melee hit probability

[Nick]

IIRC splitting evasion and absorption was how v4 combat started - if I'm wrong, I'm sure someone will correct me

[TJS]

Am I allowed to mention Sil?

The way evasion and armour is handled is brilliant. Heavy armour absorbs damage but makes you easier to hit (and the extra weight reduces your stealth too), so you have the interesting choice of weather to maximise evasion or absorption.

Armour also has damage reduction dice rather than an arbitrary AC value.

[mushroom patch]

At least in some variants, to-hit also effects your odds and quality of critical hits.

I'd just like to emphasize that if you want to-hit to make as much difference as you're talking about here, you're going to end up with a situation where a typical character spends a lot of time missing at certain parts of the game and that's really not a place you want to be. If you want individual points of to-hit to matter more, just rescale the system (so a +1 becomes a +3 or so) or increase the impact of to-hit on critical hit percentages. You should also consider the symmetry with +to-dam numbers, even though it may be an apples to oranges comparison, you probably don't want to be giving +30% chance to hit in place of +30 in the current system.

Additionally:

If you stack blessing, heroism, and berserker, then you get something like +30-40 to-hit. This will increase your hit probability by maybe 5% against a deep enemy. I'm not saying that's irrelevant, but accuracy bonuses don't make much of an apparent difference in combat.

5% more hits = 5% more damage. How much time do people spend screwing around with equipment and looking at load outs for a difference of 5%? My feeling is that the difference is more than 5%, but I've never tested it rigorously in vanilla. Certainly in tomenet and probably other variants, it's considerably more than 5%.

[Derakon]

I'd just like to emphasize that if you want to-hit to make as much difference as you're talking about here, you're going to end up with a situation where a typical character spends a lot of time missing at certain parts of the game and that's really not a place you want to be.

Whyso? If most monsters have zero evasion and most characters have +0 to-hit, then your hit rate is flat throughout the game. This needn't be one of the stats that is subject to number inflation.

5% more hits = 5% more damage. How much time do people spend screwing around with equipment and looking at load outs for a difference of 5%? My feeling is that the difference is more than 5%, but I've never tested it rigorously in vanilla. Certainly in tomenet and probably other variants, it's considerably more than 5%.

While it's true that 5% more hits is 5% more damage, in most fights 5% more damage does not make the fight end any sooner (i.e. you overkill by more than 5%). I'm not saying it's insignificant, but it's not worth the investment in terms of inventory slots to achieve that, especially for berserker. And 5% is pretty accurate from what I remember of the last time I did this test. To-hit bonuses are really not that significant in the game.

Anyway, my biggest issue with to-hit is not its significance or lack thereof; mechanically the current system is pretty well-balanced, if prone to producing rather useless items (Rings of Slaying/Accuracy). My issue is how opaque it is. The math you have to do to determine your hit rate is pointlessly complex. Angband should have formulae that are as simple as possible while still providing the desired behaviors.

[mushroom patch]

Right, you wouldn't bother with these buffs except early in the game and against endgame uniques, where 5% might actually count.

also re: sil, I've only played it a few of times, but my impression from playing and reading the manual is that sil has the kind of wonky combat situation I think angband manages to avoid -- lots of missing with hits doing a wide spread of damage, not very normally distributed (which is not good -- approximating normal distributions is key).

[TJS]

At least in some variants, to-hit also effects your odds and quality of critical hits.

I'd just like to emphasize that if you want to-hit to make as much difference as you're talking about here, you're going to end up with a situation where a typical character spends a lot of time missing at certain parts of the game and that's really not a place you want to be.

I'm not sure I agree with this. The whole point of having an accuracy and evasion (or AC in the case of Angband) is that some monsters are going to be hard to hit for certain characters. Otherwise you may as well just set the hit chance to 80% and leave it at that.

Remember that mages, priests and rangers have alternative methods of dealing damage as well as melee.

If you want individual points of to-hit to matter more, just rescale the system (so a +1 becomes a +3 or so) or increase the impact of to-hit on critical hit percentages. You should also consider the symmetry with +to-dam numbers, even though it may be an apples to oranges comparison, you probably don't want to be giving +30% chance to hit in place of +30 in the current system.

Angband has really suffered from number inflation and it's hard to quantify what most of the numbers mean without years of experience playing the game.

also re: sil, I've only played it a few of times, but my impression from playing and reading the manual is that sil has the kind of wonky combat situation I think angband manages to avoid -- lots of missing with hits doing a wide spread of damage, not very normally distributed (which is not good -- approximating normal distributions is key).

You're correct in that the player and monsters do different amounts of damage depending on the damage, accuracy and evasion of the attacker and the defender. I see this as a good thing in that you tailor your equipment to take on certain foes and avoid others.

Also there is a wider spread of damage although that is mainly due to criticals rather than anything else which could be toned down quite easily. However I enjoy managing the risk that your opponent could get a lucky hit in and it certainly makes a change from just worrying about resistances, speed and HP when big breathers are about in Angband.

[Bogatyr]

I'm not looking at the code but the "feel" of, for example, the bless prayer for a beginning game priest is that it moves the player from "hardly ever hits" to "hits quite frequently." Maybe it's just all in my mind but I found that to be the case.

[Rydel]

Monster AC can be increased to increase the number of misses, extending the time to kill, and HP increases the total amount of damage to kill it.

Only real difference between AC and HP is that AC only protects against melee attacks; magic (after resistances/vulnerabilities) and shooting directly affect HP.

There's also a difference in terms of healing - damage prevented by AC doesn't need to be healed while damage absorbed by HP does. Not having to heal that damage reduces downtime between fights, reduces time spent healing in a fight (which in turn reduces the time to kill the monster), and reduces the chance that the monster will injure you faster than you can heal.

[wobbly]

Additionally:



5% more hits = 5% more damage. How much time do people spend screwing around with equipment and looking at load outs for a difference of 5%? My feeling is that the difference is more than 5%, but I've never tested it rigorously in vanilla. Certainly in tomenet and probably other variants, it's considerably more than 5%.

Just gonna mention that +5% to hit isn't 5% more hits. If your to hit probably was 1% you'd be hitting 6 times as often.

Edit: Though if my maths is right at 75% to hit it pretty much is (an extra 6.6666 etc. % damage). I'm not sure I'd bother much getting an extra 6 damage/100, not when it takes a turn (say a scroll or spell of bless), unless I have a free turn anyway.

Edit 2: Of course if I can kill it in 1 hit, & it can kill me in 1 hit the situation is a little different....

also re: sil, I've only played it a few of times, but my impression from playing and reading the manual is that sil has the kind of wonky combat situation I think angband manages to avoid -- lots of missing with hits doing a wide spread of damage, not very normally distributed (which is not good -- approximating normal distributions is key).

This is arguably more a matter of game style then good or bad. Having a more normally distributed damage range makes the game more detirministic. It's a different style of risk management in Sil anyways. Fewer instakill enemies & less escapes.

[Timo Pietilä]

I'm not looking at the code but the "feel" of, for example, the bless prayer for a beginning game priest is that it moves the player from "hardly ever hits" to "hits quite frequently." Maybe it's just all in my mind but I found that to be the case.

It's not in your mind alone. Early game those bonuses mean a lot, especially for char that is bad at hitting things. Later, when char gets levels and it's melee skill increases the % effect is a lot less. (bless also gives +5 AC IIRC, which can matter a bit early).

I think this is quite OK, I don't care how messy the calculation is, players are not supposed to make calcs, computers have been invented for that. If you feel that accuracy means something, then use it, if not, then don't use it. It's all about feeling, just like you say IMO.

[quarague]

I think there is a quite a difference in terms of game play if one separates evasion and absorption. For a very simple example, a gnome mage encounters two monsters:
Monster A has high evasion and few hitpoints, the mage has only a 20% chance to hit it but it will die to a single hit
Monster B has no evasion and lots of hitpoints, the mage has a 100% chance to hit it but will take 5 hits to kill it
Both monsters take an average of 5 turns to kill, but Monster B will take 5 turns every time while Monster A will have a huge variance. For B you can simply estimate whether you can afford bashing away at it for 5 turns (in terms of damage taken and other monsters around). Monster A is a lot more varied, if you are desperate you could attack it even if you know it will kill you after 3 turns. If you are not you should account for the possibility that you still haven't killed it after 10 turns. Overall this makes Monster A more dangerous and calls for different strategies.

[AnonymousHero]

Just curious, how many of the posters in this thread have played the now-abandoned v4? Oangband or FAangband? (I believe O pioneered the combat system used in O/FA.) I think it might be interesting to gather opinions about those alternate combat systems.

I can't claim to have tried v4, but "O" was "transparent" in techincal terms (+X% deadliness means to +X% dealiness), but completely opaque in practice ("Deadliness" means what?). It really did work in practice: Heavy weapons hit hard if you had the strength to wield them, and Daggers/MainGauches weren't the obvious choice of starting weapon for a warrior.

That's not to say that transparency can't have its benefits. In BG2 you know *exactly* the kind of damage a weapon is going to do on average at every character level. This is great for metagaming, but it's a bit... boring(?) for non-powergamers.

[jevansau]

I thought the combat system in V4 worked pretty well - at least for melee. I'd suggest also looking at Eyangband's combat - no deadliness at all, but to hit mattered and you needed to optimize for multiple hits with lighter weapons or few (often 1) hits with a heavy weapon.

[mushroom patch]

That's not to say that transparency can't have its benefits. In BG2 you know *exactly* the kind of damage a weapon is going to do on average at every character level. This is great for metagaming, but it's a bit... boring(?) for non-powergamers.

re: "boring(?)" -- it seems to be a common theme among people interested in roguelike design that if combat isn't super random, then it's "boring."

This is totally backwards. Random combat is bad. A competitive game cannot have crucial outcomes determined by chance. It's one thing if the setup of a situation is randomly generated, but if when you get into a combat situation the rules are based on high variance random processes, it damages the tactical integrity of the game, letting you escape bad situations through luck and making you lose in good situations by chance. A tight, carefully designed game minimizes both of those situations.

It really saddens me how many people are so wedded to the idea of creating variability in combat through dice throwing alone. At this stage, randomness in damage numbers should be purely a matter of flavor, not so much as to actually determine outcomes. (I should say, I think angband does a pretty good job of keeping outcomes from being determined by damage dice.)

[TJS]

re: "boring(?)" -- it seems to be a common theme among people interested in roguelike design that if combat isn't super random, then it's "boring."

This is totally backwards. Random combat is bad. A competitive game cannot have crucial outcomes determined by chance. It's one thing if the setup of a situation is randomly generated, but if when you get into a combat situation the rules are based on high variance random processes, it damages the tactical integrity of the game, letting you escape bad situations through luck and making you lose in good situations by chance. A tight, carefully designed game minimizes both of those situations.

It really saddens me how many people are so wedded to the idea of creating variability in combat through dice throwing alone. At this stage, randomness in damage numbers should be purely a matter of flavor, not so much as to actually determine outcomes. (I should say, I think angband does a pretty good job of keeping outcomes from being determined by damage dice.)

The game currently uses dice to determine outcomes. Splitting AC into absorption and evasion doesn't mean that it would increase or decrease randomness. Having insufficient armour or evasion meaning a critical hit is possible that could kill you isn't losing to randomness, it's not judging the risk of the encounter properly.

In fact I'd say it's no different to taking on a multi-hued dragon without poison resistance and then complaining you've lost to randomness when the dice decides that it breaths poison instead of one of the other elements that you have resists for.