–door mule — & the umpteenth such mule in Maple history, at that — I couldn’t possibly remake d00r, even in principle. She is her own character, & one with more characterisation than probably a lot of “main characters” out there in MapleLegends — at least, the way I see it.

Maybe that doesn’t make any sense, but I’ll leave that for the reader to decide. Maybe I’m just “bad” at making mules, because I see the adventurer in every character that I get past character creation (see the “Simply inexplicable” section of pt. c). Or maybe my lifelong refusal to engage in leeching has bestowed upon me a very particular relationship with “EXP(erience)”.

Whatever it is, I hope that you enjoyed reading about d00r’s adventures even half as much as I enjoyed having them & writing about it. And still, d00r’s adventures don’t quite end at the “level 81” mark. We’ll be seeing a little m00r d00r in the future. 🧡

Footnotes for “Behind closed d00rs”

1. [↑] See the “Lady Palmation” section of pt. ciii for more on this name.

2. [↑] An efficient way of generating the four combos of MPQ’s 6^th stage is by simply applying a Fisher–Yates shuffle to the array [1, 2, 3, 4] ten times. The result of the first shuffle determines the first element of the four combos, the result of the second shuffle determines the second element of the four combos, & so on. This produces the four combos unbiasedly, and with time & space complexities both linear in the size of the output. Care must be taken in the implementation of Fisher–Yates, as bias (and/or poor performance) can be introduced.

3. [↑] This never results in division by zero, because the sets can logically never be empty.

4. [↑] Reactor is a very specific technical term in MapleStory that refers to a certain kind of entity distinct from items, footholds/ropes/ladders, skills/summons, portals, PCs, NPCs, monsters, etc. Reactors are typified by the breakable wooden-plank boxes (RID 2000) which are the first reactors that the player encounters (& almost certainly the first reactors added to the game), as they spawn regularly on Maple Island (& elsewhere).

   

   

   Reactors can be scripted similarly to portal scripts, although they generally respond to a different & broader set of events. Reactors are fairly diverse, & only some of them are “breakable”. The ontology is used for a variety of types of entity that need to exist spatially and/or visually within a map, whilst reacting to observable events that occur — usually within that map, & usually without altering the reactor’s position.

5. [↑] Monstercontroller is another MapleStory term of art. Monstercontroller is frequently confused with aggro, but the two are distinct, as evidenced by the fact that monsters frequently have monstercontrollers without aggroing anyone. The fact that monstercontroller is a fairly aggressively single-client-sided mechanic is the source of lots of goofy nonsense in MapleStory.

Zaqqūm

It’s time for the really exciting shit, y’all. There’s so much good Zakkin’ with so many good ppl up in this here section, you’ll have to read it twice to even believe it.

Is Zakum the tree, or the statue?

But first, let’s get a few things straight. We all know that killing Zakum consists in rock-smooshing: you throw down the Eye of Fire, a big crumbly rock statue comes out, & you hit it until it pops like a piñata.

But what’s with that big tree in the background? It’s naturally obscured by the stone statue, but still definitely visible during the fight. Contrary to somewhat popular belief, however, it does not spawn when the statue does. This is easily proven by taking a screenshot upon entering Zakum’s Altar:

You can see, based on the fact that I’ve not cropped out the expedition timer at the top, that the stone statue has yet to be spawned, & no Eyes of Fire have been dropped. Still, we can clearly see the gnarly hollowed-out leafless tree.

The tree is, bizarrely, apparently chained to the walls of the cave by a mess of bulky metal chains. Moreover, at the base of the tree, we see a variety of stone fixtures — not the least of which being a small staircase leading up to what is presumably “Zakum’s Altar” itself — each of which has a variety of what appear to be decorative patterns, glyphs of a probably ad hoc or nonexistent writing system, and/or faces carved into them.

The tree

The in-game name Zakum is from Arabic زَقُّوم (sans diacritics: زقوم; ⟨Zaqqūm⟩; MSA /⁠zaqˈquːm⁠/). Zaqqūm is mentioned in the Qurʾān, e.g. in 37:62〜68^[1] (note that the English translation chosen here^[1] translates Zaqqūm as “Bitterness” rather than leaving it as-is):

Zaqqūm is also mentioned in 44:43〜46, where it’s made clear that the fruits themselves are a scalding punishment:

زَقُّوم, and variants/derivatives thereof, have been used for various real-life plant species.

The statue

In the “Fighting Zakum for the first time in my Maple career!!!” section of pt. lxxv of this diary, I did a brief overview of some of the visual motifs that the in-game Zakum statue makes use of, using coppersan’s MapleStory Mythology Monday — Zakum as one major reference. I’ll be reiterating some of the things that I wrote in that section, but I still think it’s worth really doing one of the most — if not the most — iconic bosses in MapleStory justice:

Monumento a la patria

coppersan compares the Zakum statue to this monument, although he refers to it periphrastically as “the Mérida monument” due to its location in Mérida, the capital of the Mexican state of Yucatán. The name translates to “Monument to the homeland”, & was constructed by hand over the course of 11 years, by Rómulo Rozo, around the mid-20^th century. This makes it pretty recent, & thus its style is best described as “neo-Mayan”; Rozo intended to evoke the history of Mexico with the piece.

coppersan also describes this monument as “an Aztec monument from the Mayas”, which is, regrettably, multiple different kinds of incorrect. Because the monument is explicitly pan-Mexican, the Aztecs are definitely relevant, but not exclusively so. The location of this monument (the Yucatán Peninsula) was home to the Mayans, but never to the Aztecs; moreover, the style is neo-Mayan, not neo-Aztec. Even furthermore, this monument is not “from the Mayas” — it’s from Rómulo Rozo. Who, by the way, was of Colombian origin, although he did live most of his life in Mexico.

In any case, although most pieces of Mayan or neo-Mayan stonework would probably have been appropriate here, the Monumento a la patria is a good reference point due to its status as relatively recent, revivalist neo-Mayan architecture. As we’ll see, the Zakum statue isn’t based on any particular ancient Mexican artefact nor anything like that, so this is an appropriate vibe.

Classically (i.e. Mayan, rather than neo-Mayan), stone monuments of the general kind exemplified by the Zakum statue’s body are known as Maya stelæ.

Chāhk & Tlā́loc

Chāhk /⁠t͡ʃäːhk⁠/ & Tlā́loc /⁠ˈt͡ɬäː.lok⁠/ are the Mayan & Aztec, respectively, gods of rain & storms. Tlā́loc was probably an adaptation of Chāhk, which itself might be adapted from an Olmec deity.

coppersan suggests that the various stone faces associated with the Zakum statue — the big face of the statue with goggle eyes and big round teeth, its square-toothed & square-nosed counterpart at the very bottom of the altar (part of the map itself, not of the statue), & the Zakum Helmet — are modelled on Tlā́loc (or Chāhk) masks. This strikes me as somewhat likely, given that this certainly wouldn’t be the first time that Chāhk/Tlā́loc and their masks appeared in popular culture, especially considering the continuing cultural importance of Tlā́loc in Mexico.

The goggle eyes, large teeth/tusks/fangs, & prominent nose are common features of Tlā́loc masks which are also used for the Zakum statue/altar & for most of Zakum’s summoned minions.

Also worth mentioning is Cocijo, the equivalent Zapotec deity.

The quetzalahpānehcāyōtl of Quetzalcōātl (& of Āxāyacatl, or possibly of Motēuczōmah II)

We also have the Zakum statue’s distinctive semicircular headdress…:

…Which can also be seen on the Zakum Helmet.

The Aztec deity Quetzalcōātl /⁠ke̞.t͡säɬˈkoː.äːt͡ɬ⁠/ is reasonably consistently characterised as having a similar semicircular headdress, excepting some depictions of his feathered-serpent form. Quetzalcōātl appears in many Aztec codices, like e.g. this depiction in the Codex Borgia:

The native Classical Nāhuatl name for such a featherwork headdress is quetzalahpānehcāyōtl /⁠kɛt͜säläʔpäːnɛʔˈkäːjoːt͜ɬ⁠/^[2] (modern quetzalāpanecayōtl /⁠kɛt͜säläːpänɛˈkäjoːt͜ɬ⁠/). Here, the quetzalahpānehcāyōtl’s semicircular basis is in dark grey, with structural detailing in a combination of red, yellow, & a bit of white.

“Motēuczōmah’s headdress” is a real-life quetzalahpānehcāyōtl traditionally associated with tlahtoāni (Classical /⁠t͜ɬäʔtoˈäːni⁠/; “ruler, king, sovereign, governor”) Motēuczōmah II (/⁠mo.te̞ːkʷˈsoː.mäʔ⁠/). The exact provenance of this quetzalahpānehcāyōtl is uncertain, & it very well may not’ve belonged to Motēuczōmah at all; it was perhaps Āxāyacatl’s (/⁠äː.ʃäːˈjä.kät͜ɬ⁠/). In any case, here’s a modern reproduction^[3] of the headdress:

Quetzalahpānehcāyōtl are associated with Quetzalcōātl, and we can see the signature semicircular structure & ornate radiations in this example. The zigzag pattern on the Zakum statue’s headdress is reminiscent of the repeating patterns on the innermost part of Motēuczōmah’s headdress, & its radiating offshoots are reminiscent of feathers. Most of Zakum’s summoned lackeys share this general feature as well.

𓂀 (Horus’s eye; left Wadjet eye)

The Eye of Horus^[7] (also “left Wadjet^[8] eye”) is a very common symbol in ancient Egyptian religion. It represents Horus’s (usually left) eye after it was restored, following Horus’s arch rival Set^[9] ripping it out or otherwise destroying it. To be fair, Horus did steal Set’s testicles…

In any case, Horus’s restored eye is frequently associated with protection & healing. It’s also associated with the Moon (the Sun being his other eye and/or the Eye of Ra^[10]); the loss of the eye is the new moon, & the lunar phases constitute its restoration. It’s also associated with royalty & kingship, because the conflict between Horus & Set — in which Horus generally prevails — is over the throne.

None of this has to do with Zakum, but the statue’s summoning animation uses the Eye of Horus anyway:

Amun

Zakum’s Altar has just one NPC in it: Amon. This NPC’s name is a typical spelling of the name of the ancient Egyptian deity Amun (𓇋𓏠𓐰𓈖𓀭; ⟨ı͗mn(w)⟩; Anglicised Egyptological /⁠ɪ.mɛn(u)⁠/; reconstructed */⁠jäˈmäː.nuw⁠/; usual English pronunciation /⁠ˈɑ.mən⁠/). Amun was the chief god of the Egyptian Empire, & as a result, would later come to be identified with Zeús & Jūpiter. He was also probably originally a wind god, & later ended up identified with Ra^[10], the Sun god.

(Lemegeton) Clāvicula Salomōnis

Lemegeton Clāvicula Salomōnis, known in English as The Lesser Key of Solomon or simply the Lemegeton, is a grimoire compiled in the mid-17^th century & focusing on the art of Goetia. Goetia lies somewhere amidst the crossroads of what we would now call “magick” (note that the ⟨k⟩ is significant) or “sorcery”, and what is now commonly characterised as Western esotericism & Qabbālā (קַבָּלָה; usually spelt ⟨Kabbalah⟩ in English). Classically, Goetia was considered “base”, “low”, “deceitful”, or “evil”, in contrast to some other kinds of magic & rituals. Modern occultists like Aleister Crowley have popularised highly syncretic variations.

Using an English translation by Samuel Liddell MacGregor Mathers (edited by Aleister Crowley & Hymenæus Beta) as a reference, I found the Magical Triangle of Solomon that can be clearly seen in one of Zakum’s animations. Note that coppersan mistakenly refers to the Magical Triangle of Solomon as “the Seal of Solomon”, which we’ll get to later.

The three sides of the triangle read:

All three of these words are claimed to refer to God, but only the first (the Tetragrammaton יהוה) actually exists anywhere outside the Lemegeton, as far as I can tell.

Starting in the bottom-left corner & proceeding clockwise, the text inside the triangle reads:

Spelling the name of Michael the archangel.

Note that the Magical Triangle of Solomon is paired with the larger Magical Circle of Solomon, the former being three feet in diameter, & the latter being nine.

But Zakum’s version of the Magical Triangle of Solomon has some extra stuff attached to it: one circular seal-looking thing for each corner of the triangle. Indeed, the one at the top is exactly a seal, namely the Seal of Decarabia:

Much of the Lemegeton consists in listing, describing, & illustrating 72 demon spirits, of which Decarabia (sometimes Carabia) is #69. Note that Zakum is not the only part of MapleStory that pulls directly from this list: Andras, Marbas, Amdusias, Valefar (a.k.a. Valefor, which is the name used in GMS), Crocell, & Astaroth are the warrior, magician, archer, thief, pirate, & big-bad-boss, respectively, devils of the “Episode One” (“Devil Hunter”) expansion added in late-pre-BB MapleStory (in GMS: v79, 2009-11-11).

I also found an 1889 English translation of a Clāvicula Salomōnis on the Internet Archive, also due to Mathers. Note that this is not the Lemegeton, but rather, would be the “greater key” of Solomon.

Of particular interest are the pentacles, which are basically talisman or necklace designs, often ornate, used for magickal evocation. The ⟨penta⟩ in pentacle is unrelated to the Greek-derived prefix penta- “five” — it’s actually from OF pendre “[to] hang”, also the source of pendant — & indeed many pentacles have nothing to do with the number five.

The Clāvicula Salomōnis text refers to its “Fig. 47” as “the Fourth Pentacle of Mercury”:

Most of the animation for Zakum’s flaming pillars attack is a nearly verbatim copy of the Fourth Pentacle of Mercury:

Even the Latin-language text around the outside appears to be directly copied from the Fourth Pentacle of Mercury, despite the fact that it’s borderline illegible in the final product. The Latin text is as follows:

Sapientia et virtus in domo ejus, et Scientia omnium rerum manet apud eum in sæculum sæculi.

This version of the Clāvicula Salomōnis also adds an editor’s note about the Hebrew text & the Latin-language text:

In the centre is the Name of God, El. The Hebrew letters inscribed about the dodecagram make the sentence, “IHVH, fix Thou the Volatile, and let there be unto the void restriction.” The versicle is :― “Wisdom and virtue are in his house, and the Knowledge of all things remaineth with him for ever.”

“The versicle” being Mathers’s English translation of the Latin-language versicle.

There’s also the Second Pentacle of Venus:

Zakum’s 1⧸1 animation is difficult to parse, but it appears that the pentacle on the left-hand side is simply the Second Pentacle of Venus:

There’s also a lot of other nonsense going on in Zakum’s animations that looks like it was, or could have been, drawn from the Lemegeton, the Clāvicula Salomōnis, or something related. Take for example Zakum’s lightning–Seal attack, which features two or three pentacles (depending on whether you count the bottom-left one), plus a big fat Mercury symbol that has been merged with a Sun symbol (☉) in the middle:

Granted, the Mercury–Sun symbol is very low-contrast, so it’s difficult to pick out. Also, what’s up with the Vitruvian Man-looking thing in the bottom-left, anyway?

The Seal of Solomon (Ring of Solomon) & the Star of David (✡)

As mentioned previously, coppersan erroneously identifies the Magical Triangle of Solomon with the Seal of Solomon. In reality, the Seal of Solomon is a hexagram (sometimes pentagram) symbol that is the predecessor of the modern Star of David. Although the Seal of Solomon is, like the Magical Triangle, closely associated with occultism/mysticism, the Seal nevertheless occupies a position in mainstream knowledge, in part due to its use in vexillography (& thus the modern Star of David).

We can see something like a Seal of Solomon at the top of Fig. 15 above, in addition to a much lower-contrast yellow one near the bottom of the image. Fig. 12 displays a frame of Zakum’s flaming pillars attack animation, but not pictured is the end of the animation, which prominently displays a Seal of Solomon.

Miscellaneous “foreign” text, & ad hoc glyphs

Zakum’s altar, the statue, & its attack/skill animations feature a wide variety of text & things that are probably supposed to be text. We expect — at least, naïvely — the MapleStory audience to view graphemes outside of Hangeul, the Latin script, Chinese characters, & kana (& more marginally, the Thai script) as being “foreign”.

We already saw some Hebrew text in Fig. 12. It’s worth noting that this is far from being the only Hebrew text associated with Zakum; for example, we can also see some in Fig. 9, plus some Latin-script text reading “BETH”, the second letter of the Hebrew alphabet.

Consider also Zakum’s regen animation; within it, we can observe at least two distinct kinds of “foreign” text:

There’s a lot of Hebrew text here, some more legible than others. But we can also see, in the top-left & top-right pentacle-looking discs, two glyphs (one for each disc) that might be intended to represent one (or more) grapheme(s) each. As far as I can tell, these are what I’m going to refer to as ad hoc glyphs, meaning that they aren’t sourced from any writing system that actually exists (or existed) in real life. These two ad hoc glyphs in particular vaguely remind me of the Bengali–Assamese script, but don’t actually match up with any graphemes of that script. Maybe some comparison to Tibetian is also warranted… but probably not.

Ad hoc glyphs show up in many of the visual aspects of Zakum, including e.g. the stone inscriptions on its arms.

There’s also the stone tablet that the statue is holding, which has two identical inscriptions written vertically along its sides:

Unfortunately, the inscriptions are a bit obscured by Zakum’s hands, & are so low-resolution as to be nearly inscrutable, but the fact that the inscription is duplicated does help.

These glyphs look a tiny bit like Chinese characters, but I don’t really recognise any of them. My best guess is that the second-to-bottom one looks kinda like ⟨百⟩, except that the bottom two horizontal strokes don’t connect to anything. And that’s about all I got. It’s possible that these glyphs are designed to evoke — but not to represent — older forms of Chinese writing, e.g. small seal script.

However, the glyphs here are too geometrically regular — being almost entirely composed of line segments that are either vertical or horizontal — for these interpretations to really make sense. This particular geometric regularity suggests a comparison with cuneiform, but… this doesn’t particularly look like cuneiform either. My tentative conclusion, then, is to just assume that these are ad hoc.

Summary

We might summarise these elements as follows:

• Postclassic Mesoamerica:
  • Maya stelæ,
  • Chāhk & Tlā́loc masks,
  • & quetzalahpānehcāyōtl.
• Ancient Egypt:
  • Eye of Horus,
  • & Amun.
• Western/Jewish mysticism/esotericism/occultism/astrology:
  • Lemegeton Clāvicula Salomōnis,
  • Clāvicula Salomōnis,
  • Seal of Solomon,
  • astrological symbols,
  • & miscellaneous other pentacles/seals.
• Glyphomancy:
  • Hebrew text/alphabet,
  • & multifarious ad hoc glyphs.

Monster book info

MapleSEA provides context for Zakum’s existence within its Monster Book entry:

• Lv. : Unknown
• Form : Devil

A dormant spirit of the Zakum tree which previously was without any intellect. The tree started to show signs of life while being fueled by the passion-filled energy of the residents of the lava region. The lava region has been known to have survived by only the passions of its residents in the midst of all the darkness, burning flames and heat surrounding the region. Apparently the dark and shady emotions of the humans living in the area have somehow been absorbed by Zakum over time. Zakum used to be a small, neglected tree, but has now grown powerful thanks to the energy of the human residents of the lava region, who have harbored a great force that even exceeds that of humans…

The wording here notably implies pretty strongly that “Zakum” is the tree.

The basic structure here of “otherwise ostensibly inanimate being awakened/empowered/angered by the energies of the people around it” is something of a generic mytheme that MapleStory is fond of. See, for example, the story of Stumpy, another evil tree with a similar narrative.

The name itself

Say it with me

As mentioned in the “The tree” section above, Zakum is from Arabic زَقُّوم ⟨Zaqqūm⟩, which is pronounced /⁠zaqˈquːm⁠/ in MSA^[4]. In English, Zakum is probably(?) /⁠ˈzæ.k(ə)m⁠/, /⁠ˈzæ.kum⁠/, /⁠zæˈkum⁠/, /⁠zəˈkum⁠/, or maybe /⁠ˈzɑ.kum⁠/ or /⁠ˈzɑ.k(ə)m⁠/ (not to even mention the face vowel…), with different speakers choosing pronunciations basically arbitrarily. Just to see exactly how the pronunciation has to be mangled to get the Zakky-wakky that we know & love, let’s really quickly look at the MSA pronunciation^[4]:

/⁠z-/

This part’s the same as in English. We’re off to a good start.

/⁠-a-/

MSA /⁠a⁠/ is anywhere between [æ] & [ɑ]. In English, those are the vowels of tap & bra, respectively.

/⁠-qˈq-/

MSA has /⁠k⁠/, like the beginning of English cat. But it also has a distinct(!) phoneme /⁠q⁠/, of which English has no real analogue. The good news is that /⁠q⁠/ is basically just the same as /⁠k⁠/, but further back in the mouth: with /⁠k⁠/, the back of the tongue is on the soft palate, but with /⁠q⁠/, it’s cosy with the uvula.

/⁠ˈ⁠/ goes immediately before the primary stress of the word, so this is zaqQŪM, not *ZAQqūm.

The reason why there are two /⁠q⁠/s instead of just one is that this consonant is geminate. English doesn’t really have gemination, but the basic idea is simple: it’s a doubling of a consonant. For some consonants, like /⁠l⁠/ for example, you can just hold it for twice as long. But /⁠q⁠/ is a plosive, so what usually happens is that its release is delayed, so that the “hold” is twice as long. The best example of gemination in English that I can give is bookkeeping.

/⁠-uː-/

This is basically the vowel in English goose. However, in Arabic, vowel length is contrastive, so /⁠u⁠/ & /⁠uː⁠/ are two distinct vowels that differ only in how long they last, the latter being about twice as long as the former. Not *Zaqqum, but rather, Zaqqūm (or Zaqquum, if you prefer).

/⁠-m/

This part’s the same as in English. What a great ending. Love that for us.

The original KMS name for this boss is 자쿰 ⟨Jakum⟩ /⁠t͜ɕɐ.kʰum⁠/. Korean has no /⁠z⁠/, just /⁠s⁠/ & /⁠s͈⁠/, so /⁠t͜ɕ⁠/ (somewhat similar to the ⟨ch⟩ in English each /⁠it͡ʃ⁠/) is used to approximate it.

√z–q–m

In Semitic languages like e.g. Arabic, Hebrew, Amharic, etc., verbs & most nouns are morphologically composed of roots. A root is usually a sequence of three consonants, & is thus called a triliteral root, although quadriliteral roots do exist, & other Afroasiatic languages have biliteral roots as well. To form a word, the root is then transfixed with vowels & sometimes other consonants, in an appropriate way according to rules idiomatic to the language in question.

For example, the root √r–ḥ–m (Arabic √ر–ح–م) means something like “caring, mercy, sympathy”. Thus we have raḥima رَحِمَ “[to] have mercy [upon], [to] have compassion”, raḥīm رَحِيم “merciful; compassionate; beneficient”, marḥūm مَرْحُوم “under mercy [of God]; deceased”, etc.

As it turns out, Zaqqūm is no exception. Hans Wehr’s A Dictionary of Modern Written Arabic (1961)^[5] arranges its entries in root order. This dictionary includes √z–q–m (√ز–ق–م), with the headword زَقَمَ ⟨zaqama⟩ /⁠ˈza.qa.ma⁠/, & defines it like so:

to swallow, gulp, gobble [food]; IV: to make [someone] swallow or gulp down [something]; VIII: ازدَقَمَ āzdaqama = I

And of course, it also includes a sub-entry for Zaqqūm. Here, the Roman numerals indicate verb forms; Arabic has a fairly complex system of verb conjugation.

Last Mission

In MapleStory, each map in the game has a name, but it also has a street name, indicating a larger region (a “street”) that the map is one part of. In many cases, the street name is the generic Hidden Street (the namesake of the Hidden Street website), & in most other cases, the name has a geographical — or at least spatial — meaning.

Zakum’s map — viz. Zakum’s Altar — is, however, an exception. Its street name is Last Mission, & there are no other maps on this street; both Entrance to Zakum’s Altar & The Door to Zakum are on the El Nath street. Why the inconsistency?

The “missions” here are Adobis’s missions. Both stages 1 & 2 of the Zakum Jump Quest (ZJQ) are on the Adobis's Mission I street. All of the Zakum Maze PQ maps are also on this street; I’m not sure why they’re both “I”… Then we also have the “refinery” mission, which requires 30 Zombie’s Lost Gold Teeth. The fourth & last mission, then, is putting Zakum itself to rest.

Personally, though, I’ve always thought of this street name a little differently. At the time that Zakum was first added into the game, & for a while after that, slaying Zakum really was the last possible mission. If you could kill Zakum — which, let’s face it, you probably couldn’t; it’s very challenging indeed — then you would nearly (spiritually, at least…) be “beating” MapleStory in doing so!

Last Missive

I think that’s roughly all that I have to say about this big dumb thing. I hope that you learned at least something about Zakum, or about something else, or at least had a bit of a chuckle at how much time I spent writing about one (1) big dumb MapleStory boss.

To answer the rhetorical question in the title: Zakum (Zaqqūm) is spiritually & narratively the tree, & this is even confirmed by in-game Monster Book info; but game-mechanically & software-wise, Zakum (& variants thereof) is the name of the stony monsters that make up the boss statue. Software-wise, the tree is just a handful of pixels making up part of the backdrop, & thus has no name. Players also cannot meaningfully interact with the tree.

The reason why the question is merely rhetorical is that any answer is inevitably unsatisfying for those who naïvely seek “the real truth” or similar. If Zakum is narratively the tree, & slaying just the statue is enough to put Zakum to rest, then the only logical conclusion is that — narratively — the statue is no more than an extension of the tree; Zakum would thus be both the tree & the statue. Going any further than these simple (or simplistic) answers that I’ve already suggested plunges the Mapler into classic philosophical problems of reference & mereology (& interactivity in digital art, for that matter…).

If you absolutely insist on “the real truth”, it’s this: Zakum is a mishmash of disparate & frequently unrelated mythological, archæological/anthropological, religious, divinatory, esoteric/occult, & linguistic/orthographic elements combined by a handful of Korean videogame developers in the 2000s into a videogame boss complex of gameplay, visuals, audio, names, & a fragile, anæmic post hoc mythopœia that makes no real attempt to justify nor tie together any of these elements.

It’s also the iconic big dumb boss that we all know & love.

Suboptimal does Zak now‽

And on that note, I’m proud to say that we got to do quite a number of Suboptimal Zaks™!! Whoa!!!

In case you’re not already aware, Suboptimal is the alliance that I lead, its constituent guilds being Oddjobs, Flow, lronman, Victoria, & Ossyrians. Guilds & their alliances are great places to organise group activities, but we have a hard time doing that “as a guild/alliance” rather than just two (or three, if we’re quite lucky…) people using alliance chat to occasionally grind together one time, or to organise something smol that we do every day anyway.

The reason for this is basically simple: there aren’t that many of us (yes, even with five whole guilds), & when there does happen to be more than a few of us, none of us can play together. Sometimes not playing together is a matter of game mode, e.g. Victorians not being able to do game content outside the Victorian Archipelago. Other times, someone wants to do Papu, but others are too low-level or not “strong” enough; LPQ, & others are too high-level; etc.

But we managed to get past this! Sort of! One day, when a suspiciously large number of Zakky people were online, we pushed to make the Suboptimal Zak dream a reality. The landmark ✨Suboptimal Zak № 1✨ starred ironman hero Jonathan (daggerknight, Gumby, Kimberly, Edward) of lronman; marksman Level1Crook (Lvl1Crook, xXCrookXx), bishop MiIf (GiIf, Dakota), & nightlord Snel (LawdHeComin, AppleBasket, JunkyardBat, OrangeFungus) of Flow; STRginner Taima (Tacgnol, Gelelp, Yunchang, Boymoder) & my pugilist tarandus of Oddjobs; & shadower Qubsanity (Qub*) of… Nimbus?:

Okay, so maybe I lied just a little bit. We did generally recruit, like, one non-Suboptimal-member character each time. To be fair, Qub also has numerous Suboptimal characters, & in any case, we are still ✨technically✨ in a meta-alliance with Dream (the alliance that Nimbus is a member of). 😁 The point is that this was still definitely organised by, & composed almost entirely of members of, Suboptimal. So there.

Here we are, collectively 🔥shredding🔥 Zak’s left arms:

Whoa.

And them bodies, too!:

The keen reader may have already noticed that this is a seven-Mapler Zak squad. Does that mean that we had multiple parties? Actually, no. Remember how I mentioned that Jonathan is ironman? Yeah. Well… unfortunately Taima crashed midway through our first run, for technical reasons…:

🪦

But make no mistake: we still prevailed over the dastardly stone abomination known as Zakum. And when we did so, we saw so many helmets rain from the sky…

Woah! Ironman Zak hat!! Granted, before this point, Jon already had access to Rāvaṇa Helmets, but still. Very cool. Also, now we know what his face looks like!

Of course, that wasn’t going to be enough. Not only did this first pair of Suboptimal Zax not drop any of those sweet, precious Advanced Combo Attack skillbooks (which can only be obtained from Zak, by the way), but we simply had more Suboptimal Zakking to do.

We went again with a kinda similar lineup: instead of Level1Crook, we had Lvl1Crook the coarse hair; instead of Snel, we had AppleBasket the bishletish; instead of Qubsanity, we had another shadower, Harlez (Gock); & unfortunately MiIf couldn’t make it this time, so we chilled as a six-Mapler squad (five in party):

But what do we do after both of the Zax are all crumbled up? Why, some post-Zak hoeing, of course! The Door to Zakum is a swell place to socialise:

And o boi, did we do some Zakkin’. Like this one, starring SwordFurbs (Yoshis, Fabiennes, Furbs, SwordFurb) the DEXadin of Oddjobs, & another pally by the name of Paid2Win!:

We were having some troubles with Zak’s soaring summoned stooges, not least because of poor Taima, who was frequently one errant rightward summon-boop away from the grave. When we did have a death on our hands, we were forced to employ a certain very precisely-timed strategy of Resurrection: wait for one Zak body to die, & then in the very short window of time before the next body fully spawns, Resurrect Taima.

Great success!!

Oh, & you know what’s also a great success? ACA10 get!!:

On another occasion, we ran as a squad composed of myself (on tara again), Taima, Lvl1Crook, MiIf, Harlez, & a guy named Omphalos who had a crossbow. I don’t mean to be overly dramatic, but I think that the second run that we did can safely be referred to as The Zakkening™.

“What does that even mean?”, you ask? Well, it all started when Harlez d/ced before even a single arm died. Welp, there goes our only carry.

I mean, we could just give up, but we’re already here, & we couldn’t do another run even if we wanted to. So unless we want Zakum to know us only as cowards too pusillanimous to follow through with a rematch, we’ll have to LOCK IN & show this boss who’s boss. And probably eat some Gelts!!

It wasn’t on track to be the fastest Zak run of all time, but we still thought that we were pretty hot stuff.

Well, at least until MiIf accidentally died to Zak’s body:

Alright, well… now we have no bishop. But we could still finish anyway, right? R—right…?

Oh dear. Our random unexpectedly ran out of cheese, & has now gone to see his eternal cheesemaker in the sky.

It’s now just Lvl1Crook, Taima, & myself. So we’re trioing the shit out of this third bod—

Lvl1Crook focking died???

As you can see, at this point we’re not really having “a problem with Zak’s summons”. We’re just… having Zak’s summons. All of them. All. You can’t even take a breath without inhaling enough of the dust being spewed everywhere by the thousands of stony summons constantly bumping into each other to instantly contract asthma. There are so many of these goddamned rocky little shits that there’s nowhere in the entire map left to stand. Zakum’s Altar is like ≥99% summons by mass and by volume.

But we were already locked in. Remember? Locked in.

Eventually, after 53〜54 minutes of continuous Zak battling, that blasted stone statue finally capitulated, with only Taima & myself left to witness it:

And hey, that marks the first time that I’ve ever been able to loot the meso sacks that Zak drops…^[6]

But even after The Zakkening™, we remained undeterred. We stood strong. Or rather, sat strong. On our mounts.

This time, instead of playing tara, I hopped onto my woodsmaster capreolina. Running again with Harlez, & this time also Jonathan, I did what I do best during the arms stage: whacking arms with my PSB as hard as I can, & letting my birb do half of the work.

Ooohhhh yes. That’s a raw damage range of 1.0k〜13.2k!! With MW20, Echo, & Gelt, of course. 😊

I also took a look at my damage when I swapped to my bow. The raw damage range here is astonishingly tight, with the lower end being about ≈85% of the upper end (CV of ≈0.047)! With a sneaky @dpm 2 squeezed in between weapon-cancels, I clocked around ≈3.4M DPM!:

Other Zax

Not all the Zakkin’ that I did was Suboptimal Zak™, though. For instance, I did some Zakkin’ with course air uwu123 (stardustuwu). Here I am on tara, with uwu123, Lvl1Crook, ✨brand-new✨ hero SIIlL, bishop 2Jade, & bowmaster TaroHamster:

SIIlL, bless her heart (& her name…), really was brand-new; she’d never fought a MapleStory boss before. So it was my job to give her some tips, as the only person in the party who had considerable experience playing One O’ Them Stinkin’ Warriors That Stand There ’N’ Go “Slap Slap” All Damn Day™. I’ve never actually played a hero, but you know. I get the idea.

In another Zakventure with uwu123 & SIIlL, I was the last to join a party of not one, not two, not even three, but four SI-wanters…:

Wow. We are a heroic party indeed. I think my SI is doing twice as much damage as my actual damage…

I also ended up joining a Zak run hosted by misandrist (xRook, sissyboy) & Brokeen (Dazho, Kaoz), where we originally thought that I was needed to provide SE to the party… until another archer ended up joining anyway. Oh well. At least I leveled up for once!!:

Waow! My birb is only one (1) level away from maxing! Does my birb get Echo when it hits the level cap?

Anyway, with my very excellent birb (& perhaps some help from a Gelt Chocolate…), I had no problem whiting arm #4, even with like three other ranged attackers in the party:

😎

Footnotes for “Zaqqūm”

1. [↑] The translation used here is a 2015 translation due to free-minds.org, which can be found at Wikisource. The verse numbering is that of the Hafs reading. The translation style is a “modern literal” one, where the literal meaning is preserved to the extent possible (presumably, at least without totally butchering any idiomata), & a given word (or lemma) in the original language is translated to as few as possible different words (or lemmata) in the target language, whilst also sticking exclusively to (Late) Modern English. Compare, for example, the Modern Literal Version (1987) & the Literal Standard Version (2020) of the Christian Bible (OT & NT).

2. [↑] Somewhat humorously, English-language sources seem a bit confused about whether quetzala(h)pāne(h)cāyōtl is a Nāhuatl word for featherwork headdresses in general, or the proper name for specifically the headdress traditionally associated with Motēuczōmah II. The confusion seems to largely result from Wikipedia being the only accessible English-language source here, but the relevant Wikipedia article is not clearly worded & cites no sources. This confusion is typified by one entry in the Nāhuatl(–English) dictionary provided by wired-humanities.org, which includes *quetzalpanecayotl (note the missing ⟨ā⟩/⟨a⟩ between the ⟨l⟩ & ⟨p⟩) as a headword & gives a correct translation, but cites English Wikipedia as its only attestation & only source!

   To be fair, they also include quetzalapanecayotl (note the restored ⟨a⟩) as a headword, & that entry cites much better sources…

   Spanish-language sources are much less confused. These tend to use the word penacho (see e.g. the relevant Spanish Wikipedia article) as the Spanish-language analogue. UNAM’s Gran Diccionario Náhuatl helpfully only includes quetzalapanecayotl as a headword (no *quetzalpanecayotl).

3. [↑] The original is in Austria, because of course it is 🙄.

4. [↑] Note that MSA has no native speakers, & is acquired only through formal education. Depending on whom you ask, Arabic is either a macrolanguage with varieties of varying levels of mutual intelligibility, or a family of roughly ≈30 distinct languages. As its name implies, MSA is the modern prestige lect, & has no particular geographic associations.

5. [↑] Yes, this is an Arabic dictionary written by a literal Nazi. But even to this day, it’s still pretty much the standard, & it’s very useful in this case.

6. [↑] At some point, MapleLegends changed the looting rules so that meso sacks (but not paper meso stacks, gold coins, nor bronze coins) cannot be looted except by the PC who whited — or the PC who dropped them, if dropped by a PC rather than by a monster.

7. [↑] 𓅃; ⟨ḥr.w⟩; Anglicised Egyptological /⁠hɛɹu⁠/; reconstructed */⁠ˈħäː.ɾuw⁠/; usual English pronunciation /⁠ˈhɔɹ.əs⁠/.

8. [↑] 𓇅𓏏𓆗; ⟨wꜢḏt⟩; Anglicised Egyptological /⁠wɑ.d͡ʒɛt⁠/; reconstructed */⁠ˈwäː.ɟit⁠/.

9. [↑] 𓋴𓏏𓐰𓄡𓀭; ⟨swtẖ⟩; Anglicised Egyptological /⁠su.tɛç⁠/; reconstructed */⁠ˈsuː.tiç⁠/; usual English pronunciation /⁠sɛt⁠/.

10. [↑] 𓂋𓐰𓂝𓇳; ⟨rꜥ(w)⟩; Anglicised Egyptological /⁠ɹɑː(.u)⁠/; reconstructed */⁠ˈɾiː.ʕuw⁠/.

On the way, way up

Hellow. And welcome. To another episode of Les petites aventures d’ela et noto^®. I did some OPQing, & also… some OPQing. Like these OPQs that I did with renowned slinger Gock (Harlez):

🦄

It seems that my little red mousie notomys is a bit shy these days. Any time that he opens his mouth, only blank chatbubbles come out!

In any case, OPQ is full of other legendary phenomena aside from whispering mice.

Hmm. Well, I definitely see lots of meso coins fly out of those boxes in the bonus stage. These days, I generally don’t mark the platforms in the On the Way Up stage, preferring to instead keep track of the entire combo (up to 4 ⋅ 4 = 16 digits) in chat. After all, chat messages don’t expire, & in the worst case, I can just resend it. But usually my party members drop little meso coins to mark the platforms anyway. Maybe those are the same coins that drop in the bonus stage?

Even with at least two orders of magnitude(!) more mesos used to mark platforms in On the Way Up, we seemed to be getting the same amount of mesos from the bonus stage! What gives‽ Maybe Minerva is saving the extra mesos for next time…

Remember when ela did EPQ? Me neither. Must have been in a fugue state or something. In any event, I got her a pair of Glittering Altaire Earrings, but procrastinated on actually scrolling them. Well, I finally got around to scrolling them, & you’ll never guess what happened…

Wow… deer utterly failscrolls yet another precious item? Stop the presses. 🙄

As further evidence of the ill-fatedness of our dear ela, another attempt at On the Way Up had me in the fetal position:

Well. At least I also got to OPQ with woodsman IntoTheWoods (Thalnos, ZeroSTR, QuestOnly, Garnier, manni):

🙂

The orange bitch

It me, once again. wooza the darksterity knight. Reddie for some big bad bosses? …Okay, lez go~!

First up is my good ol’ frienemesis Burger Moth. As avid Bossin’ Wif wooza™ readers are already aware, I’ve fought this big bad Japanese robot more times than anyone can count, particularly with my Maple wife Harlez (Gock). Although it’s a great boss for me in many ways, it also has its quirks. Aside from the obvious fact that it doesn’t really give me any opportunity to cleave, its WDEF is also regrettably high. The first body, which has nearly as much HP as the other two bodies combined, has a whopping 2.2k WDEF!! Its second body is a bit better — but still a bit high — at 1.45k, & only the third body has a more modest 1k WDEF. But, you know, by the time I’m at the third body, I’m spending like 40% of my time dodging shit anyway…

The only reason I bring this up is to justify my attempts to clock higher DPMs on this tanky first body. Whilst duoing with Harlez, & benefitting from PB buff, Echo, MW20, & SI, I managed to clock 1.59M with a @dpm 6 on the first body:

Nice~!

Then, disaster struck. On our second run, Harlez accidentally unsubscribed from having more than zero HP, leaving only me & the Burger Moth to duke it out.

But I’d been here before. Back in the “rusa is a big girl now.” section of pt. xcix of this diary, we saw rusa solo a whole bunch of big bad bosses, not the least of which being Bergamot! So I pushed on to finish the run, for great glory, & for the MCP2 that dropped! Nice!!:

Oh, & it’s a Christmas-themed MCP2. Wowzers. I just hope it doesn’t have coal inside…

It’s less likely than you think.

But it wouldn’t be a proper episode of Bossin’ Wif wooza™ without me being absolutely ✨captivated✨ by Horntail-y wail-y’s li’l wiggly purple fingie (uwu^[1]), yeah? Let’s start with the part where I died in a funny way:

Dying? In my preheads? It’s more likely than you think.

But seriously, you can see that I died here after most of the sed’s duration had already elapsed; & yet, nearly up to the point where I actually bit it, I still thought that I was totally fine! Should my bishop have Healed me like once, maybe, over the course of those 7〜8 full seconds? Perhaps. It definitely would have helped, although to be fair, I technically didn’t specifically ask for any. Then again, this was piss-poor luck in more ways than just that. I was particularly unfazed as a result of having Dragon Blood active the whole time, but it just didn’t matter when I ended up somehow taking touch damage anyway. That lightning attack had to not only hit me (already not extremely likely), but also had to break through my Power Stance (10% chance), and the sed had to be jump-sed (only one of the three seds are jump).

Well, the result of this little mishap was that my bishop became much more alert…

It’s frequently a bit awkward for me, because in reality, I actually need very little babysitting the large majority of the time. I get the idea that many sed bishops are not quite in tune with the fact that, >99% of the time, 1⧸1s are the only thing that can really lead to their sed target’s death — or perhaps just not quite in tune with what a 1⧸1 attack (including the distinctive markers on the ground) looks like. In any case, I don’t like to sound like I’m advising my bishop. I could — & in all reality, probably should — say things like “I generally do not need help unless I request it”, “1⧸1s are the only threat to my survival”, & the like. And also be less hesitant to request acute help, for that matter. But a bishop in HT — sed or “reg” alike — frequently already has a lot on their plate, & in many cases, they also have considerable prior experience running HT as a bishop (& possibly as other classes as well). Am I, the random orange bitch, going to curtly advise them on it?

Maybe. Even veteran HT bishops likely have little or no experience with non-shadower sed targets. Perhaps this is the origin of the underappreciation of 1⧸1s.

In any event, the resulting dichotomy is clear: either I have a hard time getting help in the rare cases that I actually do need it, or I have a hard time getting my bishop to stop riding my be-hind like a jennet. I died for the first time ever in preheads, & now my bishop is pelting me with Heals when I’m already at 90% HP & there’s not a single 1⧸1 even remotely in sight! I’m safer than a baby in a swaddling!!

Anyhow, I’m just being goofy now.

I do get some left head+mid head cleave action when I can…:

Bānaṇa

…And it was otherwise pretty fun— Wait. Where’d Horntail go?

Is it really time for… wooza getting some Ravpap action?!? Whoa…!

That’s right. I’ve been continuing my PBventures, & now I’ve got a handful of PB buffs that I don’t know what to do with, plus a brand spankin’ new Timeless Pendant that I gotta level up! So let’s focking do it!!

Naturally, this Papuvanaing was done with course heir Lvl1Crook (Level1Crook, xXCrookXx, Macer) & Stronginner Taima (Gelelp, Yunchang, Tacgnol, Boymoder). Although it did help me use my PB buffs & level up my TLP, it also taught me a lesson about rusa’s pouch-opening fingers.

Unfortunately for everyone involved, I am a slow learner, & at first, I refused to learn my lesson:

Oh, you thought repeated failures & a defame would stop me? Think again.

Griefing MCP1s wasn’t the only thing that I got to do, though. When there was a little server downtime for routine maintenance, Lvl1Crook & I raced to the Phantom Forest as soon as the server came back up. It was time for some 🔥speedhunting🔥 of the freshly-spawned Figboot population:

And I think we did a pretty good job! That was a lot of Bigfeeties!!

This shit again?

I appreciate when my bishop lets me know that they’re under the influence, even if that influence is just melatonin & prune juice. That way, if they fall asleep mid-run, then I can at least rest easy knowing that they shat themself as well.

More seriously, these other runs were great as well. I even managed to establish a point of “I don’t need help unless I request it”. Here’s a fun little jig that I did with HT’s right fingie:

Don’t blink; you might miss it. Kinda like how right head’s 1⧸1 “MISS”ed me there. It’s difficult to tell because the beginning of the Hero’s Will animation is a bit subtle, but I do manage to get my Will off before the 1⧸1 connects. It just so happened to be unnecessary in retrospect…

Just to be clear, I cannot take any responsibility for triggering left arm’s mass sed, as seen in the video clip below. At least half of the time that I’m over on that side, I’m up hitting heads anyway. But you know, the archers get bored with left head getting weapon-cancel-happy… we’ve got two bishops spamming Genesis…

Goodness me!

Not to worry, though. We all made it through! And what’s more, I fricken levelled tf up???

Holy shiet!! Level 185!!! At this point, every level is a damn milestone. 😅

Pibbles

It looks like my TLP is fully baked: level six out of six! Let’s see how it came out…

Ok, that’s actually pretty dang good! It’s definitely better than my first (i.e. previous) attempt, back in the “Pink Bean & jelly sandwich” section of pt. cix. I’ll take it!! ☺️

…For making me feel like I’m still playing MapleStory

And finally, let’s take a brief look at one special PB run in particular. It wasn’t extremely special or anything — we certainly didn’t get two MW30s, nor anything like that… — but it was special to me, for at least two mundane reasons. One of those reasons was simply that I managed to squeeze in a decent @dpm 28 test during the statues phase:

This was a “decent” one because I don’t think I particularly messed anything up, but at the same time, I probably could have done somewhat better, & also the test ran for a little too long. By the time that I got my results back, we were already at the beginning of the main body fight, where I’m pretty sure that my DPM is even lower…

In any case, 1.13M is, I suppose, not absolutely terrible for Cidering the whole way. Not that I’m really there for my DPM, but you know, I do try.

Oh, & the other special thing about this run was that I got to do a lot of babysitting. Is that a weird thing to say? Maybe I just like babysitting. Fight me about it.

This was cores ere Danger’s (Safety) second PB run, & first one in quite a while. I also had a number of other delicate attackers in my party. Out of all five parties in the expedition, we were way up there #1 on the deaths leaderboard, so I was kept very busy during the main body fight with /partyinvites, /partyexpels, urgently-timed HBs, & the like. And make no mistake: that’s no condemnation! Shit happens, & Pink Bean is not fucking around; it will kill you. Danger in particular was at a great disadvantage especially because he has no way of surviving PB’s Big Bang attack, other than dodging it by being too far out of its reach. Of course, this isn’t a problem when PB is, properly, facing rightward, but you know how that goes…

In sincerity, it’s an awkward thing to talk about in a MapleStory implementation that has already long since utterly squandered the lion’s share of its naturally-existing game mechanics in favour of, in this particular case, HP washing. I feel like I, in this diary, constantly come back to this general tangle of topics. If & when I do, it’s only because I care. Is it fair to the relatively unwashed that they’re forced to be dependent on their teammates when their peers get to be independent? Is it fair to people like me who want to play a genuinely supportive role, that they can be stripped of such a role by the mere existence of HP washing, HP Challenges, or what have you? Is it fair to pre-BB MapleStory that we’ve eviscerated it of even the most basic structure of its character-classes & party compositions, leaving behind only feeble echoes — echoes broken regularly by the silence of the anti-mechanics that we’ve replaced it with?

When HP washing in particular is spoken of, the discourse all too often centres on the instinct that washing gets in the way of what we want to do. This is, for what it’s worth, usually itself a valid instinct; in particular, when we play a videogame, we do it for fun, not for anti-fun. But this is too superficial, & so its antithesis is unsubtle: washing is optional; & if it isn’t, then it still is for most people; & if it isn’t, then it’s actually not so difficult to do; & if it is, then we can ameliorate that difficulty; & if we can’t, then we can provide an alternative; & if we can’t, then……

But the more fundamental reason why HP washing desolates the game is hinted at by the fact that genuinely supportive roles simply do not exist throughout most (or in many lost cases, all) of the game, & when they do, they are usually not taken seriously, nor are they regarded as worth talking about at any length (unless you’re me, apparently), nor will you have habitually good luck finding people to fill those roles.^[2] The simple fact is this: although the barest offence of HP washing is its ability to get in our way, its true crime is thieving our way entirely. We don’t complain about the myriad elements & aspects of the game that washing has taken away from us, because we don’t even remember that they’re not there anymore.

Footnotes for “The orange bitch”

1. [↑] What? I’m not making it weird. You’re making it weird.
2. [↑] This is not to imply that supportive roles are the only aspect of the game relevant to HP washing in general — far from it! But this just happens to be the most relevant one here.

Volar

Ready or not, it’s time for some Maple c^O_nTeⁿT with the queen of palms & soles, ya girl tar & us tarandus.

It hurts to say every time, but unfortunately this c_oNt^EN_t will not be questing content; my apologies to the raving mad Questin’ With tara™ fans, who are now zealously foaming at the mouth, preparing their fingers to spam my email inbox with even more impassioned hatemail.

Just kidding. I think. Bossin’ With tara™ ain’t so bad, right…?

But before we get to the scawwie bosses, it’s first time to ✨witness histowwie history✨! Legendary F/P gishlet Tacgnol (Taima, Gelelp, Yunchang, Boymoder) has been holding on to a certain Maple Magician Shield that she managed to land two(!) 30% scrolls for WATK on. Scrolling shields is no casual endeavour, & especially not when the starting point is an already-expensive shield like a Maple Magician Shield. So landing two 30%s is a pretty epico start!

But now, with the money saved up from backbreaking apple farming & other activities, plus the motivation to “use some 60% scrolls on whatever I guess” provided by tier II of the HP Challenges (see the “Grab-bag of optionals” section of pt. cx), it was now time to just spam mfing 60%s for WATK on it!!:

HOLY FOCK@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@

A 20 WATK Maple Magician Shield. Twenty. Heck, mine is only 18!! And you know Tacgnol got some confused and/or angry whispers after showing people this absolute beauty of a gish(let)/STR mage shield. People just don’t appreciate what a work of art that is, which is a real shame.

Unarmed, but not defenceless

You already know that I did lots of Rabananapapalatasing with the homies Taima & Lvl1Crook (Level1Crook, xXCrookXx, Macer). The most essential component of this ritual is, of course, hoeing afterwards at the Origin of Clocktower for extended periods of time:

Often, when we kill a boss, we expect our time, efforts, & resource expenditures to be rewarded at the end by some big prize — usually EXP and/or l00tz. So it’s kinda like sucking on one of those hard lollipops with a soft sweet at the centre: you work your way slowly through the outer part — which, granted, still has a nice flavour — & then when you hit that sweet, sweet centre, you get to gobble that shit up. Yummy yum yummers. In this case, the sweet at the centre is hoeing. It’s literally just standing there, chatting, & wearing funny hats. That’s the delicious chocolatey goodness that we come to MapleStory for.

But Rabananapapalatasing wasn’t the only borssing that we did. We also killed ourselves some Biggyfeets!:

That includes a few Biggyfoots that we killed alongside Harlez (Gock), plus the stray Headless Horsemans that like to show up at The Evil Dead:

I even did some more Kreckselling alongside Harlez & xBowtjuhNL/BuccjuhNL (PriestjuhNL):

xBowtjuhNL had been trying to do Krex more often, but seemed to be having trouble finding willing participants…:

Whew. Good thing the anprims can’t see this, since they don’t have internet connections. 😊

I also did some Rabanana with Harlez & Lvl1Crook…:

…And afterwards, we considered a certain Yāosēng [武(Wǔ)林(lín)妖(Yāo)僧(sēng)]. But Lvl1Crook had yet to do the prequests, so we headed over to the Dutra Suppository for some shiny-monk slapping. Like, for example, 1–2 F:

You know, I really don’t remember the Yāosēng prequests being quite so quick & easy…

In any case, with Lvl1Crook’s prequests done, we took on the great gramps of martial artistry:

During which, I hit gaw dam level 161‽‽:

Back at Rabananana, we had… an incident. Often referred to as “The Japanese Map incident of 2024”, what really happened on this day is known only to me, Lvl1Crook, & Taima. And now you, since you’re reading this. Assuming that you’re neither Taima nor Lvl1Crook (hi).

It all started so innocently, with fun & games. We were discussing the scrolling of Japanese Maps, & Taima noted that she had a few junk ones sitting in her inventory. Lvl1Crook thought that it would be funny to pick one up & wield it for himself. Little did he know, Rāvaṇa was about to start summoning gobs. So, so many gobs.

Indeed, it was shortsighted. With gobs flooding the entire map, Lvl1Crook couldn’t catch his breath long enough to switch equipment. He had inadvertently transformed himself into a swashbuckler with really bad stats.

What happened next will shock you. We killed enough of the gobs to let him switch back to his gun, & then proceeded as normal. 😱

Interlude: “Ball tarandus”

Back to our regularly scheduled programming

I had now done Yāosēng as a duo with Harlez, and also as a trio with Harlez & Lvl1Crook. But what about as a duo with… BuccjuhNL?

Now, you have to remember, in this image, I might be level 161, but BuccjuhNL is 16 levels below me, at 145.

Does it seem ill-advised to you? I wasn’t too worried about volunteering myself in buddy chat under the caveat that my damage might be kinda peepee poopoo, but that was probably because I mistakenly thought that the character asking to “R>” — viz. xBowtjuhNL — would also be the one running. When I realised that he actually meant his buckin’ near BuccjuhNL, I figured we could still do it, but I made sure to negotiate what WATK pots we would be using. We decided on basically just using a bunch of stoppers.

In the way of an insurance policy, Danger (Safety) very kindly agreed to come along for the runs, in spite of the fact that, realistically, he could have just duoed with BuccjuhNL very easily. 🥲 Failed Yāosēng runs are indeed painful, so if necessary, Danger could step in to provide the necessary DPS to polish off Yāosēng at the last moment.

It was looking pretty good, & as a result of our both being buccs, we could use one TL per run, thus getting us each triple-ST^[1] for both runs. Then, you know, our STs ran out.

Well, okay. I have max-level Barrage at least, so I’m not doing zero DPS in my non-ST’d state… r—right?

Hmm, well, BuccjuhNL is level 145, so I suppose he doesn’t have many 4^th-job SP either… But by the time that it became clear that we were struggling to finish within the time limit, we were having trouble putting that insurance policy into effect.

Did we overestimate ourselves? Yeah, a little. But in our defence, by the time that we got booted, Yāosēng had a preciously thin sliver of its HP remaining, so we didn’t overestimate by much at all. Plus, I still feel guilty for only using six stoppers, when I really ought to have used… I dunno, probably eight or so.

I figured if I just pop an apple, then we’re definitely peachy keen; after all, I’ve been doing so many APQs that I don’t even know what to do with all these apples anymore. But popping apples just to clear Yāosēng was looking like a dumb idea from BuccjuhNL’s & Danger’s perspectives. So, okay, it’s better if we just have Danger help out with attacking…

Right, so, even with Danger only attacking part of the time, that was unsurprisingly more than enough to clear…:

All in all, that was actually four (4) Yāosēng runs (because we started not long before reset), & only one failed! ✨

In the “Other Zax” section above, I did some Zaqqing with gores eyre uwu123 & heroine SIIlL. Well, you’ll be perhaps unsurprised to learn that I also Krex’d with them (& Harlez):

Back at the Phantom Forest of Figbeet, I Figbeeted my way into an entire level 162~!:

Whoa~ Big numbor~ ✨✨

And finally, I was invited to chill out with at the local Scar Lion with pally akiwi (Flurri, SecretIy), bowmaster xRook (misandrist, sissyboy), & Harlez:

Noice. 😎

Every day is leg day for Horntail

So, tara is level 162 now…? And I still have yet to take on the biggest, baddest boss of the GMS v62 world: Horntail!!

I’ve only ever HTed on my darksterity knight rusa, so I really have very little idea of how it’s gonna go as a buccaneer. The attacking style is just radically different, & also depends a lot on skill build, including how many SP you have available to spend on attacky attacks. So basically, I’m tryna say that I’m a n00b. In my defence, I still have 16 Gelt Chocolates left to use…

But those are concerns about my outgoing damage; damage from me, to the big lizard. What about the converse? Well, my main concern is those legs. Sure, the legs are the first thing to die, & they generally don’t last that long, but I don’t really trust my ability to dodge its attacks. If I were truly fearing for my life, then I could skip out on legs entirely, in favour of attacking heads until legs go down, & then switching to tail (at least, as long as I’m ST’d, & thus don’t have to get close to touching the deadly thing).

As I’ve just implied, the legs’ attacks can be dodged, specifically by jumping. This can work well for ranged attackers, who aren’t getting anywhere close to touching the legs anyway, & who usually can jump freely enough to jump-dodge on command. As a buccaneer, it’s a little more complicated — notwithstanding the fact that most buccs have more than enough MAXHP to tank all leg attacks anyway. When not in ST, my only options for doing reasonable DPM all involve cuddling up with the leg, which could risk taking touch damage. And when in ST, using Demo (literally just once) is enough to stick me in my animation for a whopping ≈2.4⁢ s or so. Using Demo is, thus, a little bit like intentionally stunning oneself. The problem is that, by the time that the leg even starts its ranged attack animation, there’s a good chance that I’ve already pressed the Demo button. Uh-oh! I can’t jump now! If I’m lucky, my iframes will absorb the hit. But the key phrase here is “if I’m lucky”…

Note that this latter problem still exists for Barrage, in spite of it also having iframes (largely due to its delay^[2]). In Zaqqūm fights, I can (& do) circumvent this limitation via the strategic use of CSB, but in this context, doing so is typically a suicidal strategy.

Data spelunking

So I guess the question, then, is: how much damage can I actually take?

To answer that question, we’re going to have to look in 🗃️The Data🗃️. Yes, that means the WZ data. Your everyday Mob.wz file can be cracked open with an off-the-shelf copy of HaRepacker, or you can use browse-old.maplestory.io if you hate yourself. Either way, you’re just looking at the game data in the same way that your game client does.

To find the data for HT’s legs, we look at Mob.wz/8810008.img, where 8810008 is the MoID of Horntail's Legs. Under Mob.wz/8810008.img/info, we can see lots of goodies related to HT legs generally, including PADamage & MADamage, which are 1 160 & 560, respectively. Here, PA & MA are physical attack & magical attack, respectively. We can confirm that these are the same values listed for WATK & MATK, respectively, on the ML Lib entry.

But we can also see Mob.wz/8810008.img/attack1, alongside three other attacks attack2, attack3, & attack4. Just by looking at the raster-graphics data, we can see that the first two attacks look like mirrors of the latter two; we just don’t often see the right-side leg attacking, for the obvious reason that we don’t generally do “tail first” runs…

Looking at attack3, we see its info node. Here, we see the hit node, which mostly stores raster-graphic data for what it looks like when you’re struck by the attack — but also includes an attach node, which presumably makes it so that the hit animation attaches to your PC’s position, rather than staying in the original position where you were initially struck. We also see a jumpAttack node, which presumably indicates that the attack can be dodged by not being on a FH when it goes off. The knockback node indicates that this attack causes super-knockback if it hits you. The range node indicates the hitbox for this attack, which is specified by the lt & rb subnodes, which stand for left top & right bottom, respectively. Both are (x, y) coördinates. These two points uniquely specify a single axis-aligned rectangle.

There’s also attackAfter, & honestly, I cannot find any info on what this one means. More relevant is, of course, PADamage, which is 950. That’s actually not even that high, although we’ll get back to that later…

attack4 is the one that stuns instead of super-knockbacking, so instead of having knockback, it has disease, which is set to a value of 123. If we check out Skill.wz/MobSkill.img/123, we can dig a bit to see that Skill.wz/MobSkill.img/123/level/1/affected/0 is one of the frames of the stun animation. However, the PADamage value for attack4 is also 950, so both attacks are basically the same for our purposes.

The plot leg thickens

But we also know that HT has the ability to empower itself, which gives it a WATK buff. Bishops have to dispel HT’s self-buffs anyway, so we expect that HT body parts will be unempowered most of the time, but still not all the time.

We can look under Skill.wz/MobSkill.img to find the empower skill. As it turns out, there are confusingly two empower skills that kinda look the same. I don’t claim to know why this is, but their IDs are 100 & 110. The wings do the empowering, so we can peep Mob.wz/8810007.img/info/skill/4/skill and see that it’s 110; none of its other skills have an ID of 100 or 110, so this is it. Mob.wz/8810007.img/info/skill/4/level is 5, so we look at Skill.wz/MobSkill.img/110/level/5/x to find that the magnitude of the buff is 130.

This effectively leaves us with a matrix of four possible WATK values:

Bean counting & number crunching

The way that we get from these WATK values into actual damage numbers is unnecessarily complicated, due to MapleLegends’s incoming physical damage formula being the GMS v62 version. Although I will walk step by step through one example calculation (for empowered leg touch damage), it’s optional reading that will be hidden inside a <details> element. Note that magical attacks are not quite so complicated, but HT legs don’t have any magical attacks.

The bird’s-eye view of the formula looks like this (note that this is not the actual formula):

…Where $𝒰$ is used as a placeholder for a “random variable” with the specified continuous uniform distribution, & $\mathrm{protect}$ is a real number that is usually small & usually positive.

Of course, in order to calculate $\mathrm{protect}$, we need my stats. For the purpose of these theoretical calculations, I’ll use actual numbers copy-&-pasted from in-game, but I’ll pick conservative numbers that involve no buffs — particularly, no WDEF buffs, & no MW. This will get a fairly accurate picture of “the realistically worst case” scenario.

• WDEF: 316.
• STR: 891.
• DEX: 113.
• INT: 37.
• LUK: 37.

Being a level ≥100 pirate, my StandardPDD is Base.wz/StandardPDD.img/5/100^[3] = 309.

Repeating this process for all four quadrants, we get:

Live to punch another leg

At the time of this writing — which includes a few event-only temporary equips, unfortunately — I have just barely (barely) over 12k MAXHP on tara, unbuffed. That puts me pretty firmly in the “can tank” realm for all these attacks, excepting empowered touch, where I’m pretty firmly in the “can’t tank” camp unless I’ve HB or something.

That’s pretty good news, although obviously it would be a treat to be able to tank empowered touch as well. When in ST, I don’t have to worry about touching, so I’m pretty much all good. Outside of ST, I think I have basically two options:

1. Just be careful about getting close enough to legs to Barrage it. I know from experience that Zaqqūm (or Barrage, rather) is fairly generous with its hitboxes, making it not so difficult to Barrage its body without taking touch damage. With HT legs, I might consider trying to get iframes from an incoming attack before walking towards the legs with the Barrage button held down. This is probably(‽) a fairly safe way to get in position; far enough right to hit legs with Barrage, but only just so.
2. Give up when not ST’d, Barraging left head (or whatever) instead.

Considering that the legs aren’t alive for that long anyway, & that my ST generally has a ≈66.6% duty cycle, I feel better about just doing (2.).

Footnotes for “Volar”

1. [↑] See the “A grasp nullary” section of pt. cviii for more on this mechanic.
2. [↑] Not to be confused with its cooldown (it has none) nor its attack period.
3. [↑] The 5 is due to pirate CIDs starting with the decimal digit ⟨5⟩.

速 makes waste

Bosses are cool & all, but I know why you’re really here. You’re here to see everybody’s favourite stabby-stab nightlord McDaggerbitch alces explore more of the Maple World, & stab tf out of even more MFs. Mostly for the cards. The sweet, sweet cards. And hey — we just might learn something along the way.

Well, you’re in luck. Buckle ur seatbelts, tighten ur chinstraps, safety glasses on. Because it’s time to… sleep?

I’ve been sleeping a lot. Or at least, being in bed a lot… So I put together an outfit that I could wear in-game whenever I managed to crawl out of bed for a bit.

In the last episode where we saw alces cardhunting — viz. the “Moose xing” section of pt. cx of this diary — I left off by starting Elpam Gorlab’s quests: “Lost!” & “ReVersal”. To complete “ReVersal”, I would need to farm Temporal Fragments from the Tick-Tocks of the Whirlpool of Time:

I dunno. This “Temporal Fragment” looks an awful lot like a gold Rolex^® or something. If I have to hunt for 60 of them… What’s the plural, even? *Rolices??

In any case, I also got a Springy Worm, as you can see. I suppose I may as well do Clock Tower Monster Wipe-Out, yeah? It is a fantastic quest, after all. So that means hunting down some Tick-Tock’s Eggs…

Whoa! They even have the little twin bells on them. That is adorable.

…Plus some Chronos’s Eggs…

…And then, of course, opening them up to get the Timer’s Eggs that’re tucked inside. I perhaps ate slightly all the Chick Cookies, & now my tum is full of carbs. I think I need a nap.

Dīlūdium sopōris

Requiēvit

[sniff] Okay, I’m ready. Let’s defeat those monsters.

With the Tick-Tocks, Chronoses, Platoon Chronoses, & Master Chronoses scrapped for parts, I just needed that one (1) Timer kill.

Of course, there were no Timers. Whilst I waited for one to spawn, I decided to take on “The Clocktower Headache”, by dismantling some 80 Ticks at Crossroad of Time:

Thankfully, I didn’t have to collect the parts, so I just left the exploded Ticks on the ground.

For “The Clock Workers’ Batteries”, I collected 200 of them from the Robos in my favourite little Lūdi map: Toy Factory ⟨Aparatus Room⟩:

Yes, they did misspell apparatus.

For my efforts in recovering the batteries, Grandpa Clock awarded me with one Purple Shadow…:

C—cool…

And of course, “Mac the Mechanic’s Maintenance Manual” went missing. Again. Here we go…

Okay, I’ve broken like half a dozen of these dumb boxes & haven’t seen anything interesting. What gives?

Oup. There we go…

There it is! The mauve Maintenance Manual! Mac will be very pleased to get this back. At least, until he misplaces it yet again…

Hmm. Still no signs of Timer. Welp, I’ll head down to meet with Ghosthunter Bob again, & this time, I’ll do “The Soul Collector”:

The Spirit Vikings never stood a chance. Now I’ve got their little can of ghost… spaghetti?

I’m kidding. Why would you store spaghetti in a can like that? It’s just pasta sauce.

Unfortunately, now matter how much pasta sauce I farm, I can never seem to catch a glimpse of the elusive Timer. Oh well. I’ll just head back to Masteria now.

O, Masteria!

Whoa! Completing “ReVersal” also let me finally complete “Urban Warrior”! Now I have a cosy chair to… probably fall asleep in.

But completing Elpam Gorlab’s pair of quests was merely a prelude to doing the other Masteria quests, too. I’ve already done a number of them, but with “ReVersal” out of the way, I’m on my way to “Finding Jack”, wherever he might be up in these blasted Crimsonwood Mountains…

alces & the Valley of Heroes

Speaking of the Crimsonwood Mountains, have you ever seen the flora up here? It’s not so much “flora” as it is “floran’t”. I mean, look at this poor Crimson Treestump…:

Drier than a soda cracker in a windtunnel — & just as broken apart, as well.

At least I found something inside:

Maybe there’s a reason why the Crimsonwood Mountains look the way that they do.

The palæogeology of CWM

Looking at the detail of these mountain ridges, which apparently form a mountain range that we know as CWM, I see a handful of features that immediately pop out to me:

• These ridges are very steep/sheer.
• The ridges are also characteristically narrow/thin & tall.
• There’s no visible vegetation, making it easy to see the rock surface directly.
• The ridges seem to peak at similar, vaguely uniform elevations.
• Nevertheless, the rock surface is seriously uneven & jaggèd in most places, including a few smaller spire-like formations that jut out.
• The rock surface looks remarkably uniform.
• The presence of at least two ridges that look to be somewhat parallel, combined with the fact that these maps are referred to as the “Valley of Heroes”, implies that these mountain ridges are implicated in the formation of at least one valley.

Taking these observations collectively, I’d like to speculate about the geomorphology of CWM.

If the valleys are U-shaped, then that would make this ridge an arête. Arêtes are formed when multiple glaciers carve valleys in parallel, leaving sharp ridges that separate the valleys. Given the extreme dryness of the CWM that we see, it might seem unlikely that there was lots of ice here, but it seems plausible given that the formation of arêtes happens over geologic time. A good real-life example of an arête is Crib Goch in Gwynedd (it’s not lost on me that lowercase cwm /⁠kʊm⁠/ is Welsh for “valley”…).

If the valley(s) are V-shaped, then it’s more likely that these formations are fluvial. Either way, huge quantities of flowing water of some sort, over geologic time, are involved in the formation of the CWM structures pictured in Fig. 2.

I have doubts that these formations are the result of specifically oceanic–continental orogenic/uplift processes, given that, in such a case, we would expect there to be an arc–trench complex. This would imply a vulcanic arc, as well as a nearby subducting ocean with the corresponding oceanic trench. Given that there are no signs of vulcanism & no signs of a nearby ocean, this is doubtful. If there were a nearby ocean, it would likely need to be separated from CWM by something producing a rain shadow, to explain the extreme dryness of CWM.

Nevertheless, an orogenesis of these structures is still plausible if it was one of continental collision. This would imply two distinct continental plates colliding in a way that has sutured them together, leaving evidence in the form of these mountains, & possibly the exhumation of UHP rocks (although there is no evidence of the latter). Given the obscure history & geography of Masteria, this is also plausible.

Looking at the sheerness & thinness of these formations, however, I still lean more towards the glacial/fluvial hypotheses.

Cataclysm

The almost complete lack of vegetation & the uniform rock surface both suggest to me that CWM as we know it resulted from some kind of cataclysm — & a recent one, at that. To attain such a uniform surface, even at all elevations of each ridge, either of two things had to’ve happened:

1. These formations were carved up from a single monolithic stratum, & that carving left behind little evidence of additional deposition on top.
2. The surface that we see was deposited recently.

(1.) doesn’t seem likely to me. For starters, that would be one huge stratum!! Plus, the idea that there hasn’t been any deposition since then would seem to imply that these formations are neither glacial nor fluvial, & presumably that the formation was very recent.

Then again, (2.) doesn’t seem particularly likely either. Deposition of this kind would presumably be due to wind transport, or perhaps to precipitation of mineral dust. Still, this could be plausible, especially given the thick reddish-brown cloud cover that we can also see in Fig. 2.

We can imagine (1.) happening because these were formed as mid-ocean ridges, later followed by some cataclysmic event of unknown description that caused the ocean to disappear — effectively drastically lowering the local sea level. Of course, this requires the ocean to disappear somehow, & I also have my doubts that such a process would produce such tall & thin structures.

For (2.), we need some source of the sediment that was transported here via wind and/or rain dust; although we don’t know of one, that’s not necessarily a dealbreaker, because the geography of Masteria is obscure anyway. If rain dust was involved, then we also have to explain how CWM now ostensibly lives in a rain shadow.

Either way, we need some way to explain the absence of vegetation. Even very hostile environments such as CWM generally have some vegetation that is specially adapted to the conditions. This is fundamentally why we need some kind of cataclysm to explain what we observe.

(A.) Raised from the sea

We’ve heard in myths & tales of Masteria sinking into the ocean (or part of what we now know as “Masteria” already — from the beginning — being part of the ocean, perhaps?), & subsequently re-emerging some time within generational memory.

This could explain how CWM was either some kind of mid-ocean-ridge-like formation, or how it could have been formed in the obvious way (glacially and/or fluvially), deposited upon recently, & then voided of basically all viable lifeforms.

(B.) Seismic catastrophe

Perhaps what we see is just ordinary rock that was deposited in a mundane way, but was then very rapidly uplifted into ridges by a catastrophic seismic event.

This could explain the morphology, & perhaps a sufficiently catastrophic — & importantly, very recent — seismic event could explain the extinction.

(C.) Victim of vulcanism

Perhaps the eruption(s) of a volcano caused a localised extinction event via abrupt & drastic climatic changes, vulcanic ash pollution/poisoning, etc.

This would allow for the structures that we observe to have arisen organically (presumably glacially/fluvially). Then, later on, the vulcanism would have caused the extinction, as well as the deposition of the reddish-brown rock surface that we see.

For better or worse, (A.) is largely contradicted by what we know about the history of Masteria. IsaacGS & Inugami’s What Do We Know About Masteria So Far (2020-11-08; archived) has this to say (emphasis in the original):

Masteria did not, in fact, sink into the ocean. According to Taggrin and Lukan, there was a huge flash of light and it was warped to “another world”. This other world is obviously Versal. This explains a lot of odd things, like how the flora and fauna could have survived while the continent was under the ocean.

This is an unfortunate detail partly because it invalidates (A.), but moreso because it adds a clear element of world-bending magic. Speculating about geomorphology & such is not quite as interesting when arbitrary things can be “just magic”. Still, it’s worth noting that (A.) can be partially salvaged by blaming the extinction on whatever happened whilst Masteria was in Versal.

Being somewhat more mundane, (B.) feels more believable, but does require a sui generis seismological event. It also seems somewhat unlikely that a seismological event on its own — even a fantastically extreme one — could cause an extinction of this kind.

This brings us to (C.), which I personally think is the most appealing of these three. It makes the origin of the observed geomorphology fairly straightforward & believable. Vulcanism is also known to be a relatively common cause of extinction events. Moreover, although the deposition of the vulcanic ashes/material does appear suspiciously uniform, it’s very possible that said deposition was aided by the thick clouds that we observe.

Fossil record

That’s not to say that there are absolutely no signs of life. There always are.

On the right-hand side of Fig. 3, we can clearly see two mammalian skeletons. It’s difficult to tell, but judging purely by their size & by the shape of their horns^[1], they appear to be bovids.

Near the top-left corner of Fig. 3, we see some kind of crimson-coloured bush-looking kinda thing. It’s prickly, & it hurts if you touch it, but those are about the only features that we can observe. If we wanted to identify it as a plant, we’d want to see some kind of reproductive structure… or maybe some leaves, roots… But we just don’t have a lot to work with here.

Still, the presence of these fossils without any corresponding living specimen, nor really much life at all, supports the “recent extinction cataclysm” hypothesis.

Erosion landforms

After climbing through, beyond, & above the Valley of Heroes, I arrived at the Lower Ascent. Although it may be “lower” indeed, this is nonetheless the first part of the ascent to the Crimsonwood Keep itself, which is at a very high elevation. It’s along this ascent that we start to see some more interesting things.

The increase in interestingness is partly just due to the proximity to the Keep; it was here where I bumped into so many of the corrupted myrmidons of the Twisted Masters. But the considerable increase in elevation also means a decreased impact of the vulcanism (or what have you) that caused the extinction event & its associated geomorphic changes. Along the ascent, I spotted many relatively untouched sandstone formations, like the one in Fig. 4.

The banding is typical of sandstone. The denudation that produced this structure was not mass wasting, but the kind of aeolian processes that might produce an alcove. Given just a little more geologic time, these might become little more than hoodoos.

Calcareous caverns

No simple alcove, this sandstone structure gives way to a vast interior cavern. One where I found the errant Jack Barricade, & of course, the wickèd creatures who would hold him captive:

At the behest of Lukan the Stormcaster, for “The Brewing Storm”, I collected ten of the Stormbreakers’ badges.

Did I mention that the caverns are vast?

Carbonates, do you think? Column-like speleothem formations could explain the consistently blocky & uneven nature of what we see in the representative Fig. 5. Carbonate rocks come in a variety of hues, so there’s no issue there. The formations hanging from the ceilings look awfully suspicious, though.

And what’s up with that flame on the right-hand side? This neatly-carved & symmetrical lighthouse thing means that someone’s been living here for a while, & they’ve been modifying the cavern as they see fit. The brilliant blue colour might mean that this flame is burning very cleanly (no soot), or perhaps there’s some copper or other blue-burning sort of salt in there. I suppose burning an extremely clean flame would make sense; wouldn’t want to smoke yourself out of your own cave.

Javelins, knives, & Shadowknights

It was in these caverns where I realised how much Taunt & Venom complete my arsenal. It’s a bit strange, right? My “main attack” is still a 1^st-job one, & yet these two 4^th-job skills fit like a glove.

The result is an utterly unmistakable combat style: FJing around, collectively immobilising & draining the HP of my enemies with SW, cleaving through them & applying Venom with so many Taunts, & powerfully finishing off individual monsters with my Double Stab.

Here, I’ll show you a neat trick:

I can keep moving whilst I’m waiting for my Taunt animation to end, if I can time a jump-Taunt correctly as soon as I land.

And there it is: ten Stormbreaker Badges. That was quite the workout, I must say.

Perhaps it was their difference in guild membership during the old days that has kept Lukan apart from Taggrin & the rest of the Raven Ninja Clan: Lukan a diehard Stormcaster, & the Raven Ninjas the most upright, wary, & conservative post-disappearance splinter of the Shadowknights. Or perhaps it was just an omok [오목] grudge…:

In any event, I had seen the Elderwraiths with my own eyes: bare skeleton torsos, robbed of their legs & skulls, but totally obscured by thick black cloth wrappings enveloped in a supernaturally purple misty flame. But I wasn’t eager to get anywhere near them.

According to Taggrin, these wraiths are the souls of brave Crimsonwood Keep soldiers who fell in battle whilst protecting the Keep from the invading Krakians. Not to place undue blame on the Krakians; it’s difficult not to invade the Crimsonwood Keep when you’re under the impossibly powerful magical thrall of The Alchemist. Taggrin promised a reward of trust if I could lay these fell spirits to rest for good, bringing back the ashes for a proper burial.

The gentler Phantom Forest

Although legend has it that the Phantom Forest is home to a very terrifying beast indeed, it’s otherwise markèdly less hostile than CWM. Compare Figs. 1 & 3 to what we see in Fig. 6 below:

It’s difficult to say much about the pedology here, for reasons that are perhaps obvious. Still, in contrast to what we see in Figs. 2〜4, Fig. 6 is not only an earthier, darker, moister-looking regolith, but it’s also visibly made of particles of varying sizes (rather than being monolithic). Granted, the bulk of it appears to be pretty large chunks of rock, some of which are large enough to be considered boulders. But this is still workable, & there are pebbles in there as well — plus something approaching real soil, to hold it all together!

These Phantom Trees are much healthier-looking than those Crimson Trees (especially the one in Fig. 1…), although they’re still conspicuously leafless. More interesting is the fact that the ground is covered in a layer of low-lying plants. I’d suspect that there could be lichens involved — lichens can, after all, live in just about any environment — but the pinnate leaves towards the bottom of the image don’t look like the lobes of a foliose lichen to me.

This leaves^[2] us with various plant possibilities, like e.g. mosses, which would presumably be in addition to some kind of graminoid plant(s) (Poales) responsible for the bulk of the foliage. These might be true grass(es), sedges, or possibly even rushes. Who knows, really. All of these families are pretty large (especially Poaceae) & diverse, including many species that are hardy & adaptable.

Still, the Phantom Forest appears to be lacking in biodiversity…

In any case, I killed quite a few of these sooty flying napkins with my dagger:

They weren’t forthcoming with the cards for a while, but I eventually got to them…:

…And brought their ashes back to Taggrin, who handled them gingerly, in preparation for burial rites.

El camino de la fortaleza

Lukan was still upset at the idea that the Keep, & the ascent to it, had been overtaken by soulless pretenders, & wasn’t entirely convinced that I wasn’t also on their side. To prove my loyalty, & to get rid of some of these pests, I stormed the castle to kill a number of Windraiders, Firebrands, & Nightshadows:

It was here at The Path of Strength where I learnt just how much I hate Nightshadows. I mean, they look cool & all, but with that much HP, the immunity to poison/venom was really killing my vibe.

El camino a la desertización

At the base of the PoS, I encountered one of the only true signs of life up here in CWM:

A real plant? In my CWM‽ It’s more likely than you think.^[3]

The armed (in this case, four-armed) shape of this cactus is the signature appearance of saguaro cacti (Carnegiea gigantea). The name is via Mexican Spanish saguaro /⁠säˈgwä.ɾo̞⁠/, ultimately of Uto-Aztecan origin. The Uto-Aztecan origin is due to saguaros only being found in the Sonoran Desert, in the northwestern bits of the Mexican state of Sonora & the southwestern bits of the U.S. state of Arizona.

There are no woody ribs showing, & this individual appears to be very much alive. It is, however, suspiciously short. It’s barely taller than me, which means that it’s either fairly young — perhaps only three decades or so old — or has had its growth rate stunted by a lack of water. Yet, it still managed to produce four arms; not all individuals grow arms at all, & those that do tend to be more mature at the time that their first arm begins growth. Our little cactus friend is trying real hard out here.

ℹ️ Not keen on the details? Feel free to skip ahead to the “Cactus ænigmaticus” section below.

Then again, saguaros are not the only cacti that can grow in this general sort of shape. Others would also be members of the subfamily Cactoideae. In the simplified table below (viz. Table 1), I’ve classified cactus species based on their habits, according to the following five-armed^[2] scheme:

Features relevant to Fig. 7 are highlighted. Taxa marked with a ※ are monotypic.

Like with most biological organisms, the taxonomy of these cacti is not necessarily well-understood or well-documented, & furthermore, is subject to change as more work goes into studying them. In cases where sources disagreed, I have — at the time of this writing (2024) — defered to Wikispecies for the correct names. Although it is arranged phylogenetically, Table 1 is neither exhaustive nor a phylogenetic tree, & intentionally omits several taxa that would necessarily be included in a full phylogenetic tree.

You might have noticed that the vulgar names are rather confusing: the same (or substantially the same) vulgar name may be used for multiple distinct species, & moreover, many species — or in biology generally, almost all species — have no common names at all.

Biogeographic division

You also might have noticed that all these species are only found in the Americas; & at that, none of them can even be found significantly north of the Mexico–U.S. border. Broadly speaking, this distribution encompasses what is in Udvardy’s scheme^[5] the Neotropical realm, in addition to the southernmost bits of the Nearctic realm. Similarly, in One Earth’s scheme^[5], this would be the combination of the Lower Neotropic, Upper Neotropic, & southernmost bits of the Lower Nearctic.

But biogeographic realms are vast & general, carving up the planet’s surface based largely on evolutionary ties (keeping palæogeography in mind), without really considering ecosystems & geological features. One Earth’s realms are composed of 52 subrealms, each of which is itself composed of one or more of their 185 bioregions, where bioregions are clustered into subrealms based on geographical relevance. This puts our region of interest largely into a combination of five subrealms: Mexican Drylands (#7), Caribbean (#11), Central America (#12), Amazonia (#14), and Andes & South Pacific Coast (#17). However, of these, the Mexican Drylands (#7) and the Andes & South Pacific Coast (#17) subrealms are overrepresented, forming the habitat of most armed cacti.

Maybe if I narrow my focus, I can go further. Looking at the entries in Table 1 that have useful highlighted features, I see that although a few members of Echinocereeae are good candidates, the genus Leucostele looks the most promising. Members of Leucostele are distributed roughly around the trijunction of Bolivia, Argentina, & Chile, plus regions of Chile further south, all the way down to the Coquimbo Region or so. In One Earth’s scheme, this is part of the Andes & South Pacific Coast subrealm, and is more specifically encompassed by these three bioregions: Andean Mountain Grasslands (NT5), Chilean Matorral Shrublands & Savanna (NT6), and South American Coastal Deserts (NT8). These bioregions are further composed of ecoregions, of which the relevant ones are: Central Andean Dry Puna (#587), Central Andean Puna (#588), High Monte (#592), and Chilean Matorral (#596).

Puna here refers to the puna grassland, a biome-defined ecoregional designation used specifically for the central Andes. Matorral here is a Spanish word referring to shrubland, originally that kind found in Spain. In the Americas, matorral is used to refer not just to shrublands & wooded areas in Mediterranean climates, but also to xeric^[7] ones, as well.

Biomes & climates

Now that I know what kinds of ecoregions I’m working with, I can attempt an overview of the environmental conditions that many members of Leucostele live within. In the table below (viz. Table 2), climates are classified according to the Köppen–Geiger–Trewartha climate classification (KGTC). Temperatures are rounded to the nearest degree Celsius (°C), & precipitation is given in millimetres (mm). Elevations are given in metres (m) above mean sea level.

And, since the saguaro is the point of comparison here, I may as well do the same for Carnegiea:

The species here (excepting C. gigantea) mostly represent just two biomes^[6]: (a.) Mediterranean forests, woodlands, & scrub; and (b.) montane grasslands & shrublands.

In the former case (a.), I’m really talking about scrub, & not so much forests nor woodlands; this is the Chilean Matorral ecoregion. But don’t be fooled by the biome nomenclature; the climate in these environments varies considerably, & not all of it is well-described as “Mediterranean”. Table 2 has three entries associated with Mediterranean scrub, but only one has a climate characterised in this way: the Quillota Province has a KGTC of Csak, & the most rainfall of any entries in Tables 2 & 2.1: 400〜500⁢ mm annually. It is, however, on the hottest end of Cs (corresponding to Köppen–Geiger Csa). On the other hand, Ovalle is considered semiarid — in other words, a steppe climate — with a KGTC of BShk, & Vicuña is perhaps the most desert-like climate of any entries in Tables 2 & 2.1, with a scorching dry KGTC of BWal.

On the other hand, (b.) is well-described as a steppe or shrub–steppe — although the latter term appears to mostly be used in North America, & maybe not so much South.

Table 2.1 includes the “deserts & xeric^[7] shrublands” biome, although the “representative location” that I chose for C. gigantea gets enough rainfall to be classified by KGTC as BShl. ≈300⁢ mm annually isn’t much, but it’s quite a bit by xeric^[7] biome standards, which perhaps explains why the saguaros in that park are so damned enormous…

As can be seen in Table 2, members of Leucostele tend to be found at considerable elevations, even upwards of ≈3⁢ km in extreme cases. This is, of course, due to the influence of the Andes. But, at any elevation, the finer details of climate are important for the survival of cacti, as explained with relation to C. gigantea by the National Park Service (which oversees the Saguaro National Park):^[8]

Because they are not keen on the cold temperatures, saguaros are also limited by elevation. They are generally found growing from sea level to approximately 4 500⁢ feet [≈1 372⁢ m] in elevation. Saguaros growing up to 5 000⁢ feet [≈1 524⁢ m] in elevation are usually found on south-facing slopes where freezing temperatures are less likely to occur or are shorter in duration.

Xeromorphy

This inability to cope with cold temperatures is related to what makes cacti so special: their xeromorphy. The members of Cactaceae have evolved to live in xeric & seasonally hot environments, which means adapting to increase water-use efficiency, as well as to decrease damage due to heat & UV exposure. Despite being ordinary flowering plants just like any other member of Angiospermae, cacti — in particular, focussing on the kind seen in Fig. 7, which I’ve loosely identified with the real-life Leucostele — have a number of fun xeric survival tricks up their sleeves:

• Being a succulent enables the cactus to store large quantities of water for later use as needed. This is why the cactus’s stems have ridges: when the stem fills with water, the ridges flatten out to accomodate the expansion.
• Many cacti have shallow & wide-reaching root systems that allow them to gather even the smallest amount of water that is deposited on the soil’s surface.
• The cactus’s leaves are prickly spines instead. The first thought might be that this serves to defend the cactus from herbivores, & it probably does. But importantly, it also greatly reduces the surface area of the place where the plant would otherwise lose large amounts of water via transpiration.
• You may have noticed that many of these cacti, like e.g. L. atacamensis, are visibly hairy/fuzzy (tomentose). This greatly reduces airflow, further reducing the possibility of evaporation. An extreme example of this is Cephalocereus senilis “old man cactus”, which is covered in a thick mat of white hair reminiscent of the hair of an old man.
• When water loss cannot be tolerated, the cactus may close its stomata; thick, impermeable cuticle helps the cactus to retain water when its somata are closed. Epicuticular wax is responsible for the bulk of this coating, which in turn is chemically composed mostly of long hydrocarbon chains that may be fatty acids, or other compounds depending on the functional groups at the ends. The wax can also help to protect against UV damage.
• The cactus makes use of Crassulacean acid metabolism (a.k.a. CAM photosynthesis) instead of C₃ or C₄ carbon fixation, to not only improve photosynthetic efficiency, but also to allow the cactus to keep its stomata closed during the day. By opening its stomata at night, capturing CO₂ at this time, & then storing that CO₂ as malic acid (HO₂CCH(OH)CH₂CO₂H; note the HO₂C- & -CO₂H, both carboxyl groups), the CO₂ can later be used during the day — whilst the requisite sunlight is available — to photosynthesise. Being able to photosynthesise without opening the stomata is a massive win in xeric environments, because opening them during the day is how most plants lose ≈97% of their water.
• The cactus is hesitant to germinate, & to grow. Many cacti are absolutely tiny, even many years after germination; nevertheless, they can get very, very large, & their lifespans may be measured in centuries. This allows the cactus to be strategic about when it commits water & other resources to tissue growth, & also to be more efficient when it does so.

Many of these adaptations are quite drastic & require considerable evolutionary time. From this, I can infer that the population to which the individual in Fig. 7 belongs (or belonged…) was already firmly established in CWM prior — presumably long prior — to whatever cataclysmic event occurred (as discussed in the “Cataclysm” section above). Either that, or it was introduced artificially…

Other material conditions

I’m definitely getting somewhere here, but I’m still missing at least two other ambient features of great importance: the soil, & what other species the cacti associate with.

I can pretty clearly see in several of the images above — not the least of which being Figs. 2 & 3 — that the “soil” of CWM is really more like bare, solid rock. The surface of CWM is thus characterised more by whatever regolith it might have, plus the stone substrate itself. Still, for many plants, growing upon little more than rock is perfectly doable: these are lithophytes. For example, some members of Crassulaceae — for which “Crassulacean acid metabolism” is named — are lithophytic.

Cacti are not usually lithophytic. Nonetheless, some species, like e.g. Epiphyllum oxypetalum “queen of the night; Dutchman’s pipe cactus”, are habitually epiphytic, and thus must acquire their resources from rain, the air, & any debris that they manage to passively collect. This can make such cacti situationally/opportunistically lithophytic, as they already don’t need soil to begin with. But these cacti don’t look anything like our armed friend in Fig. 7, nor like any in Table 1.

However, a 2004 paper by M. E. Puente, Y. Bashan, C. Y. Li, & V. K. Lebsky^[9] found wild examples of four species — viz. Pachycereus pringlei, Stenocereus thurberi, Cylindropuntia cholla, & Ficus palmeri — that were enabled to live lithophytically by the presence of bacterial & fungal colonies on their rhizoplanes that were responsible for nitrogen fixation & the provision of important mineral nutrients. The paper doesn’t identify all species involved, but the basic idea is similar to the way in which members of the nitrogen-fixing bacterial genus Rhizobium enable legume growth: the association is basically mutualistic, with the bacteria fixing nitrogen & making already-existing nutrients more available to the plant’s roots, & the plant providing carbohydrates & similar to the bacteria.

Of the four plant species in the 2004 paper, two (viz. P. pringlei & S. thurberi) are in Table 1, one (viz. Cylindropuntia cholla) is a cactus species not closely related to any of those in Table 1, & the remaining species (viz. Ficus palmeri) is not closely related to cacti at all — it’s a fig tree. Moreover, the paper has this to say about the root-colonising species:

Cell and colony bacterial morphotypes among the species were many and varied. Because similar culture media used for recovery of culturable bacteria from different plants yielded different bacterial populations, we concluded that different plants support different levels of colonization, and that root colonization by culturable bacteria is not homogeneous.

Taken together, it seems that although this process of rhizoplanar colonisation allowing plants to live lithophytically (effectively aeroponically) is seemingly uncommon, & seems like it requires at least some series of coincidences, it nevertheless likely occurs repeatedly, & with a wildly different community of plant, fungal, and/or bacterial species every time.

Cactus ænigmaticus

The only other place where we see cacti is in the Nihāl Desert, which leads us to two competing hypotheses:

(I.) Exotic species

This cactus is native to somewhere in the Nihāl Desert, & someone — presumably The Alchemist — brought it here, intoducing it to CWM.

This is the simplest & most convenient explanation. It also adds another connection between Masteria and Nihāl, a connection implied by Masteria’s original — but fragmentary — mythopœia.

(II.) Recent lone survivor

The recent cataclysmic event (as discussed in the “Cataclysm” section above) wiped out virtually all lifeforms previously inhabiting CWM — excepting perhaps those maintained by magical/spiritual forces… — but was more likely to spare that organic matter found at higher elevations; that is, closer to the Keep. A fortunate cactus seed (or seedling) along the Ascent, together with at least some microörganisms, survived in a condition viable for the (re-)establishment of lithophytic growth of the cactus along with its microbial rhizoplanar colony.

This is the most natural explanation, & implies more information about CWM as a biogeographical area.

Although (I.) is certainly convenient, it’s perhaps too convenient. We don’t actually know any details about the connection between Nihāl and Masteria, which is made worse by the fact that these two locales are greatly geographically separated. Recall that the Nihāl Desert is on the continent of Ossyria, making the Nihāl–Masteria separation akin to the IRL separation between the Old World & the Americas.

Worse still, the identification of the Nihāl region with the IRL Middle East — up to & including Persian-language names, Arabic-language names, & explicit references to Arabic literature — is already biogeographically problematic to begin with. Excepting Rhipsalis baccifera “mistletoe cactus”, there are no cacti outside the Americas! And unfortunately, R. baccifera couldn’t look more different from what we see in Fig. 7 — & what we see in Ariant, for that matter.

Just to add another nail in (I.)’s coffin, we have to consider how much less likely the “lithophyte via microbial rhizoplanar colony” hypothesis is if (I.) is assumed to be true. This is already a fussy arrangement that would likely only come about in response to the conditions found on CWM. Yet Ariant is geologically, ecologically, & meteorologically wildly different from CWM in a number of ways.

On the other hand, the real-life case of R. baccifera inspires a synthesis of (I.) & (II.): although we can’t know for sure, it’s speculated that R. baccifera made its way to the Old World via migratory birds, via early sailors/rafters, or perhaps even by emerging prior to the breakup of Gondwana. In any case, an ancient transplantation of this kind — analogous to a transatlantic crossing that occurred many hundreds, thousands, or perhaps millions of years ago — is a tantalisingly plausible explanation for both why we see a cactus in Fig. 7, & also why we see naturalised cacti species across much of Ariant.

This would essentially be (II.), with the addition of the converse of (I.) also being true: rather than cacti being Nihāl-native species very recently introduced to Masteria, they would be Masteria-native & introduced to Nihāl in ancient times.

CWM today

If this synthesis of (I.) & (II.) is true, or even if (II.) is true at all, then the specimen in Fig. 7 tells us quite a lot about CWM. It is worth noting that, even if (II.) is utterly incorrect, this does not imply that the specimen in Fig. 7 tells us little or nothing about CWM; the mere fact that this cactus survives in CWM as we know it tells us plenty of things about CWM, even if it ultimately tells us more about the cactus itself.

In any case, we can summarise these inferences like so:

• If the specimen in Fig. 7 is native to CWM, then the corresponding closest real-life analogue of the CWM environment would (likely) be somewhere within One Earth’s^[5] “Mexican Drylands (#7)” and/or “Andes & South Pacific Coast (#17)” subrealms. This spiritually places CWM firmly within the Americas, & more specifically South America — or perhaps somewhere within the North American historical region of Mesoamerica.

• The specimen in Fig. 7 most closely resembles a member of the genus Leucostele, specifically a member of one of the six (sub)species listed in Table 1. Leucostele is only found in South America.

• The climates (expressed in terms of Köppen–Geiger–Trewartha classification) in which this sort of cactus may be found range, broadly speaking, from something like Csak to something like BWhk. A general characterisation of this would be:

  • A hot desert (i.e. arid) climate (BWh), or perhaps hot steppe (i.e. semiarid) climate (BSh).
  • A mean annual precipitation level of 100〜400⁢ mm, so perhaps somewhere around ≈250⁢ mm.
  • Hot-monthly mean temperatures downwards of roughly 24〜29⁢ °C, & cold-monthly mean temperatures upwards of roughly 8〜11⁢ °C.

• The distribution of the species represented in Fig. 7 is across an elevation range of probably something like ≈1 500〜3 500⁢ m or so.

• CWM’s biome would then likely be best characterised as “montane shrublands”, although “(semiarid) Mediterranean scrub” or “deserts & xeric shrublands” are also possible.

• The geomorphology of the mountains themselves is likely fluvial or glacial in origin, depending on whether the corresponding valleys are V-shaped or U-shaped, respectively. A recent volcanic cataclysm is the likely reason why the mountain surfaces are surprisingly uniform & barren of life.

• Aeolian formations sculpted from sandstone can be found somewhat frequently at the higher elevations of CWM.

• The caverns of CWM are characterised by calcareous formations, especially speleothem-like structures.

• Cacti, like that represented in Fig. 7, are probably indigenous to CWM. Their presence in the Nihāl Desert is attributable to either their being introduced to Nihāl at any time in the distant past, or their having naturally spread to Nihāl in (pre-)ancient times, prior to the great geological separation of Masteria & Ossyria that we see today.

Desertización

Although CWM has likely always been arid and/or semiarid, it perhaps wasn’t always a “desert” in every possible sense of the word. The cataclysm discussed in the “Cataclysm” section above — regardless of whether or not it was fully or partially anthropogenic — is easily a driver of catastrophic desertification effects, before even considering any other causes that might also contribute to desertification.

We even see some of the classic effects of desertification firsthand, like the huge duststorms that loom large over the CWM skies.

The result is that CWM is now utterly biologically unproductive, & shows no signs of recovery in the near future. This is especially true given that the inhuman minions of the Twisted Master are unlikely to give a single solitary fuck about desert greening or whatever. In fact, because of the geological changes that resulted from the cataclysm, desert greening on its own isn’t gonna cut it. Massive soil regeneration is only the first step…

Anyway, I think that’s enough staring at a certain cactus for now…

El símbolo de heroísmo

Even with the sweat running down my face as I painstakingly slew each Nightshadow, I was able to get those 75 kills of each monster that Lukan requested, & without too much trouble. I even managed to get one (1) Nightshadow card along the way…:

I also found a Firebrand card, which is a set that I might actually finish:

I also found a great monetary mystery:

What are the meso sacks doing? Why are they floating out in the void, taunting me with those shiny meso coins that I can see peeking out at the top? Why is it called Hollowed Ground, instead of Hallowed Ground? These are the questions that will be with me forever.

Now that Lukan was sufficiently impressed with my ability to kill random bad guys on his behalf, he had a better idea for what I should do: make the Marker of Heroism, so that I can teleport around the Valley of Heroes like some kind of ancient Masterian wizard. For that, though, I would need a few Typhon Feathers:

Damn those Typhons. Damn them for also having wings. I’m supposed to be the one with wings!!

It’s fine, though. I only fell once… or twice. Or— Nevermind. The point is that I have the materials to forge the Marker of Heroism. Acquiring the Gold Ore was especially easy, as I’d only to melt down the gold grills of a couple unsuspecting Pigs. And would you look at that!:

Wowie. The Marker of Heroism, the Map of Phantom Forest, and the Crimsonwood Keystone‽ Looks like I have just one other order of business to take care of in Masteria…

A card in the hand is worth two in the mountains

That’s right. Did you forget already? I came to Masteria for the card sets, you silly goose! I was keen on finishing the Windraider & Stormbreaker sets, so I headed back to the Cavern of Pain:

But it wasn’t exactly all sunshine & rainbows. When they named it the “Cavern of Pain”, they were not fucking around. The basic vibe is this: I run around SWing, Double-Stabbing, & Taunting for a while, and then kinda pass in & out of waking consciousness, occasionally being awake enough to continue SWing, Double-Stabbing, & Taunting.

Oh, my god. The Stormbreakers do drop cards! How kind of them… I— I definitely need another big fat nap…

Interludio soñoliento

Descansada

Le sigh. Really, I’m still tired, but I guess I’m awake now, & there’s nothing that I can do about that. At least I can probably finish the Stormbreaker set within the next two or three naps…

…Okay, okay. I’m done counting sheep. Or… naming sheep. “Mumph” is a cute name, right?

In any case, I’m 5⧸5 on Stormbreakers, so it’s time to take the walk of shame to the Cavern of Shame.

It’s time to farm Firebrand cards, & if I’m lucky, I might just finish the Windraider set as well. Unfortunately for me, however, the Windraiders had other plans; rather than dropping cards, they dropped GFD15s:

So, a few big naps & a few big Firebrand cards later, I headed back to Lower Ascent to finish the Windraider set. Although I did find some Windraiders, I also found a mysterious floating inmyleague:

He never responded. Kind of rude to just float out there in the void, staring at me, without having the common courtesy to say “hi” back.

Eventually, the Windraiders faltered. Their lengthy & arduous campaign to withhold cards from me had met its match in my unyielding attrition:

Phewf. You know, I could try to farm the Nightshadow set. I really could.

Blue in tooth & fin

So anyway, I went to Aqua Road, right? Upper Aqua Road is a pretty good spot for card sets, & I haven’t seriously farmed it on this character yet. I have Taunt now, so maybe it won’t be so bad…

Thank you, Poopa. 🙏🏽

I have occasionally idly wondered if the name of this piscine species is intended to be the English pooper /⁠ˈpuː.pə(ɹ)⁠/ “anus; fun-killer [in phraseme party pooper]; toilet; one who poops”. The spelling ⟨Poopa⟩ suggests the pronunciation /⁠ˈpuː.pə⁠/. Notice how they sound the same? That’s because it’s literally the exact same audio; in non-rhotic accents, /⁠-ə(ɹ)⁠/ and /⁠-ə⁠/ are the same thing (hence the round brackets).^[10] This is reinforced by the Kenta’s Research questline, where the corresponding DNA sample is called Pooper’s DNA Sample.

The original name is 푸퍼 ⟨Pupeo⟩ /⁠pʰu.pʰʌ̹⁠/. /⁠ʌ̹⁠/ is the closest thing that Korean has to English /⁠ə⁠/, which has the stressed variant /⁠ʌ⁠/ in many accents of English^[12]. Alternatively, English has /⁠ɔː⁠/ as in paw /⁠pɔː⁠/ — note that /⁠ʌ̹⁠/ & /⁠ɔ̜⁠/ denote the same thing. This latter interpretation would probably get the spelling ⟨Poopaw⟩. Vowels are fun! Right‽‽ (😭)

More seriously, the variant spelling ⟨Pooper⟩ was likely written by someone who interpreted 푸퍼 as an adaptation of non-rhotic English pooper, since that’s an actual word in English. The spelling ⟨Poopa⟩ takes it more at face value as a fanciful name. There’s also the word pupa, but that refers to a particular develomental stage of an insect, & it’s pronounced /⁠ˈpjuː.pə⁠/ anyway. However, as far as I can tell, 푸퍼 is primarily used in Korean as a borrowing of English puffer /⁠ˈpʌ.fə(ɹ)⁠/. Go ahead & try putting 푸퍼 into your search engine of choice; you’ll be greeted by many images of people in insulated “puffer” jackets.

If that seems weird to you, consider the fact that Korean has no equivalent of /⁠f⁠/. This makes /⁠pʰ⁠/ — which, in Korean, contrasts with unaspirated /⁠p⁠/ — the closest alternative. But because Korean /⁠pʰ⁠/ is already equated with English /⁠p⁠/, & Korean /⁠p⁠/ with English /⁠b⁠/ (Korean makes an aspiration distinction, whereas English makes a voicing distinction), we have English /⁠p, f⁠/ ↦ Korean /⁠pʰ⁠/. This rather unfortunate many-to-one mapping was enough to confuse the poor GMS localisers, who couldn’t put together the fact that 푸퍼 was an adaptation of English puffer, as in puffer fish 🐡, this whole time!!

And indeed, in English, pufferfish is the most common vulgar name for members of the family Tetraodontidae. Although they do have a lot of other vulgar names as well, including ridiculous ones like honey toad. Honey toad? The fuck?

Pufferfish are distinguished by their unique methods of deterring predators. They’re neurotoxic, & when threatened, can become ball.

They seem kinda like cute bouncy-balls, at least until you die of whole-body paralysis. Oh well.

Luckily for me, I don’t need the Poopas themselves — I just need those sweet, sweet cards. And the Jr. Seal cards that come with ’em:

Did I mention that they’re extremely neurotoxic? Poison Poopas are less cuddly than they look…:

Overall, it’s really not too bad out here. Sure, even Taunt is pretty bad at hitting swimming monsters, but at least the card droprates here are comparatively good.

Over at Forked Road : West Sea, I began my hunt for Freezer…:

…And Sparker cards:

By the look of those pinkish seal pups on their heads, I suppose Freezers are the “Sr. Seal” counterparts of the Jr. Seals. Sparkers, on the other hand, look more like walruses; in fairness, walruses (Odobenus rosmarus) are a kind of seal (a.k.a. a pinniped, i.e. of the parvorder Pinnipedia).

Wait a second… that doesn’t make sense. Pufferfish are mostly found in the tropics, & can’t be found anywhere near the polar zones. But seals can’t be found in the tropics, & walruses are only found in the polar zones! Hello??? Immersion ruined. Absolutely unplayable.

Castles made of sand

Moving on to the eastern wing of the Upper Aqua Road, I hunted for Flower Fish cards at Sand Castle Playground:

Sand Castle Playground is “that one really tall map”, & true to its name, it does have something like eight or so similar-looking sandcastles scattered across it.^[13]

This sandcastle is mostly made of very golden-coloured sand, but it also appears to have some seashells lodged in it. A seashell is usually the disembodied exoskeleton of a mollusc (member of the phylum Mollusca), which is calcareous & thus formed by everyone’s favourite type of mineralisation: biomineralisation. I’m not really seeing any molluscs down here — other than Squids & Risell Squids, which are molluscs, but don’t produce shells — although these shells are quite smol in comparison to any of the monsters that I’m card-hunting, so maybe I’m just not seeing them, or they mostly live in shallower waters or something.

It’s known that members of Torquigener albomaculosus — a species of pufferfish — create elaborate art in the sand as part of a mating ritual. With this in mind, it’s not at all out of the question to suppose that the sandcastles in this map could be created by fish native to this region, presumably Flower Fish or Bubble Fish.

There’s also the aquatic shrub on the right-hand side of Fig. 9, which is similar to the other shrubs that can be seen throughout this map. I don’t really claim to know anything about this organism’s taxonomy, as I don’t reckon I’ve ever seen anything like it before, but there are at least a handful of useful observations to be made:

• This looks very similar to a shrub, complete with a woody-looking stem system that has a clear trunk & subordinate branches. This would at least classify it as a plant (a member of the clade Embryophyta^[14]), but unfortunately that’s an extremely general taxonomic rank.

• This plant is fully submerged & attached to a substrate on the seabed. This would make it a submerged macrophyte. This is, however, purely a niche- & morphology-based classification, so it doesn’t tell me anything about the taxonomy.

• Given the presence of shrimp (Krips) & pufferfish, the water here is at least saline enough to be that of an estuary — generally speaking, brackish water (salt content upwards of 1〜30⁢ g∕L). Unless the salinity in this map is unusually low, the plant in Fig. 9 cannot be a glycophyte (a plant intolerant of salt), & is thus a halophyte. If(?) this is ordinary seawater, then the salinity is probably more like ≈40⁢ g∕L, which is about 20 times more than the upper limit of what your typical glycophyte can tolerate. Being able to tolerate more than about 10⁢ g∕L would make this plant a euhalophyte.

  The combination of being a (eu)halophyte & being a submerged macrophyte makes this plant a hydro-(eu)halophyte. Real-life examples of hydro-euhalophytes include members of the genus Zannichellia “horned pondweed” and some members of Althenia “water-mat”, especially A. marina. However, these real-life plants are morphologically nothing like the one seen in Fig. 9, & are better described as filiform herbs.

• The flowerage of this plant ranges in hue between dark pink & light pink. This is reminiscent of the showy pink blossoms of a cherry blossom tree like e.g. Prunus serrulata “Japanese cherry; サトザクラ ⟨satozakura⟩”.

• Although the flowerage is certainly pretty, I don’t see any foliage. Where’s the chlorophyll? I don’t see any green!! Unless all these shrubs are simply on the brink of death, they must be deciduous, meaning that their leaves seasonally fall off.

  I suppose it’s possible that at least some of the white/pink nodes are actually unusually-coloured foliage. Some deciduous trees produce pigments in their leaves before shedding them, leading to autumn-coloured foliage. But anthocyanin is a deep purple colour that I wouldn’t describe as “pink”. It’s plausible that a pinkish effect could be achieved by diluting with something very whitish or reddish-white, which might(‽) be explained by the leaves being naturally whitish in the absence of pigmentation, along with perhaps some influence of carotenoids. But… probably not.

• The stems of this plant are quite white. Since I already decided that the stems are woody, it would be safe to further assume that this is white bark. Whitish bark is not terribly uncommon, although of course none of the relevant species can be found underwater: Pinus albicaulis “whitebark pine” of the North American regions of Cascadia & the northern Rocky Mountains; various birches (members of the genus Betula) have white bark, e.g. B. platyphylla “Asian white birch” of East Asia, B. papyrifera “paper birch” of northern North America, etc.; certain members of Corymbia that’re known as “ghost gum” trees, e.g. C. aparrerinja, of mainland Australia; etc.

I suggested that the plant in Fig. 9 is a “plant” in the sense of Embryophyta^[14], but I think that I can go at least a bit further by calling it a flowering plant (a member of the clade Angiosperms). Considering that, in the above list, I compared the plant fairly closely to monocots (members of the clade Monocots, e.g. the hydro-euhalophytes mentioned above), but also to various eudicots (members of the clade Eudicots), it’s not really possible to go any narrower. (If I excluded monocots entirely, then I could go down to the clade Rosids…)

A mushroom coral hill to die on

In any case, I of course did the Bubble Fish set as well:

Moving on to Blue Seaweed Road, I did the Krappy set:

The name of this species is 크라피 ⟨Keurapi⟩ /⁠kʰɯ.ɾɐ.pʰi⁠/, from English crappie (sometimes croppie) /⁠ˈkɹæ.pi, -ɹɒ-⁠/, the vulgar name for the genus Pomoxis. The further etymology is uncertain, but it might be from dialectal Canadian French crapet /⁠kʁɑ.pɛ⁠/ or crappé /⁠-e⁠/.

But you know, they’re clearly supposed to be clownfish (a.k.a. “anemonefish”; subfamily Amphiprioninae) or something, except with one weird horn. Right? Like a… uniclownfish. 🦄🤡🐠 Unfortunately, crappies & clownfish are not closely related, & their distributions aren’t even close to overlapping: eastern North America vs. the Indian Ocean.

Along this same road, I started the Cico set…:

…And accidentally finished the Pinboom set, as well:

Moving along to Mushroom Coral Hill, I also started the Seacle set:

Mushroom Coral Hill is a map with an unusually high level of visible diversity. I don’t even know what half of this shit is!

In the background — including mostly silhouettes, but also some full-colour fish like the one on the far right-hand edge of Fig. 10 — I can see some fish that aren’t found as monsters within the world, some of which might even be small sharks (cartilaginous fish of the subcohort Selachimorpha).

I can also clearly see some kind of seaweed along the floor. These are quite green — only a few laminæ are slightly brownish, at best — & all laminæ form distinct long, finger-like blades that float upwards. The term “seaweed” is a vulgar name that doesn’t really mean much, but the finger-like blades of this particular seaweed suggest that it’s probably a kelp — that is, a member of the order Laminariales.

Like most organisms that are considered “seaweeds”, kelp is not a kind of plant, except perhaps in a centuries-old obsolete sense of the term.^[14] In vulgar terms, kelps are “protists”; unfortunately, however, the kingdom *Protista is somehow even more useless than *Plantae.^[14] I didn’t even know that that was possible… The problem is that “protist” is really a catch-all term for any eukaryotes that aren’t animals, aren’t plants, & aren’t fungi, either. Defining things negatively (i.e. in terms of what they are not) tends to produce murky & unnatural groupings, & the so-called “Protista” are no exception.

Kelps are better described as stramenopiles (members of the superphylum Stramenopiles), & more particularly ochrophytes (members of the subphylum Ochrophytina). This means that their motile cells have two very different flagella, & being in Ochrophytina makes them stramenopiles of the photosynthesising kind^[15].

Each group of kelp laminæ in Fig. 10 looks to be that of a kelp bed. Kelp anchor themselves to the seabed with their holdfasts, which differ from the roots of plants in that they only anchor the organism to its substrate, without uptaking any resources. Still, if the holdfast is like a root system, then the analogue of a stem would be the kelp’s stipe. Stipes can even have phloëm-like structures that actively transport resources through the kelp, although stipes are a few hundred million years of evolution behind phloëm, as a result of being much younger. And finally, the analogue of a leaf would be the kelp’s thallus, of which we mostly see the blades — that is, the laminæ. Unlike leaves, however, thalli consist largely of undifferentiated cells, & thus lack vascular tissue. All of these similarities (analogous traits) are the result of convergent evolution.

And of course, most of what I see in Fig. 10 are the eponymous mushroom corals! “Mushroom corals” (or “plate corals”) is the vulgar name of a real family of corals: Fungiidae. Unlike most corals, most of the Fungiidae are solitary, & thus have their own hard skeletons. Corals are more usually known for their clonal colonies of genetically identical individuals called polyps. I’m not so sure that what I see in Fig. 10 is in Fungiidae, as it looks more like a fanciful depiction of the calcareous castle structure formed by a colony of many polyps — along with, of course, the brilliant colours given by the zooxanthellae that they live in symbiosis with.