Species Design Considerations

[Quazifuji]

Big part of why "fire dude" doesn't work as a race is...well two things. (1) there are a lot of races; so small changes compared to every other race doesn't really exist. and (2) since almost every fire spell is Fire/Conj, 3! Fire is unlikely to outclass 3! Conjurations (Tengu) or 3! Spellcasting, 1 Conj, 1 Fire (Deep Elf)without a lot of Int...

Yeah, I think in general, there seem to be two opinions that are pretty widely held among many of the regulars and devs on this forum:

1. Aptitudes alone are not enough to justify the creation of a new race. It needs a more interesting and novel concept than having high aptitudes specialized in something other races don't (such as high fire aptitude).

2. Versatile aptitudes are more interesting than special aptitudes. They allow the race to be used in more different ways, and encourage you to think about how to use a race. They also allow the race to adapt better to whatever you find when playing the game, and adapting is a huge part of DCSS's strategy. This is why many recent races like Vine Stalkers and Formicids have relatively flat aptitudes. If a race is created specifically for the purpose of having good aptitudes for a Fire Elementalist, then that just means 90% of people who play them will make a Fire Elementalist (or maybe Wizard of Conjurer), and that's boring.

This is the problem with most of the "fire" or "ice" race proposals that pop up in this forum. Point 1 is why "fire" or "ice" isn't interesting enough, and point 2 is why "fire" or "ice" may actually be a bad thing that makes the race less interesting even if it does having some other novel feature. That's not to say a fire or ice race will never exist, but I'm not sure how likely one is to be created starting from the premise of wanting to create a fire or ice race.

[duvessa]

Okay, so here's the thing about skill aptitudes: by Crawl's official standards, they are a bad design.

  Code:

Major design goals
  * challenging and random gameplay, with skill making a real difference
  * meaningful decisions (no no-brainers)
  * avoidance of grinding (no scumming)
  * gameplay supporting painless interface and newbie support

Minor design goals
  * clarity (playability without need for spoilers)
  * internal consistency
  * replayability (using branches, species, playing styles and gods)
  * proper use of out of depth monsters

Avoiding no-brainers is a major design goal. The whole point of the skill aptitude system is to create no-brainers. When a species has unequal aptitudes in weapons/spells/melee vs. ranged combat vs. conjurations, you obviously default to the option with the highest aptitude, assuming the options are balanced against each other in the first place. So there are only a few things that could potentially make skill aptitude a reasonable feature:

• Skilling options aren't balanced in comparison to each other. For example, back in 0.5, maces were total crap and if you had +0 maces and +0 axes you'd always use axes. But that's an obvious balance problem in its own right that needs to be corrected for all species, so trying to fix it using a species-dependent system makes no sense.
• All of a species' aptitudes are the same, so you have species that are generally better at all skills and species that are generally worse at all skills. Mummies vs. demonspawn vs. demigods vs. humans are almost an example of this. But then why are there separate aptitudes for different skills? If you're going to do this, there should just be one aptitude that controls all skills.
• A species has some special feature that makes a skill abnormally useful/useless compared to other species. Ogres and trolls are good examples; maces skill and unarmed combat skill are abnormally useful to these species. If ogre's maces aptitude was lower than all its other weapon aptitudes, and troll's unarmed aptitude was lower than all its weapon aptitudes, then the aptitude difference would actually help balance them. In these examples, that's literally the opposite of the status quo (ogres' maces aptitude is their highest one, and troll's unarmed aptitude is their highest one), but hey, that could change.
  Since this pretty much never happens in the game currently, I think the devs regard it as a clumsy approach to balance or something like that. And if that's the case, I'd agree with them.
• You are willing to sacrifice meaningful decisions (a major design goal) to make species more different from each other, presumably for replayability (a minor design goal). This isn't a great trade, which is why you see it less and less often these days.

I believe at least some of the DCSS devs basically agree with this, because when skill aptitudes change, it's usually either to reduce their impact (because their impact on the game is bad), or to compromise a major design goal (no no-brainers) in deference to a minor one (replayability) or something else that isn't even a design goal (flavour).
The result is that skill aptitudes that actually do something are mostly restricted to older species like HO and HE and DE and Mi, which a lot of people don't really want to change because "human with shallower strategic decisions" is their entire design, and getting rid of their skill aptitudes makes them just "human with sexier ears" which would probably get removed from the game pretty fast. If you want to add a new species with lopsided skill aptitudes you have to first attach some gimmick to it so it's a candidate for addition at all, and then push through the stupid aptitudes on the basis of flavour or replayability. The success rate of this is actually pretty good, it worked for Fo/Gr/VS and even for Op's stealth aptitude.

I specified "skill aptitudes" here, but MP aptitude probably has a similar problem. HP aptitude is fine. Experience aptitude is basically redundant with HP/MP aptitude, since the only consequential effects of XL are HP/MP and occasionally spell levels.

[Lasty]

I agree with the spirit of, if not the full text of, duvessa's post. I think flat aptitudes are generally much more interesting design than varied ones, and in particular when the aptitudes are of otherwise quite similar things -- among melee weapons, among ranged weapons, among (most) spell schools.

[archaeo]

But that said (e: by which I mean I also agree broadly with the idea that flattish apts are better than hugely lopsided apts), I'm puzzled by the idea that lopsided aptitudes fail the "no-brainer test," given that Crawl doesn't make you play any given species. Sure, after you've picked a species like, say, DE, your number of meaningful decisions is reduced, but isn't the act of picking a species a meaningful decision?

[duvessa]

But that said (e: by which I mean I also agree broadly with the idea that flattish apts are better than hugely lopsided apts), I'm puzzled by the idea that lopsided aptitudes fail the "no-brainer test," given that Crawl doesn't make you play any given species. Sure, after you've picked a species like, say, DE, your number of meaningful decisions is reduced, but isn't the act of picking a species a meaningful decision?

No, because if you treat species/background choice as a strategic decision, then for winning the game it is a no-brainer to pick DDEE (or some centaur or whatever) every time.
Even if species/backgrounds were balanced, which they clearly aren't (and I am routinely told, by developers, that this is intended and species/background choice isn't supposed to be a strategic decision; you can't have it both ways), the decision would be completely identical every time; your species/background choices aren't randomized like the dungeon is.

[archaeo]

I went ahead and split this off from the Salamander topic because it's starting to veer off into its own interesting discussion, one I felt like deserved its own thread.

Even if species/backgrounds were balanced, which they clearly aren't (and I am routinely told, by developers, that this is intended and species/background choice isn't supposed to be a strategic decision; you can't have it both ways), the decision would be completely identical every time; your species/background choices aren't randomized like the dungeon is.

I don't see "meaningful strategic decisions" in the design goals, though? You're obviously right, but I can't quite wrap my head around the idea that a lopsided species is inherently bad design. If you want a "balanced" species that demands you make decisions about what the dungeon has to offer, Crawl has that in spades; if you want a "lopsided" species with more directed gameplay (which, judging by !lg stats and whatnot, players often do), Crawl offers that too.

Out of curiosity, if "lopsidedness" makes skill aptitudes inherently bad design, what's the response? Do you just remove skill aptitudes altogether and focus on differentiating species based on stats, innate mutations, etc.?

[duvessa]

Well, if you're asking what I personally think, I think it would be fine to either remove aptitudes altogether, remove aptitudes other than HP altogether, or remove aptitudes other than HP and a "skill aptitude" that is a single number affecting the rate at which all skills are trained.
If you're asking what I personally think will happen in the near future regarding aptitudes, the answer is nothing.

I was just trying to explain why proposals for new species that are mainly differentiated by aptitudes - which are very common here - are unpopular.

[archaeo]

well, and a damn good job you did explaining, too!

I'm curious what other people think about species design. It would be enormously helpful if, instead of just shooting down bad ideas (which is like, 90% of GDD activity as far as I can tell), we could also have some threads on what good design looks like.

[Siegurt]

what I consider good race design:

One or two defining features (typically in the form of mutation(s) it might be a single mutation, or a group that act in concert)
Where said defining feature *changes how the game plays for a player of that race when compared against other races*
And also *doesn't pigeonhole the player into a single play style, that is to say it's still reasonable to play a range of styles*
And also *isn't substantially the same as the defining feature of another race* (that is to say, the same but with a different flavor)
Aptitudes can*compliment* racial design, rather than define it, again without pigeonholing the race, which may or may not mean mostly-flat aptitudes

This is a very challenging proposition, and IMHO why 95% of all racial designs *should* be discarded out of hand, that isn't to say they shouldn't be *proposed* only that they should be further taken apart and discarded.

Whys:
One or two rather than "a slew of vaguely related features" for several reasons:
1. too many competing features makes each one less distinct, and
2. it overly reduces the distinctiveness of races as a whole, if all races have 10 'special things' about them, none of them are special
3. it reduces the design space available to create not-overlapping races.
It should change the way the game plays because otherwise what's the point, if there's no difference other than 'flavor', it doesn't add anything.
It shouldn't pigeonhole characters because a reduction of strategic choices available makes the race itself less replayable.
It shouldn't duplicate the defining feature of another race because again, if there's no difference other than flavor, it doesn't add anything.

[WildSam]

How many already implemented races can be considered well designed?

[partial]

I see race choice as a difficulty slider for the class/play style you're looking to play.

Want an easy Wz or a hard Fi? Play a DE.
Want the opposite? Play an Mi.

It's part of crawl's meta more than the game itself.

[Sokengo]

How many already implemented races can be considered well designed?

In my personal experience Demonspawn, Deep Dwarf, Draconian, Formicid and Octopode have very unique mutations and play styles from other races and are very well designed, but what I really don't understand is the part of "Aptitudes alone are not enough to justify the creation of a new race" that many people seem to think that should be the norm when we have races like Hill Orc and Merfolk that the only difference that they have with normal Humans is their aptitudes and a very situational ability that don't really affect their play style at all.
I would also argue that Kobold, Halfling, Minotaur and Tengu are in the same boat, they are simply defined by their aptitudes and what equipment they are able to wear.

[Kaelii]

I don't think aptitudes are strictly bad.
I do think that we might have TOO many aptitudes, and they are easy crutches to try to make a species follow how the species plays different from other species in the designer's head. However emphasizing aptitudes often forces species to play one background rather than actually gives the species a different style of play for most/all backgrounds because of the differences inherent to the species.

Well designed species, but may not be perfectly designed:
Human -- a baseline may not be particularly interesting to experienced players, but it is helpful for newer players and is important as a check for other species. If they are strictly outclassed in all ways by another species that other species needs to be addressed.
Naga -- Slower movement radically changes how the player must approach the game. They are given some tools to help cope with that change. Even if they had entirely flat aptitudes they would feel very different from humans.
Centaur -- Faster movement radically changes how the player CAN approach the game. They are encouraged to explore different paths to deal with challenges, and given their own challenges. Even if they had entirely flat aptitudes they would feel very different from humans.
Felid -- As much as I hate playing them, I think they are a fairly well designed as a low defenses race. On the other hand, I think Spriggans actually fill a similar design space to them, only with vastly superior defenses.
Deep Elf -- Good at spell, bad at weapon (tm).
Minotaur/Hill Orc (One maybe) -- Good at weapon, bad at spell (tm).

Decent designs:
Demonspawn and Draconian (Possibly just one) -- Adapt to what the game gives you. Its fun!
Octopode -- They're adaptable, and have a different difficulty curve as the game progresses from most other species, however their complete dependence on an overabundance of one type of item (GOOD rings) generating is at times very awkward, and may tilt them unreasonably towards magic. (Although I think a lot of species are unreasonably tilted towards magic because magic offers more than just killdude)
Formicid -- Permastasis, free access to premeditated digging makes them have a much more tactical approach to combating enemies.
Deep Dwarf and Vinestalker -- Your health is a resource to manage and use.

Designed; maybe need to look at again.
Ghoul/Troll (One) -- Fight often or you're gonna have a bad time.
Demigod -- Gods are major parts of the game, foregoing a god choice is very interesting. Boosted stats however might not be the best way to do this. Human^Chei vs Demigod?
Gargoyle -- Less HP for a vast boost to AC, as well as several resistances. With fairly general aptitudes, I think they may need something to further alter their playstyle from human, unless we specifically want one or two races that are superior to humans.

I want a redesign:
Spriggan -- I feel felids cover almost all of the interesting space spriggans fill, other than HERBIVORE. Given all the noise I've seen about eventual hunger redesign, it might be time to look at a different way of approaching the IDEA of herbivore: a species limited by their exploring capabilities to survive rather than their combat capabilities (ghouls/trolls are opposite to this idea)

[amaril]

Avoiding no-brainers is a major design goal. The whole point of the skill aptitude system is to create no-brainers.

Although I agree that in most cases flat aptitudes give the player more compelling strategic decisions (whereas if I'm playing a Merfolk, I'm going to use Polearms almost certainly, and probably charms), I think that uneven aptitudes can add gameplay variety by encouraging the player to make 'undesirable' choices. For example, if no species had good Hexes/Short Blades/Stabbing aptitudes, I would almost never play a stabby character. However, when I am playing a Kobold, stabbing becomes far more appealing. The choice of [training hexes further to increase spellpower vs. investing xp in a totally untrained weapons skill] becomes a real strategic decision in the midgame. Stealth in particular feels like a meaningful aptitude because choosing fights (for me, at least) is one of the most difficult decisions in the game--stealth is good on pretty much any light armor/armorless character, but a species with a high stealth apt (eg. octopode) has the interesting early game choice of [train combat skills vs. train ability to avoid combat]. It is important that different playstyles are viable under different circumstances, because blasters/meleedudes/stabbers/etc. all impose unique tactical challenges.

Having written that, perhaps a better way to support varied playstyles would be to make them more even in power level (not that that's an easy thing to do). For example, 'optimal strategic play' on almost any flat-aptitude race is Trog>Heavy Armor Throwing notTrog>Medium Armor Charms>etc (the exception to this being armorless races). As is, because alternate playstyles are fun but suboptimal, it is (in my opinion) a good thing that there are species with imbalanced aptitudes (such as DE) that serve as 'difficulty-levels' for these playstyles. Ultimately, the 'fun' of crawl lies much more in the tactical decisions than in the strategic ones.

[milski]

Duvessa, you've ignored one of the major design goals that aptitudes are necessary for: Newbie support. If aptitudes are removed or simplified as you suggest, the only options are to massively increase or decrease the difficulty of the game for new/badplayers, because you neuter the newbie-choice races. New/bad players are in a much weaker position if Crawl lacks invincible minotaurs or deep elves reliably casting IMB at level 4.

As a less relevant aside, if you hold "no no-brainers" above all else, then the only favorable Crawl is one where everybody is a HuWa and attributes are removed. Otherwise there are varying degrees of pigeonholing due to racial mutations and starting skills.

[bel]

Oh someone already did this, nvm. Delete.

[duvessa]

Duvessa, you've ignored one of the major design goals that aptitudes are necessary for: Newbie support. If aptitudes are removed or simplified as you suggest, the only options are to massively increase or decrease the difficulty of the game for new/badplayers, because you neuter the newbie-choice races. New/bad players are in a much weaker position if Crawl lacks invincible minotaurs or deep elves reliably casting IMB at level 4.

This doesn't make any sense. You don't need skill aptitudes to make overpowered species. Centaur and troll are two of the easiest species in the game to win, and certainly moreso than the two you mentioned, but this has nothing to do with their skill aptitudes. Trolls have a total of one above-average skill aptitude (unarmed combat).
Even if that weren't the case, your argument doesn't provide any reason for different skills to have different aptitudes.

[tabstorm]

Races that should be removed for redundancy or being unfun due to a gimmick:

Ce
Fe
Ha/Ko pick one
HE
Vp
Te

Note: This isn't due to apts. Ce and Fe are basically redundant with Sp as fast-move races and no one likes Ce anyway. Ha and Ko don't really do anything interesting aside from have an EV bonus and less HP. HE is a bad version of DE or a bad version of Mf/HO depending on how you build, etc.

Also, I don't think anything is bad about apts, at least for things that aren't weapon schools. Lopsided weapon apts are more like a part of power level rather than a decision-forcer.

Not everyone wants to play "OMG I'm gonna adapt 2 the dungeon floor!!!", people sometimes know they want to play X character archetype or challenge before they start the game and choose their race/bg accordingly.

[Sprucery]

My personal opinion is that the species in Crawl are currently quite ok. Every species plays differently enough to me. Even Ko and Ha. Then again, I guess I play the species quite differently than most people: aptitudes affect quite much how I train skills.

What I would change: DD (take away the wand recharge ability, or make it take 10 MP), Fe (take wands away again), Vp (give an ability to switch to bloodless near-instantly instead of having to wait).

I really don't understand the arguments that Ce/Sp/Fe are too similar just because they're fast.

[duvessa]

Movement speed is really, really important, to the point that it makes a species' other properties seem irrelevant. Nagas have weird defenses and poison spit but it's barely noticeable in the face of the movement speed penalty.
Fe is also pretty well summarized by "worse Sp but with extra lives", honestly.

[tedric]

I don't think that it's inherently a problem to use aptitudes to differentiate species. But I do think that pretty much all of the interesting possibilities for this already exist, because you only get a race with varied gameplay possibilities when you group like apts together.

Mi pushes you firmly toward weapons instead of spells (and Ce pushes you firmly towards ranged weapons); DE pushes you firmly toward spells instead of weapons. Lots of species (Ko, Ha, Vp, Sp, Fe, etc.) strongly suggest stealthy gameplay with good apts for Stealth/SBl/Hexes plus small/tiny size or other mutations; but since stabbing alone can't carry the game in quite the same way that tabbing or a broad spellset can, it makes sense that there are a variety of approaches here -- each of which admits a variety of less-stabby playstyles as well, and differentiates itself from the other stealth-suggestive species in significant ways (with the probable exception of Ko/Ha).

When "aptitude differentiation" gets much more specific than those three broad categories, you start to pigeonhole the player into specific choices that are so optimal, they basically cease to be choices (see: Mf's Polearms +4). That's why new race proposals that focus on apts generally fail.

Innate mutations offer better possibilities for differentiation between races, and for introducing new gameplay mechanics that have interesting effects across a variety of playstyles. But they can also contribute to pigeonholing, especially when combined with highly specific apts (see: Tr's UC +0 combined with terrible weapon apts, even worse spell apts, and Claws 3; Og's M&F +3 combined with ability to wield GSC). Most new race proposals that focus on innate mutations seem to fail because they are tailored to too specific a playstyle.

All of which tells me that the best way to design a species is to either give it flat apts (whether good-flat or poor-flat), or to skew its apts toward weapons/spells/stealth via broad groupings, and then do further differentiation through mutations that don't pigeonhole.

We've already got several excellent species that use innate mutations on top of flat apts (Hu, Dg, Ds, Dr, Op; to a lesser extent Mu, Na, DD, VS, etc.) to produce well-differentiated species that are each capable of a variety of playstyles; more of these will always be viable. As mentioned, the stealth-suggestive species is already a pretty large and differentiated group. I think there is probably room for a few more species with innate mutations that interact interestingly with the weapon/spell split that Mi(/Ce) and DE are currently filling -- mutations that change your relationship to weapons in general, mutations that change your relationship to melee positioning in general, mutations that change your relationship to spells/MP in general, etc.

[tasonir]

I vastly prefer Ce to Sp, but I'm not really sure it's a game design worthy discussion. Ce is a far more powerful race imho, and I like winning.

On topic: I agree with generally flat aptitudes being a good thing, but I thought I'd mention one use of non-flat aptitudes that isn't mentioned much: using required skills as EXP sinks. You don't want the race as a whole to have -3's across the board, you want them to be able to pick up weapon skills and magic easily, but they're too strong with 0's everywhere. So you toss -3 into armor, dodging, shields, and maybe even give them a bad fighting aptitude. Now they can get their early killdudes going, while having lower defenses and being "forced" into training low aptitude skills. But they only have to train them a bit later on - it doesn't make getting to the temple any harder. Nagas and Centaurs are the general examples of this: Centaurs get -3 but have some good aptitudes like +3 bows, and Nagas are a bit more flat except for their -2 defenses and good poison and stealth. Imho it's a valid use of having different aptitudes.

That being said, I do like Duvessa's idea of just having one skill aptitude for the entire race. I wouldn't mind an alternative which was broken down into categories: A weapons aptitude (includes fighting/unarmed), defenses aptitude, magic aptitude, general aptitude (stealth invo evocations). With only four it'd be pretty easy to memorize the aptitudes for at least a few of your favorite races, and I wouldn't be asking myself if nagas had the -3 defenses or centaurs did (I was wrong when I checked the table).

[WingedEspeon]

HE
HE is a bad version of DE or a bad version of Mf/HO depending on how you build, etc.

I will argue that DE is a bad version of HE.

On another note, I think that uneven apts can in some cases create interesting choices. Sometimes you will find a really good weapon in a class that is not favored by your apts. Sometime you will end up with high Str, no spells, but a terrible armour apt and a great dodging apt. (MfBe, for example). sometimes the best spells you are offered are in schools you don't have good apts for. If you had entirely flat apts, these situations would usually be no-brainers.

Also, to completely eliminate strategic no-brainers you need all strategic choices to be equally good which is not only impossible, but makes strategic choices completely irrelevant.