Discussion of Gnolls 2.0 commit

[tasonir]

So here's the commit which is linked in the .20 tournament post about a possible Gnoll rework: https://github.com/crawl/crawl/pull/532 ... c6fad56192

To be explicit upfront: I think this is a bad idea, however, I fully support merging it to trunk and having players actually test it out. Experiments are a great thing.

That being said, I don't see how it's possible to hybridize a character with 5 int. Strength and dexterity aren't really *required* for melee effectiveness/dodging, they are merely *helpful*. But you can't really cast spells with 5 int - you'll never reach any significant spellpower, even on the low level spells a 'hybrid' with 'diverse' magic skills should excell at, and your spell hunger will be essentially unmodified. Unless they become carnivores or otherwise ignore spell hunger, playing a significantly magic-using character is going to be largely impossible. They'll be melee and ranged weapons, with possibly some low level magic support (aka, regeneration, blink, etc). It'd probably be very easy to play a crossbow/longbow hunter, with some early game melee until you collect the ammo. I don't see 5 int gnolls ever going significantly into magic, which means they'll learn roughly 6 skills: Fighting, armor (limited by your armor choices with 5 strength), dodging, a melee weapon, a ranged weapon, and evocations. Then possibly invocations depending on god, and possibly shields if you want to use slings as a ranged weapon, although this is probably worse than bows/crossbows. Sure they'll have non-zero magic skills, but I don't see those skills making much of an impact at 5 int.

Am I wrong? Has anyone tried them in an experimental branch yet? Do you think it's an improvement over the current version? For what it's worth, I think the current version works pretty well - they are absolutely weaker than most races and the late game is much tougher, but assuming it's meant to be a difficult race, then that's fine. They also succeed in making the early game easier, which is interesting.

[Vajrapani]

I don't like it, at least on paper anyway. Consider Minotaurs: they typically will only train a combination of fighting,armour,dodging and weapon skill throughout a game. Every single one of those aptitudes is +1 or higher. Minotaurs also have the added benefit of not being capped at 5/5/5 stats(or 7/7/7 judging by a recent commit) so I don't see why Gnoll is just made into a bargain basement Minotaur.

+3 aptitude's across the board might be nice,but other races will always outclass whatever role the gnoll is trying to play. For example if you want to play an axe user of Uskayaw, you'll typically want a species with high axe and invocation aptitudes, so you end up with a choice between Gnoll & HO. They both have +3 in Axes and Invo but you will always go with HO every single time(if your goal is to win) considering the fact that HO's will have much more options available to them by virtue of their bigger stats and thus a bigger toolbox to deal with threats.

I quite enjoy Gnolls as they are, and I honestly doubt this will a change for the better. Instead of being a generalist, this gnoll is more of a cheap knock-off of all the races at once.

[duvessa]

That being said, I don't see how it's possible to hybridize a character with 5 int. Strength and dexterity aren't really *required* for melee effectiveness/dodging, they are merely *helpful*. But you can't really cast spells with 5 int - you'll never reach any significant spellpower, even on the low level spells a 'hybrid' with 'diverse' magic skills should excell at, and your spell hunger will be essentially unmodified. Unless they become carnivores or otherwise ignore spell hunger, playing a significantly magic-using character is going to be largely impossible.

The high aptitudes are intended to offset the spell failure problems arising from low int (I know +3 isn't necessarily enough to do that, I was not the person who chose that number). Newgnolls having low spellpower is intentional, that's kind of the essence of the "jack of all trades, master of none" gimmick. There are tons of spells that are great at low spellpower: apportation, conjure flame, invisibility, freezing cloud, mephitic cloud, sticky flame, the good form spells, most of the summoning and necromancy schools...
Spell hunger is meaningless and may be on the way out anyway.

5 str means you do a little over 50% as much melee or missile damage as a character with 20 str...before the effect of AC.
5 int means you have 25% as much pre-stepdown spellpower as a character with 20 int. This means that if a character with 20 int has 50 spellpower, you have 12.5...but if they have 100 spellpower, you have 37.5, and if they have 150, you have 73. Most spells also have significant constant factors that mean reducing spell power by 50% won't reduce conjuration damage by 50%.

If 5/5/5 stats is unbalanced, they can just, y'know, be changed to 7/7/7 or 5/10/5 or whatever. Like how the gnoll aptitude curve has changed about 5 times.

+3 aptitude's across the board might be nice,but other races will always outclass whatever role the gnoll is trying to play. For example if you want to play an axe user of Uskayaw, you'll typically want a species with high axe and invocation aptitudes, so you end up with a choice between Gnoll & HO. They both have +3 in Axes and Invo but you will always go with HO every single time(if your goal is to win) considering the fact that HO's will have much more options available to them by virtue of their bigger stats and thus a bigger toolbox to deal with threats.

Yes, that is the price you pay for being a generalist: not being as good in a specialty as a specialist.

I quite enjoy Gnolls as they are, and I honestly doubt this will a change for the better. Instead of being a generalist, this gnoll is more of a cheap knock-off of all the races at once.

Gnolls as they are aren't generalists. They're humans with a free early game and a slight difficulty increase later. The point of this change is to make them more generalist.

[NhorianScum]

Low stats high aptitudes seems like enough of a niche to differentiate knolls. Making the stats completely un-modifiable just seems like a not even remotely fun thing to deal with.

[Floodkiller]

Am I wrong? Has anyone tried them in an experimental branch yet? Do you think it's an improvement over the current version? For what it's worth, I think the current version works pretty well - they are absolutely weaker than most races and the late game is much tougher, but assuming it's meant to be a difficult race, then that's fine. They also succeed in making the early game easier, which is interesting.

At this point it doesn't matter about the current version, because the dev team's decision about current Gnoll is a no. Regardless of whether I feel those criticisms are valid (for example, I still think the skill system itself is at the same level of micromanagement as Gnolls), it is a decision that needs to be accepted by either pushing forward with different versions or stopping development (leaving it to either have someone else pick it up or get removed). Since the species is still in trunk (only left out of stable), I'm opting for Option A to take advantage of the wider audience trunk provides to find something that sticks. I started with one of Duvessa's ideas because it was simple to understand and easy to implement. It also still fit the idea of what a 'gnoll' is to me. If this approach doesn't work, I'm willing to keep working. After all, there's plenty of time before 0.21 to get something both acceptable and fun/unique.

I don't like it, at least on paper anyway. Consider Minotaurs: they typically will only train a combination of fighting,armour,dodging and weapon skill throughout a game. Every single one of those aptitudes is +1 or higher. Minotaurs also have the added benefit of not being capped at 5/5/5 stats(or 7/7/7 judging by a recent commit) so I don't see why Gnoll is just made into a bargain basement Minotaur.

+3 aptitude's across the board might be nice,but other races will always outclass whatever role the gnoll is trying to play. For example if you want to play an axe user of Uskayaw, you'll typically want a species with high axe and invocation aptitudes, so you end up with a choice between Gnoll & HO. They both have +3 in Axes and Invo but you will always go with HO every single time(if your goal is to win) considering the fact that HO's will have much more options available to them by virtue of their bigger stats and thus a bigger toolbox to deal with threats.

I quite enjoy Gnolls as they are, and I honestly doubt this will a change for the better. Instead of being a generalist, this gnoll is more of a cheap knock-off of all the races at once.

These same arguments have already been made about previous versions of Gnoll and will probably continue to get made unless the direction changes massively. If the species is better than minotaurs at melee and better than deep elves at spells, what do you take away to balance that? If you don't, what's the point of having any other species in the game if Gnolls are better than them at it? As for the aptitude and stat numbers on this particular implementation, they are mainly placeholders at the moment (see the adjustment of 5/5/5 to 7/7/7 after some quick playtesting while ensuring the implementation was free of bugs) and can easily be adjusted after seeing more data.

Low stats high aptitudes seems like enough of a niche to differentiate knolls. Making the stats completely un-modifiable just seems like a not even remotely fun thing to deal with.

If it's just low stats without a lock, the differentiation goes away the moment the player worships Zin/Chei or finds enough stat boosts on artifacts, turning them into a species that is just straight up good/better at everything (see above). I also can't just ban those specific instances and leave others available, because then the player can be confused as to what is kosher and what isn't. Locking down all stat alterations (positive and negative) is easier to understand, as well as doubling as another unique feature (immune to stat drain).

[Sprucery]

If it's just low stats without a lock, the differentiation goes away the moment the player worships Zin/Chei or finds enough stat boosts on artifacts, turning them into a species that is just straight up good/better at everything (see above). I also can't just ban those specific instances and leave others available, because then the player can be confused as to what is kosher and what isn't. Locking down all stat alterations (positive and negative) is easier to understand, as well as doubling as another unique feature (immune to stat drain).

I think it would be perfectly understandable that the only way to increase stats would be on level-ups. The starting stats could be lowered to 4/4/4 and then the normal choice to raise a stat every 3 levels* (no random raises). This would give the player some more control on how to develop the character based on what armour/spells etc. are found.

*or 4, or 5, whatever would be best for balance

[VeryAngryFelid]

If it's just low stats without a lock, the differentiation goes away the moment the player worships Zin/Chei or finds enough stat boosts on artifacts, turning them into a species that is just straight up good/better at everything (see above).

Why do you stop here? You can make them unable to wield branded weapons, weapons with positive enchantments, armour with positive enchantment, spell power enhancers, wizardry items, rings with AC/EV/rF+/MR etc.
In case it is not clear, I don't understand why you did that, I don't think Zin or Chei are OP and even if they are, then they should not be changed just for Gnolls, they should for every species.
Big part of having fun in crawl is to find and check loot, I imagine it can be very frustrating to waste so many good items.

[Floodkiller]

I think it would be perfectly understandable that the only way to increase stats would be on level-ups. The starting stats could be lowered to 4/4/4 and then the normal choice to raise a stat every 3 levels* (no random raises). This would give the player some more control on how to develop the character based on what armour/spells etc. are found.

*or 4, or 5, whatever would be best for balance

I don't think I'll lower the starting stats any further than they currently are (7/7/7) because it makes the start of the game too frustrating for certain backgrounds. However, I think re-enabling the player based stat growth could be okay to do (disabling it on Gnolls and nothing else may be equally as confusing as leaving it enabled for a species with 'unchanging attributes').

[4Hooves2Appendages]

I don't see the benefit of banning stats from items. It's not like they are guaranteed. It will also have almost no effect on the early game. Later it will be a trade-off in all likelihood. Why not offer people choices?

[watertreatmentRL]

A deeper dive into the influence of stats and aptitudes on gameplay is the last thing crawl needs. "High skill compensated by low stats" is a bad concept.

The earlier versions of this gnoll concept were more compelling at least in the direction they went: Stronger early game with potential for real late game limitations. From what I remember, that potential was never realized. Too trigger shy on max skill levels, produces characters that are even more homogeneous than normally occur in crawl.

To make this concept good, you have to break out of standard mechanics and do something actually different. I still say the starting skills + experience level-based straight bonus to all skills with a low cap like 12 and normal stats are going to be your best bet.

[Floodkiller]

I don't see the benefit of banning stats from items. It's not like they are guaranteed. It will also have almost no effect on the early game. Later it will be a trade-off in all likelihood. Why not offer people choices?

It's based on the feedback from what I see in most current Gnoll games. Current Gnolls change the skill system, but don't limit options that modify skills. In response, the majority of games and wins for Gnolls worship Ash, Oka, or Chei. What I infer from this is that opening these options to the player actually limits overall choice, since most players would prefer to shore up weaknesses to use a playstyle comfortable to them over trying to work with their weaknesses and excel elsewhere. You can see the same issue with Deep Dwarves: even though there are no restrictions on god options, the player feels like it is almost mandatory to choose a god that provides divine healing to fix the weakness of no natural healing.

Banning the effect in this case slightly restricts options (items that only provide an attribute change and nothing else become useless), but considerably improves player choices. Since they cannot modify their stats, they do not feel obligated to 'fix' them. Instead, they are able to work with them and make decisions around that weakness. The player can still worship Chei and Zin for their other abilities if they desire, and can still wear artifacts even if the stat boosting parts don't apply. Additionally, some options previously unavailable can be considered: an artifact that would normally be dropped immediately due to ridiculous stat drops like -9 Str/Int can now be considered as a usable item.

A deeper dive into the influence of stats and aptitudes on gameplay is the last thing crawl needs. "High skill compensated by low stats" is a bad concept.

The earlier versions of this gnoll concept were more compelling at least in the direction they went: Stronger early game with potential for real late game limitations. From what I remember, that potential was never realized. Too trigger shy on max skill levels, produces characters that are even more homogeneous than normally occur in crawl.

To make this concept good, you have to break out of standard mechanics and do something actually different. I still say the starting skills + experience level-based straight bonus to all skills with a low cap like 12 and normal stats are going to be your best bet.

I already explored a hard cap version and rejected it. The player felt obligated to backtrack and grab useless spells and items in order to avoid losing XP, even if they never actually planned on using them. Players also had every non-overlapping skill trained to max cap by the end of three runes, and this was with +2 apts and a cap of 14. I may try an XL only based skill level in a future revision if this approach doesn't work as expected, but I find the idea pretty unfun at first glance.

[watertreatmentRL]

Yes, backtracking is bad, but that is a deep structural issue in crawl that is affected only marginally by this skill cap business. Inventory based skill allocation restrictions are a much less deep issue, but still basically a crawl problem not a gnoll problem. (Worth noting here that the straight xl-based bonus without skill point growth after character creation eliminates this inventory issue.)

What is unfun is constantly entering the skill menu to modulate your experience allocation when you know that eventually your skills will all look the same. If what you have is something that differs from humans primarily in apts and stats, that just isn't a new thing, it's one more option on the character creation screen.

[tasonir]

At this point it doesn't matter about the current version, because the dev team's decision about current Gnoll is a no. Regardless of whether I feel those criticisms are valid (for example, I still think the skill system itself is at the same level of micromanagement as Gnolls), it is a decision that needs to be accepted by either pushing forward with different versions or stopping development (leaving it to either have someone else pick it up or get removed). Since the species is still in trunk (only left out of stable), I'm opting for Option A to take advantage of the wider audience trunk provides to find something that sticks. I started with one of Duvessa's ideas because it was simple to understand and easy to implement. It also still fit the idea of what a 'gnoll' is to me. If this approach doesn't work, I'm willing to keep working. After all, there's plenty of time before 0.21 to get something both acceptable and fun/unique.

Agreed, and I hope I didn't sound too negative - I included the disclaimer about wanting to see the experiment done because testing these sorts of things is exactly what trunk is for, and I'll certainly give new gnolls a shot (after the tournament ends, of course).

Really my main complaint about low stats is not a gnoll issue as much as it is the balance between the stats - casters are much more heavily dependant on int than melee, or especially ranged characters are. I feel like gnoll 2.0 is best played (theoretically) as a sling hunter, using early stones to kill everything at range, until you find a centaur, pick up a bow, and start using bows. Later on when you find a crossbow and lots of bolts in vaults, use that as well, it's going to be pretty cheap to get all your ranged weapons to min delay, and using all three is plenty of ammo. Alternatively just start with a melee weapon instead of slings and transition into bows/crossbows later on.

I'd probably never learn much magic at all, although some low level support spells would be fine. In other words, it just sounds like it's going to be a standard fighter with slightly more emphasis on ranged weapons, since those don't really require much in the way of stats to use well. You might be forced into somewhat lighter armor with only 5 strength, but having ranged attacks will help with that.

If you've raised the stats to 7/7/7, maybe something like 5/5/5 but let them pick one stat on every 4 levels (6 stats total) would end up with the same 21 stat total, but allow for them to optionally take one stat all the way to 10?

As much as it pains me to lose most of chei's purpose, I agree that you basically have to block chei's stats or else the race just worships chei all the time. There is an alternative of instead reducing chei/zin's stat gains to something like 1/3, so in the base 7/7/7 version, you would instead reach 12/12/12 at max chei piety, but I'm fine with just blocking all of the stats if you think that's more simple. And honestly, as a frequent chei player, I don't think even I would really want to deal with slow movement for only +5 in stats.

[tasonir]

Also while duvessa said spell hunger might be on the way out, if it still remains for long enough that gnolls are added first, it should be fairly simple to give them a mutation which cuts spell hunger in half (and then subtracts their int * spellcasting from that half) to help them use semi-high level magic (ie, level 4-6) while still not being suitable for 7+ magic.

[VeryAngryFelid]

That's how crawl gods work, there are always better gods and worse gods for every combo. If you take out Ash, Oka and Chei, people will start use mostly Makhleb, Kiku or Fedhas. I am surprised we are going deeper into "Make some gods almost useless" paradigm when we had so many hot discussions about undead characters with 3 good gods, non-HO with Beogh and new gods required to support multiple combos. Anyway, my post seems to be ignored so I don't think it makes much sense for me to continue posting ITT especially provided the species looks so boring that I have never even tried it.

[4Hooves2Appendages]

I don't see the benefit of banning stats from items. It's not like they are guaranteed. It will also have almost no effect on the early game. Later it will be a trade-off in all likelihood. Why not offer people choices?

It's based on the feedback from what I see in most current Gnoll games. Current Gnolls change the skill system, but don't limit options that modify skills. In response, the majority of games and wins for Gnolls worship Ash, Oka, or Chei. What I infer from this is that opening these options to the player actually limits overall choice, since most players would prefer to shore up weaknesses to use a playstyle comfortable to them over trying to work with their weaknesses and excel elsewhere. You can see the same issue with Deep Dwarves: even though there are no restrictions on god options, the player feels like it is almost mandatory to choose a god that provides divine healing to fix the weakness of no natural healing.

While I accept that many DD players choose Makh for obvious reasons, I don't think the argument applies to Gnolls. In my view Ash and Chei are clearly weaker than Fed and Kiku, Gnoll or no Gnoll.

So if players choose Ash/Chei over Fed/Kiku with the genuine expectation that it will boost their victory chances, then I view that as a mistake. There are of course many other reasons not choose Fed/Kiku every game. I don't enjoy herding mushrooms/zombies for example...

[NhorianScum]

Chei/ask/oka are strictly worse at winning the game than Srank gods on old gnoll and newgnolls +3 all with bad stats just makes them favor optimal gods even more.

These gods are boring as fuck though so most folks took oka/ash/chei to roleplay the intended hybrid build.

[tasonir]

So heroism gets around the 14 skill cap, and ash also boosts skills, but how is chei related to the skill system at all? Doesn't seem like chei is getting around gnoll's weakness, it's just that chei is an overall excellent god who everyone should worship all the time (I kid, I kid). If anything I'd say chei can be a bit dangerous as most chei gnolls are going to be transmuters with unarmed, and unarmed is dangerous when you can't get it over 14 skill due to having a higher delay than it normally would, where as you could have gotten a demon weapon to min delay. That's a bit indirect of course, but it was certainly an issue for me.

If the claim is just that having high stats lets be stronger with lower level skills, that's kind of a "well duh" concept to me. Stats are good, and they should be.