complete revision of stats

[Derakon]

The problems that you point out with CON and HP are just as bad, if not worse for SP.

SP is actually significantly better -- the curve goes linear at 18/70 (and tops out at 18/180). It is pretty poor until you get to that point, though not worse than CON.

But more importantly, everyone needs as big a CON score as they can possibly manage. Only mages and priests need large mana pools.

[fizzix]

Thanks for the vote of confidence, Mikko.

Despite all the discussion, I do think that Vanilla's stats work pretty well as it stands. Or more importantly, the game is fairly well balanced with the stats working the way they do. For example, if you made all spellcasters need both INT and WIS then suddenly mages and priests have been badly nerfed, and they already have the hardest starts (though priest levels out sooner than mage does). And of course most of the races are biased towards only one of the two stats, so there would be no good races for your chosen class...

Agreed. Changing spellcasters to need both INT/WIS or INT/CHR or whatever is a bit of a problem if stats were as they are. However, if you *didn't* need to invest any points in STR to carry around typical mage gear, or any points in DEX to actually be able to wield a weapon, things look different. Basically, the early game would have to be rebalanced so that a character with all stats equal has a reasonable trek to stat gain regardless of what class.

The problems that you point out with CON and HP are just as bad, if not worse for SP.

[CunningGabe]

At the moment stealth is not a skill, and I don't think perception/searching/disarming are either (though I think CunningGabe's trap changes may have introduced a clev factor there).

You are 1/3 right Searching and Disarming both increase as you level in V; approximately 1% per level (but this varies based on race and class). Perception -- which controls search frequency in V and search radius in v4 -- doesn't increase.

Also, rogues do get better stealth as they level. (These are all governed by the X lines in p_class.txt.)

[Mikko Lehtinen]

Ooh, nifty... those are some ginormous irregular rooms! And what's the purple zero surrounded by a bunch of #'s and ='s? An altar perhaps?

The basic room shape (and the placement of many terrain features) comes from Un. In my new dungeon generation several rooms may join together.

It's an altar of secrets on a stone platform. ='s are steps.

You say it's a new setting? So all the Tolkien monsters and artifacts are going away, as in ToME? Hmmm...

Look here:

http://www.mikkolehtinen.net/mist/?p=79

[Derakon]

Thanks for the vote of confidence, Mikko.

Despite all the discussion, I do think that Vanilla's stats work pretty well as it stands. Or more importantly, the game is fairly well balanced with the stats working the way they do. For example, if you made all spellcasters need both INT and WIS then suddenly mages and priests have been badly nerfed, and they already have the hardest starts (though priest levels out sooner than mage does). And of course most of the races are biased towards only one of the two stats, so there would be no good races for your chosen class...

Personally my biggest bugbear with stats are the HP bonuses from CON, and specifically the way they accelerate as your CON increases:

Point range   average HP gain per point at level 50
10-18         9.06
18-18/80      15.63
18/80-18/160  25
18/160-18/200 56.25

The way that CON is basically entirely irrelevant in the early game / mid range of stats, and then each point becomes more valuable than the last one until you max out, is pretty weird-looking. I think the same thing happens with a number of other stats, but CON is the most evident.

Of course, changing this is easy but would unbalance the game -- if you flatten the curve out then suddenly you have more hitpoints in the early game, or characters who would never have gotten their CON maxed will have more hitpoints than they "should".

[Magnate]

Rebooting stats so that they can be used for shaping different character concepts would require radical changes to the game, and I don't think people want Vanilla Angband to change that much.

This is why v4 was invented. You're quite right that this kind of radical rethinking would be inappropriate for V, at least until it's been thoroughly tested and refined in v4. If anyone wants to step up and code up their ideas for stats and push a v4-newstats branch to github, I'd happily take a look and merge it if it worked. This is how v4 got a completely new combat system, and could work equally well for stats.

Personally I think the D&D heritage isn't really important, though I'm not averse to stats being re-named. I think we need a careful think about what are "skills" in Angband - things that characters learn and get better at as they level up - and what are just quantities arising from race/class/stats/equipment that change as those things change but not with level. At the moment stealth is not a skill, and I don't think perception/searching/disarming are either (though I think CunningGabe's trap changes may have introduced a clev factor there). Melee and missile combat are skills, as is spellcasting (failure rates reduce by 3% per clev in addition to reducing with increasing spell stat).

Until we've got some clarity on what ought to be level-dependent and what not, I don't think we'll come up with a coherent set of impacts for stats. But I could be wrong, and I don't want to stop anyone having a go.

[ekolis]

Ooh, nifty... those are some ginormous irregular rooms! And what's the purple zero surrounded by a bunch of #'s and ='s? An altar perhaps?

You say it's a new setting? So all the Tolkien monsters and artifacts are going away, as in ToME? Hmmm...

[Mikko Lehtinen]

Hey Mikko, where is this magical new version of Fay that has all these features I've never heard of before? Or did I just never get far enough in the existing version to see any of them? For reference, the version I have right now is 1.11

Yeah, I know... I'm so excited about my development version that I have to talk about it all the time, and then people get confused. Sorry.

The game is completely playable. At the moment I'm having my own personal playtesting gauntlet. I'm playing 10 random characters slowly and carefully to the bitter end. If five of them get to DL 20 the game balance is good enough for public beta. (DL 48 is the new bottom.) I've done several of these gauntlets already, failed to get far enough, and made the game easier...

The other problem is documentation. For that, I'm very soon starting a new blog for the game. I'm going to write some important info about the new features so that people can play the public beta. Updating the help files will probably take much longer.

At the moment you can only find some randomly generated dungeons there. Feel free to ask questions about them in the blog if you like:

http://www.mikkolehtinen.net/mist/

The game is now set in a brand new world. I'm changing the name to Halls of Mist.

[ekolis]

Hey Mikko, where is this magical new version of Fay that has all these features I've never heard of before? Or did I just never get far enough in the existing version to see any of them? For reference, the version I have right now is 1.11

[PowerWyrm]

This change would make races with low INT completely unsuitable for clerics no?

[Mikko Lehtinen]

It may well be that Derakon's idea is the right choice for Vanilla!

Rebooting stats so that they can be used for shaping different character concepts would require radical changes to the game, and I don't think people want Vanilla Angband to change that much. For example, my idea doesn't really fly if every character eventually maximizes all stats.

But the benefits are huge, too. Here's an example of the various possible warrior or rogue builds in FayAngband:

You probably want to take a high score in either STR (heavy weapons) or DEX (light weapons). You may want to take high scores in both stats: it makes you more flexible in the choice of weapons, but doesn't really boost your combat skills that much. Actually taking just average scores in both STR and DEX is okay, too -- you'll do just fine if you find a good medium weight weapon.

All the other stats are quite equal in importance. Put your ability points where you want!

• INT for Mapping and Alchemy, both really useful skills. It also affects Lore, which lets you identify objects.
• WIS for Perception and Saving Throw, also essential skills. It also affects Lore.
• CHR for Magic Device skill, wand (and spell) range, and cheaper prices.
• CON for hitpoints.

It's almost as if you were playing Sangband! Even better, in this stat-based system your equipment choices become more interesting.

For spellcasters the only thing that changes is that you probably want a high score in your spell stat. That's a no-brainer in Fay. Otherwise the choices are pretty much the same as for non-spellcasters.

As Derakon suggested, I have considered renaming my mental stats. Perhaps Logic, Awareness, and Aura?

[Mikko Lehtinen]

STR: prowess, carrying capacity
DEX: finesse, hit chance
CON: hitpoints (already the God-King of stats; no need to put more stuff here to make it more powerful)
INT: spell failure rate, number of learnable spells
WIS: mana pool, saving throw

Even though I have totally different ideas, I like this too. This approach doesn't allow as much flexibility when choosing your own "build" for your character, but it's simple and nice!

[Mikko Lehtinen]

Like Derakon, I favor transparency and simplicity of impact very strongly. My approach in Fay has been to:

1) Invent a couple of new skills, with one of the stats having a very strong effect on them.
2) Stats as such don't have many effects. For example, Wisdom governs just Saving Throw, Perception, and priestly magic. (INT+WIS together give you Lore points.)
3) Remove unimportant skills to make room on the character sheet for new, important skills. For example, I joined Disarming into Perception.
4) Invent lots of scenarios in the dungeon where the skills are really important! Usually it's a pure 1d100 under skill score, and often you can try only once.

For example, Perception tests are made when you:

• search for traps
• disarm nonmagical traps
• pick locks (only one try! you can close and lock the doors again, which is very helpful for escaping)
• search for warding runes in walls
• search for hidden items in closets
• search for hidden torches, mushrooms, or faery portals in "interesting vegetation" grids
• identify the effects of fountains, which can be very beneficial or harmful. "You smell disease."
• try to notice an invisible monster next to you
• *Identify* non-potion items with your Lore proficiency

When your starting warrior has a high Wisdom, you will really notice it while adventuring in the dungeon, and feel good about your choice! Similarly, low-Wis characters will suffer for it.

I disagree with Derakon that class should be made more important. Your class already governs almost everything that your character can do, including the most important things, spellcasting and fighting. Since I wanted to make stats and races really matter, I made the class influence on "secondary" skills much less pronounced.

[fizzix]

LOL! I'm still trying to wrap my mind around how that would work, too. Perhaps fizzix has a perfectly good explanation.

Well, it's really no more than an extension of Mikko's idea of aura. For simplicity, let's assume a 1v1 battle. Both @ and the monster have auras. These auras fall off with distance, and the fall off (or initial strength) is determined by player CHR and monster level. If the player is really strong and the monster is weak, the player's aura completely dominates the entire region and the monster cannot cast spells anywhere. Similarly, if the monster's aura is strong and the player's is weak then the monster can't cast spells anywhere. This is the basic idea, but it doesn't really work because there are hundreds of monsters, all with auras, and one @.

This made me simplify things so that if the situation would be that @'s aura is dominant, then the monster has a reduction of range. If the monster's aura is dominant, then it may be able to cast spells far enough away that @ cannot, so it has an edge. It's not perfect, but it's easier than trying to extrapolate the 1v1 approach to multiple monsters and not make a muddled mess.

and Nick, don't worry, I'm not actually planning on doing this, I certainly won't have the time to. I just found it a lot of fun to think about.

[Nick]

There is another reason I am loathe to git rid of charisma. If we were to get rid of it, then any future variant that wanted it would have to add it back in. I suspect that's a good deal harder than making a variants remove charisma or ignore it if they don't want it. So we might as well keep it, and in the meanwhile, why not try _something_ with it?

Agreed. For example, FAangband has some monsters which can be either hostile to the player or neutral; I plan to have CHA influencing that.

On the whole, while it's great to have the discussion, I'm in favour of incremental changes to the stat system.