Combat redesign - taking a step back

[jens]

Combat redesign is overdue. I'd like to get some significant changes done in time for 3.4. Since combat changes need much discussion, and much play-testing, we need to start with it now. There are probably lots of threads about combat changes lying about, but I want to start by taking a step back. Since combat redesign is such a large topic, and probably filled with differing views, it would be helpful if we could first get to the same page. I'd like this discussion to be about what we want from a combat system, rather than how to change our current system. Once we have a consensus on what we want, it will be much more straightforward to actually design it.

My basic assumption in this post is that we already have a nice set of 'ingredients' for our combat system, we just need to change the recipe a bit. What I mean with 'ingredients' are things like “we have AC, and a higher AC makes the bearer take less damage”, or “additional blows can be achieved as a character develops”. These are things that most players would never consider removing, since they are so ingrained in what Angband is. This is not to say that nothing new can be added, but please wait with those suggestions for the moment.

I'll make a set of statements, questions and answers concerning what I feel is relevant to get combat design right. Please help me flesh this out, challenge any and all points were I've missed some issue, or got something totally wrong. I know it's hard to predict the life of forum threads, but please avoid concrete suggestions for changes. Combat is too large a topic to contain in one thread, so any such suggestions will bring chaos ;-)

Enough preamble :-)


We currently have 4 combat modes: melee, spells, ranged, and throwing. Should we have parity between them?
- No: melee > spells, ranged >> thrown
- All classes/races should be able to melee successfully.
- Some classes/races should be able to use spells as the main combat mode.
- Some classes/races should be able to use ranged as the main combat mode.
- In close combat (i.e. when standing next to a monster) melee should be the most effective combat mode for all classes.
- Throwing is mostly for flavour.


Increasing combat abilities is achieved through character level, stats, and equipment. To keep things clear I include stats gained through equipment in the category stats, and not in equipment. Should we have parity between them?
- Yes. I don't feel strongly either way, but the more parity we have, the easier it will be to balance the growth of combat abilities.


The growth curve of player combat abilities should match the growth curve of dungeon difficulty. That is, if you keep a steady pace in your dungeon exploration, you should, on average, have the same level of challenge throughout the dungeon.
- Yes. (This statement includes a parameter that does not really belong to a combat system, but it is useful as a reminder: any system we design needs to include plans to achieve parity with dungeon development.)


The growth curve of player combat abilities should be stable. That is, no matter what happens in your game, you should quite steadily increase in combat abilities.
- No: though in a long series of games the average power level should follow a smooth curve, in any given game big variations from this curve can, and should happen.


Should all combat modes have the same characteristics when it comes to development?
- Yes.


And remember, try to keep responses to general combat issues I have missed, and discussions to lead to a consensus on what we want.

[fizzix]

There is a possibility of porting O style combat over to Angband. In this style, + and - to weapons are given as percentage increase to hit dice instead of straight additions. If you're planning to do any major changes to combat, it might be a good idea to consider changing to O style.

[d_m]

Originally Posted by jens

We currently have 4 combat modes: melee, spells, ranged, and throwing. Should we have parity between them?
- No: melee > spells, ranged >> thrown
- All classes/races should be able to melee successfully.
- Some classes/races should be able to use spells as the main combat mode.
- Some classes/races should be able to use ranged as the main combat mode.
- In close combat (i.e. when standing next to a monster) melee should be the most effective combat mode for all classes.
- Throwing is mostly for flavour.

I strongly disagree about throwing--one of my goals is to try to fix the mechanic so you can have critical hits when throwing (like melee/ranged) and also so that the breakage rules make more sense. I actually think fixing potions to effect monsters when thrown, and changing throwing damage calculations are a way to add depth to the game.

I don't agree that we must have/maintain melee > spells/ranged > throwing. If you plan to do a redesign then I think it should actually challenge the status quo in this regard.

Finally, I think mentioning race here confuses the issue. I think you just mean that warriors can't cast spells, and that rogues don't really get attack spells. But I don't think any race/class should be forbidden from doing melee/ranged/throwing, and I think any race/class should be able to be use any of those as the main mode, modulo stats, equipment and interest. (For instance, some players may hate keeping track of ammo for ranged/throwing).

Originally Posted by jens

Increasing combat abilities is achieved through character level, stats, and equipment. To keep things clear I include stats gained through equipment in the category stats, and not in equipment. Should we have parity between them?
- Yes. I don't feel strongly either way, but the more parity we have, the easier it will be to balance the growth of combat abilities.

Your question kind of hems us in--the spell mechanic is (currently) tied to equipment (via books) and level (via slots) but you could imagine it working differently. Also many variants give classes (like warriors) abilities at various points (either chosen, predetermined or random) which might help.

We should NOT require parity between character level and stats/equipment. If we end up with it and things feel fine, OK, but I don't think it should be built in as a requirement.

Originally Posted by jens

The growth curve of player combat abilities should match the growth curve of dungeon difficulty. That is, if you keep a steady pace in your dungeon exploration, you should, on average, have the same level of challenge throughout the dungeon.
- Yes. (This statement includes a parameter that does not really belong to a combat system, but it is useful as a reminder: any system we design needs to include plans to achieve parity with dungeon development.)

I don't agree that this is a useful invariant to work toward, and as the person interested in dungeon generation I feel free to ignore it. I don't think we have this now, and since there is (intentionally) no concept of "steady pace" we aren't balancing for it.

In my opinion the game difficulty is (and should be) a sawtooth wave pattern (http://en.wikipedia.org/wiki/Sawtooth_wave), where the game gets harder as you descend until you find a game-changing item (artifact, dungeon book, ring of damage, dagger +9,+9, etc) which then shoots the difficulty straight down, and you proceed.

In this model the amplitude (height) of the wave corresponds to diving speed: if you go fast, things get harder quickly but you still find incredible items that make things easier (e.g. =speed +8). If you go slow, then the wave is much flatter, and the period is also longer, because you are exploring more slowly.

Originally Posted by jens

The growth curve of player combat abilities should be stable. That is, no matter what happens in your game, you should quite steadily increase in combat abilities.
- No: though in a long series of games the average power level should follow a smooth curve, in any given game big variations from this curve can, and should happen.

I think most people (including myself) agree with this. The implications of this are the gaining levels should NOT be the primary way to improve combat abilities.

Originally Posted by jens

Should all combat modes have the same characteristics when it comes to development?
- Yes.

I don't have any idea what this means, so I'm not sure if I agree with it.

[Derakon]

Originally Posted by jens

I'll make a set of statements, questions and answers concerning what I feel is relevant to get combat design right. Please help me flesh this out, challenge any and all points were I've missed some issue, or got something totally wrong.

We're nothing if not good at arguing.

We currently have 4 combat modes: melee, spells, ranged, and throwing. Should we have parity between them?
- No: melee > spells, ranged >> thrown

Melee should have better DPS (a.k.a. damage per [s]second[/s]turn) than spells, but spells offer massively more flexibility as well as the ability to avoid melee range. So I'm not sure that you can say that one is conclusively better than the other unless you're only considering DPS.

- All classes/races should be able to melee successfully.

Eventually. Gnome mages are going to be worthless meleers until the near-endgame.

- In close combat (i.e. when standing next to a monster) melee should be the most effective combat mode for all classes.

When appropriately geared up. A midgame mage will probably favor a weapon with +5 INT over one with better melee damage, for example, in which case he'll probably be doing rubbish melee damage.

- Throwing is mostly for flavour.

I certainly don't mind throwing being made effective, but it's sort of orthogonal to the rest of the problem.

One thing you missed:
- The combat algorithms should be as simple and transparent as possible.

Increasing combat abilities is achieved through character level, stats, and equipment. To keep things clear I include stats gained through equipment in the category stats, and not in equipment. Should we have parity between them?
- Yes. I don't feel strongly either way, but the more parity we have, the easier it will be to balance the growth of combat abilities.

I tend to prefer the player's power to derive in largest extent from equipment, as that reduces the incentive to grind for power. I recognize others may feel differently, but I'm going to complain any time someone suggests moving stats from gear to "internal".

The growth curve of player combat abilities should match the growth curve of dungeon difficulty. That is, if you keep a steady pace in your dungeon exploration, you should, on average, have the same level of challenge throughout the dungeon.
- Yes. (This statement includes a parameter that does not really belong to a combat system, but it is useful as a reminder: any system we design needs to include plans to achieve parity with dungeon development.)

I read this as being more a statement about the dungeon difficulty curve and the item generator. It seems more or less irrelevant to general combat mechanics overhauling.

The growth curve of player combat abilities should be stable. That is, no matter what happens in your game, you should quite steadily increase in combat abilities.
- No: though in a long series of games the average power level should follow a smooth curve, in any given game big variations from this curve can, and should happen.

Thank you.

[bulian]

I strongly disagree about throwing--one of my goals is to try to fix the mechanic so you can have critical hits when throwing (like melee/ranged) and also so that the breakage rules make more sense. I actually think fixing potions to effect monsters when thrown, and changing throwing damage calculations are a way to add depth to the game.

d_m does this mean quivers in V will eventually be able to hold throwable items, such as the 'thancs etc?

I tend to prefer the player's power to derive in largest extent from equipment, as that reduces the incentive to grind for power.

Agreed. This is currently true for melee characters but not true for casters. A warrior finding =damage(+10) at DL30 likely has his damage output increased by 25-30%. Casters have no corresponding ring, and spell damage is directly tied to CL.

Melee should have better DPS (a.k.a. damage per [s]second[/s]turn) than spells, but spells offer massively more flexibility as well as the ability to avoid melee range.

I agree this is how things currently are but I'm not sure that this is ideal. In other games mages are more "glass cannons", capable of dealing large amounts of damage but being somewhat worthless in longer fights. Would doubling mana cost and damage output of damage spells be interesting? I don't know, but I have no desire to ever play a mage with their current setup.

[Tibarius]

We currently have 4 combat modes: melee, spells, ranged, and throwing. Should we have parity between them?
- No: melee > spells, ranged >> thrown

Why not? Regardless if you play a spellhurling mage, an arrow shooting ranger or a hack&slash warrior - will the difficulty not vary very much if the damage output does not lie on the same average?

How is ranged/spell/throw attacks balanced to melee anyway? The first attack modes reduce the probabily of getting hit and result in damage only by ranged attacks of monsters.

All classes/races should be able to melee successfully.

How is successfully defined? For classes i consider as non-melee classes (mage, ranger) i would say no, basically only as emergency-option but far inferior to their primary way to deal damage.

-Some classes/races should be able to use spells as the main combat mode.
- Some classes/races should be able to use ranged as the main combat mode.

No, i would go further and say for some classes (not race dependant) spells / ranged should be possible to be used as ONLY combat mode

In close combat (i.e. when standing next to a monster) melee should be the most effective combat mode for all classes.

No, mages, rangers, priests (?) should be able to deal more damage via spells/ranged/prayers.

Throwing is mostly for flavour.

Throwing knifes (or axes, or the like) could work similary to ranged attack.
Throwing potions (bombs) or the like could develop to an option. But for most classes except thieves do not fit into the picture i personally have from those classes.

Increasing combat abilities is achieved through character level, stats, and equipment. To keep things clear I include stats gained through equipment in the category stats, and not in equipment. Should we have parity between them?

a) Counting stats gained from equipment in the category stats is bad under the aspect of that equipment is limited to what you have and the number of available slots (requiring player decisions), while stats are basically grinded with stat-potions until they are all at 18/100.
b) Combat power has two sources in my eyes - equipment and knowledge about the combat style (class dependant). Equipment can be found. Knowledge basically is gained through learning-by-doing - and thus can be raised through (or even requires) grinding. Nevertheless for a steady character-development (and to even out lucky finds) i think it is necessary to have also the 2nd aspect in raising character power.

[Derakon]

Originally Posted by bulian

Agreed. This is currently true for melee characters but not true for casters. A warrior finding =damage(+10) at DL30 likely has his damage output increased by 25-30%. Casters have no corresponding ring, and spell damage is directly tied to CL.

Mages will probably end up using =INT, which may not directly increase their damage but certainly directly increases their ability to use magic to deal damage. The mage's problem isn't usually DPS but rather combat endurance (a.k.a. mana), and equipment plays a major role there. Likewise, priests have "WIS.

I agree this is how things currently are but I'm not sure that this is ideal. In other games mages are more "glass cannons", capable of dealing large amounts of damage but being somewhat worthless in longer fights. Would doubling mana cost and damage output of damage spells be interesting? I don't know, but I have no desire to ever play a mage with their current setup.

Mages in Angband are already made of glass; you can't reasonably reduce their hitpoints much more. Heck, they get zero additional hit dice on top of their racial base, in addition to a -2 CON penalty.

Nor is their damage rate is bad right now, honestly. Keep in mind that the "500 damage/round" estimates we throw around here for melee don't take into account hit chance. When fighting a big target you're probably not going to be doing much better than a 75% hit rate, but spells always hit once you get your success rate to 0%. Devices are also pretty much a sure thing for mages, and they get a damage bonus based on their device skill (as does everyone else, but mages are awesome at devices). A level-50 gnome mage probably has better DPS than most warriors when using wands of annihilation. Their device skill is 13 from INT, 22 from race, 101 from class/level, minus 60 for the object level, = 1.76x damage multiplier on a flat 250 damage/use, or 440 damage, every time.

[Magnate]

Originally Posted by jens

Combat redesign is overdue. I'd like to get some significant changes done in time for 3.4. Since combat changes need much discussion, and much play-testing, we need to start with it now. There are probably lots of threads about combat changes lying about, but I want to start by taking a step back.

It's great that you're keen to work on this - the more the merrier. But please please make sure you go back and read at least most of the long threads on this topic over the past year or so. It is really irritating to have the same conversations over and over again because people think they're making great points for the first time. In particular, there have been threads about +dam and threads about AC, as well as threads about melee/spell parity and threads about archery.

We currently have 4 combat modes: melee, spells, ranged, and throwing. Should we have parity between them?
- No: melee > spells, ranged >> thrown
- All classes/races should be able to melee successfully.
- Some classes/races should be able to use spells as the main combat mode.
- Some classes/races should be able to use ranged as the main combat mode.
- In close combat (i.e. when standing next to a monster) melee should be the most effective combat mode for all classes.
- Throwing is mostly for flavour.

While I agree that we do not need parity of damage output, we need overall parity of risk/reward. So that means that melee ought to do the most damage while spells, archery and thrown ought to be about the same as each other. You get in close to do more damage but in return you're exposed to nastier attacks. (This is somewhat predicated on AI improvements and monster mana: at the moment closing to melee is often a good way to stop a monster using its nastiest spells.)

I disagree that throwing is mostly for flavour, I think it ought to be a viable alternative to archery. But this means we'll need special throwing weapons (including artifact throwing weapons), which can go in the quiver. But we needn't tackle everything at once - reworking throwing can follow on from the fundamental overhaul of melee combat.

I also disagree that all classes should be able to melee successfully, if you intended that to apply to all opponents. I see no reason why magi should be able to melee greater titans/demons and other tough melee opponents - or at least not without a significant reliance on buffs.

Increasing combat abilities is achieved through character level, stats, and equipment. To keep things clear I include stats gained through equipment in the category stats, and not in equipment. Should we have parity between them?
- Yes. I don't feel strongly either way, but the more parity we have, the easier it will be to balance the growth of combat abilities.

No. A strong consensus emerged in the stealth thread that Angband should lean more towards abilities being boosted by stats/equipment than by clev, to discourage grinding.

The growth curve of player combat abilities should match the growth curve of dungeon difficulty. That is, if you keep a steady pace in your dungeon exploration, you should, on average, have the same level of challenge throughout the dungeon.
- Yes. (This statement includes a parameter that does not really belong to a combat system, but it is useful as a reminder: any system we design needs to include plans to achieve parity with dungeon development.)

This sounds a little like a statement of the obvious, but I'm not sure it's terribly important either way. Angband has had a complete disconnect between these two things for most of its life, and it hasn't stopped it being an enjoyable game. Too much smoothness can lead to dullness - some residual variation in the challenge level wouldn't be a bad thing.

The growth curve of player combat abilities should be stable. That is, no matter what happens in your game, you should quite steadily increase in combat abilities.
- No: though in a long series of games the average power level should follow a smooth curve, in any given game big variations from this curve can, and should happen.

Agreed - this is linked to the anti-grinding point above.

Should all combat modes have the same characteristics when it comes to development?
- Yes.

What does this question mean? My instinctive answer is no, the four combat modes should have hugely varying characteristics. For example, it is traditionally quite hard to kill with spells at the start of the game - magi have only MM, and priests have nothing (except Bless for buffing their melee). Throwing oil, on the other hand, is quite effective at the beginning and rapidly diminishes in utility.

[Tibarius]

A quick reply on Derakon's thought about Mages - since i only play gnome mages currently (currently having one at clvl 50, all 9 book, hunting uniques)

Mages will probably end up using =INT

No, first of all mages max CON, else they are suspectible to one-hit kills. CON is MUCH MORE important than INT.

Mages in Angband are already made of glass; you can't reasonably reduce their hitpoints much more. Heck, they get zero additional hit dice on top of their racial base, in addition to a -2 CON penalty.

True, see point above.

Nor is their damage rate is bad right now, honestly. Keep in mind that the "500 damage/round" estimates we throw around here for melee don't take into account hit chance. When fighting a big target you're probably not going to be doing much better than a 75% hit rate, but spells always hit once you get your success rate to 0%.

Not true!
Mana Storm stays at 14% failure with an INT of 18/***(!) but it is almost impossible to max INT AND CON at the same time. Mana Storm has 400 damage not 500 (playing version 3.2.0).
Reliable spell with 0% fail and no resist is meteor swarm with a damage output of 220 "only". This requires a lot of restore mana potions / staves on almost every bigger monster, even non-uniques.

Devices are also pretty much a sure thing for mages, and they get a damage bonus based on their device skill (as does everyone else, but mages are awesome at devices). A level-50 gnome mage probably has better DPS than most warriors when using wands of annihilation. Their device skill is 13 from INT, 22 from race, 101 from class/level, minus 60 for the object level, = 1.76x damage multiplier on a flat 250 damage/use, or 440 damage, every time.

Not true, devices are utterly useless for mages (that could be a bug tho).
The time i find wands / staves or dragon breath/frost/power/annihilation i already had books with more powerfull spells.
My level 50 gnome mage with int 18/200 and clvl50 has a device rating of 132 (what does that exactly mean anyway?). If i look at a wand of annihilation the green number says 250 hp damage (not verified with rods of probing). And those high level staves often backfire and explode even with 0% spell failure rate, if i try to recharge them. And they have a remarkable lower success rate than my spells.

There are a couple of things which, in my eyes, are not well designed about mages - if mages should be a spell based class. I will open an own thread on this, once i have winner status

Not about mages, but combat in general, currently a character can use a weapon (regardless of weight), a bow/shooter, a light and a shield. Besides being un-realistic this enforces no thinking from the player.
Make lanterns / torches use up 1 hand, leaving only 1 hand available for small or medium weapons in the start. Once a mean of magic light (phial, star of xxx, maybe more other gear like amulet of magi etc) is found 2 hands are available for 2-hand weapons, shields or a 2nd smaller weapon. Big weapon = less hits but higher damage, 2 weapons = more hits with not so high damage, shield = more protection at the cost of offensive power).

[Derakon]

Originally Posted by Magnate

While I agree that we do not need parity of damage
output, we need overall parity of risk/reward. So that means that
melee ought to do the most damage while spells, archery and thrown
ought to be about the same as each other.

Keep in mind that mages, who have the best attack spells, are also inherently at a continually higher risk level due to their low survivability.

You get in close to do more damage but in return
you're exposed to nastier attacks. (This is somewhat predicated on AI
improvements and monster mana: at the moment closing to melee is often
a good way to stop a monster using its nastiest spells.)

To my knowledge, current monster AI chooses first between spells and movement, and only attacks in melee if movement takes the monster into the player's tile. That is, being in melee range has no effect on spell frequency. Has this changed?

Originally Posted by Tibarius

No, first of all mages max CON, else they are
suspectible to one-hit kills. CON is MUCH MORE important than
INT.

Everyone uses CON rings and has a similar decision to make, though granted mages are at a disadvantage due to their inherently lower HP. Still, given the choice between a =Damage that increases my nominal damage/round by 25% or a =CON that increases my HP by 100, I'll go with the =CON, every time.

Not true!
Mana Storm stays at 14% failure with an INT of 18/***(!) but it is
almost impossible to max INT AND CON at the same time. Mana Storm has
400 damage not 500 (playing version 3.2.0).

86% of 400 is 344; 75% of 500 (for the warrior) is 375 and you have to stand in melee range. I wasn't talking just about the endgame manastorm, though; similar calculations hold for Rift and Meteor Strike for earlier fights.

Reliable spell with 0% fail and no resist is meteor swarm with a
damage output of 220 "only". This requires a lot of restore mana
potions / staves on almost every bigger monster, even non-uniques.

Which is why you use devices.

Not true, devices are utterly useless for mages (that could be a bug tho).
The time i find wands / staves or dragon
breath/frost/power/annihilation i already had books with more
powerfull spells.

You're ignoring the utility of being able to store up your offense in those wands. Wands require no MP to use, just to recharge, and you don't have to recharge in the middle of a fight (but if you can find time to do so, it's still cheap MP-wise). Plus Annihilation has a better success rate than Manastorm and higher damage.

My level 50 gnome mage with int 18/200 and clvl50 has a device rating
of 132 (what does that exactly mean anyway?). If i look at a wand of
annihilation the green number says 250 hp damage (not verified with
rods of probing). And those high level staves often backfire and
explode even with 0% spell failure rate, if i try to recharge them.
And they have a remarkable lower success rate than my spells.

You subtract the object level from 132 to get your damage multiplier, which unfortunately is not reflected in the 'I'nspect screen. For wands of Annihilation, which are level 60, that means your multiplier is 1.72, giving you 430 damage/shot at no mana cost (at the time of using the wand). Damage from devices is much higher for mages than is immediately apparent. The only bug here is that the 'I'nspect screen is misleading.

I recently played a mage through to victory and relied heavily on devices pretty much the entire time. If you aren't using devices as a mage, you're seriously hampering yourself. The mage's mana pool is simply too small to be able to kill big targets using only attack spells.

Not about mages, but combat in general, currently a character can use
a weapon (regardless of weight), a bow/shooter, a light and a shield.
Besides being un-realistic this enforces no thinking from the player.
Make lanterns / torches use up 1 hand, leaving only 1 hand available
for small or medium weapons in the start. Once a mean of magic light
(phial, star of xxx, maybe more other gear like amulet of magi etc) is
found 2 hands are available for 2-hand weapons, shields or a 2nd
smaller weapon. Big weapon = less hits but higher damage, 2 weapons =
more hits with not so high damage, shield = more protection at the
cost of offensive power).

Dual-wielding, as typically seen in RPGs, is utterly idiotic. The few fighting styles that use two weapons typically have one mostly for attack and one mostly for defense (e.g. rapier/main gauche).

There's a few problems I have with trying to track hands accurately:
1) Switching to ranged mode involves swapping inventory around, which is a pain.
2) Suddenly bucklers become awesome shields because they strap onto the forearm instead of requiring a hand.
3) Reducing the number of equipment slots has serious ramifications on balance, since every slot contributes significantly to a player's power level.
4) Not finding a handsfree lightsource seriously hoses the player.