Derakon's combat revamp

[Magnate]

This is an unfortunate artifact of having somewhat separate build paths for different operating systems. For you, make -f Makefile.cocoa tests from the src directory should work. If you want to have make -C src work from the top level, you'll need to run sh autogen.sh && ./configure with appropriate arguments, to create mk/buildsys.mk.

Just for general info, noz (he of utf-8 fame) expressed some frustration with this recently (after spending fruitless hours trying to help someone build in Eclipse on Windows), and threatened to rewrite the whole system in a more coherent way. Now that really would be a Christmas bonus for Angband development.

[myshkin]

Magnate: sorry about breaking the unit tests! Unfortunately, if I do "make -C src" I get the error "../mk/buildsys.mk: no such file or directory". Indeed, buildsys.mk is nowhere to be found. Whoops?

This is an unfortunate artifact of having somewhat separate build paths for different operating systems. For you, make -f Makefile.cocoa tests from the src directory should work. If you want to have make -C src work from the top level, you'll need to run sh autogen.sh && ./configure with appropriate arguments, to create mk/buildsys.mk.

[Derakon]

Excellent, thanks Magnate (and everyone on IRC) for all your help with this project.

For everyone else, I want to make clear how much balance testing I've done on this, which is: very little. I've tossed level-50 kobold paladins (the default race/class combination for some reason) around the dungeon a bit, and took a level-1 half-troll warrior to the weapons store for some window shopping, but that's about it. Balance is almost guaranteed to be bad. I would have done more playtesting, but a) I'm on vacation, and have precious little time to spend with my family before everyone has to scatter to separate parts of the country, and b) I'm on a laptop and I don't know the roguelike controls.

At the moment I suspect that damage will tend to be too low rather than too high, but we'll see.

Here's the change log summary since my last summation on the 18th:

* Multiplied all finesse and prowess bonuses on equipment by 10 (including the combat rings, penalties on heavy armor, the random (+5,+5) on mithril shots, etc.). Well, the combat rings (=Finesse, =Prowess, =Slaying) aren't exactly comparable to their old values; they range from 20 to 120 now.

* Changed crits as I described earlier in the thread.

* Improved paladin combat skills (level 50: 140 finesse, 300 prowess).

* A bunch of bug fixes.

Magnate: sorry about breaking the unit tests! Unfortunately, if I do "make -C src" I get the error "../mk/buildsys.mk: no such file or directory". Indeed, buildsys.mk is nowhere to be found. Whoops?

[fizzix]

Whenever you're ready I have evasion working and rebased off of Derakon's commits. (hopefully I did this correctly!) Pull request sent.

I made a separate function to calculate monster hitting player probability and put it in melee2.c This is exactly the same as the old test_hit function since both players and monsters used it in the past. Now the calculations are completely different until we want to use player evasion.

right now the naming nomenclature is a bit funky. The evasion value is still referred to as m_ptr->ac. That'll be fixed when I implement the armor (absorption) half.

[Magnate]

With many thanks to Derakon for his blazing fast development, the first iteration of this new system is now in v4, available from the usual place. Please be aware that this is brand new work which has had no testing in actual games - it's hot off the keyboards (well, Derakon's keyboard). So, for optimum development, please split your feedback into

1. Bugs - things which aren't working at all - e.g. you can't hit anything, you do no damage, the numbers shown are wrong compared with wizmode, you never get crits, or something else has broken, or you can crash the game.

2. Balance - not actual bugs, but you think that your character's combination of race/class/stats/equipment has things too easy or too hard, either against certain monsters or in general.

Please don't worry unduly about balance at this stage, because fizzix's work on EvAbs will be going in pretty shortly, so we won't be doing fine balancing until after that. At the moment we're primarily looking for actual bugs in or caused by the new code. But overwhelming balance problems are of course worth catching sooner rather than later.

[Estie]

Sounds good then. The paladins traditional secondary attack is orb, at least late in the game its fairly effective.

[fizzix]

If finesse has a noticable impact on to hit chance, a difference in a factor of 3 is going to mess up the paladin. If you apply a nonlinear function to those numbers to calculate to hit chance which mitigates the effect, what have you gained for gameplay by coupling to hit and finesse ?

A first run through of the monster list with new evasion scores is up github

https://github.com/fizzix/v4/tree/evasion

armor (absorption will come later). Basically, paladins will have issues with a small class of monsters and shouldn't notice much difference for other monsters. The standard evasion value in the early game is 0 and the standard value in the late game is 10, meaning paladins will have a 75% chance to hit these monsters and the higher finesse characters will almost never miss.

All highly evasive monsters (except the special case of the will o the wisp) are vulnerable to devices. Paladins may want to carry some.

[Estie]

To add onto what Magnate said, `,o

* a level 50 warrior will inherently have 295 finesse and 375 prowess.
* a level 50 rogue will have 350 finesse and 165 prowess
* a level 50 paladin will have 90 finesse and 250 prowess

If finesse has a noticable impact on to hit chance, a difference in a factor of 3 is going to mess up the paladin. If you apply a nonlinear function to those numbers to calculate to hit chance which mitigates the effect, what have you gained for gameplay by coupling to hit and finesse ?

[Magnate]

At first, I thought it sounded weird too, but having thought more about it, it would make sense (in my opinion) to have the racial bonuses scale with level too. Realistically speaking, there is no 1 universal warrior for everybody in the world that fights the exact same way. Having racial bonuses increase with level would simulate the fact that each race/nation has different fighting styles and as their members improve with skill, they would normally do so in a manner which mirrors the fighting style of that race. One race may rely more on strength while another relies more on speed. As they improve in skill, the race that practices strength will get stronger, while the race that relishes speed will get faster.

I agree with this. At the moment no racial traits are tied to clev, but I think it would be good for this to be the first one. But let's give Derakon a chance to get the basic system working first. Adding racial bonuses will make balancing harder. I suspect that racial biases (hobbits and elves towards finesse, dwarves and half-trolls towards prowess) should make less difference (i.e. smaller numbers) than class biases.

[Derakon]

I'll freely admit that the numbers for the different classes are, so far, very rough. I'll probably refine them a bit before my pull request (which is tentatively scheduled for tomorrow) but they'll doubtless need more balancing later. That said, rogues have the worst casting of the hybrid classes, followed by paladins, then rangers. Assuming we don't want to change that at this juncture (please don't expand the scope of this project any more than it already is. ) that means that rogues should have the best combat, followed by paladins and then rangers.

I may well have overdone it for the paladins though. They should have better endgame prowess anyway.

[tg122]

the plan is to have race+class contribute about 1/3rd of your finesse/prowess, STR/DEX 1/3rd, and equipment 1/3rd.

This sounds like it's going to be a great combat system. Combine this with the fact that the stats sound like they're going to be harder to max, and the result will be a lot more variation in how different characters are played.

* a level 50 warrior will inherently have 295 finesse and 375 prowess.
* a level 50 rogue will have 350 finesse and 165 prowess
* a level 50 paladin will have 90 finesse and 250 prowess

I understand that these are still preliminary, but here is a suggestion if you want them. Why not just make the paladin have the same prowess/finesse ratio as the warrior but with less total points (maybe 305p/240f)? Then, the race and stats bonuses can determine whether the paladin wants to lean towards prowess or finesse. The paladin is typically a very skilled and versatile combatant. It just seems artificial to give them such a low finesse rating and make it difficult for them to wield anything but heavy weapons (why should it be more difficult for a paladin to wield a longsword without having to load up on dex). A rogue, on the other hand would make more sense to have a skewed ratio towards finesse because that is what rogues are all about.

Just something to consider when it comes time to focus on balancing... either way, awesome work!

[tg122]

Certainly you could increase the racial bonuses with level, though intuitively that seems a bit weird

At first, I thought it sounded weird too, but having thought more about it, it would make sense (in my opinion) to have the racial bonuses scale with level too. Realistically speaking, there is no 1 universal warrior for everybody in the world that fights the exact same way. Having racial bonuses increase with level would simulate the fact that each race/nation has different fighting styles and as their members improve with skill, they would normally do so in a manner which mirrors the fighting style of that race. One race may rely more on strength while another relies more on speed. As they improve in skill, the race that practices strength will get stronger, while the race that relishes speed will get faster.

[Derakon]

Certainly you could increase the racial bonuses with level, though intuitively that seems a bit weird -- your elf gets even more elfy with every level? Oh well; if it makes good gameplay sense then why not.

One random thing I just realized: finesse fighters will be balanced at least in part by the finesse penalties on heavy armor. It's hard to be an agile fighter in full plate armor. Meanwhile the guy swinging a heavy hammer doesn't notice so much. Which works out just like we wanted, completely serendipitously.

[sethos]

Note that currently racial skill bonuses are very minimal -- they affect your starting skills slightly and don't compound with level, unlike class bonuses. If you really wanted to increase the variation by class, you could make the racial skill mods affect the on-level class skill bonuses instead. Or you could magnify the STR/DEX bonuses and penalties.

Why not have the racial bonuses go up with level, just like the class bonuses? It seems rather intuitive to just do it that way - and it would provide an added thing to consider when you create your character (hmm... high elf will be better with finesse weapons, dunadan walk the balance, and trolls are all out prowessr... what to choose for my fighter?)

might make your race/class combo more important.

[Derakon]

To add onto what Magnate said, `,o

...

*picks up laptop off floor*

Whoops.

Anyway, the plan is to have race+class contribute about 1/3rd of your finesse/prowess, STR/DEX 1/3rd, and equipment 1/3rd. Currently, for example (and these numbers are subject to change because I really didn't give them much thought):

* a level 50 warrior will inherently have 295 finesse and 375 prowess.
* a level 50 rogue will have 350 finesse and 165 prowess
* a level 50 paladin will have 90 finesse and 250 prowess

Meanwhile, as I mentioned earlier in the thread, STR and DEX can give up to +260 to prowess and finesse, respectively. If you decide to focus heavily on DEX gear then you might have, say, 8 more points there than in STR, which will tend to mean you get about 80-100 more finesse bonus than prowess bonus. Of course your weapon will presumably magnify the finesse bonus and minimize the prowess bonus in this situation.

Note that currently racial skill bonuses are very minimal -- they affect your starting skills slightly and don't compound with level, unlike class bonuses. If you really wanted to increase the variation by class, you could make the racial skill mods affect the on-level class skill bonuses instead. Or you could magnify the STR/DEX bonuses and penalties.