Riposte: the post

[Hellmonk]

Lately I have seen a lot of threads on various forums suggesting that longblades are best suited for use on a high ev character or a "melee evasion build," whereas high ac characters should prefer maces or axes. I think that this advice is fundamentally misleading, so I'm going to explain to the best of my ability A) how riposte works, B) why the "melee evasion build" is a bad idea, and C) why riposte is not a good game mechanic.

A) How riposte works, probably excluding some edge cases.
Riposte activates on 1/3 of the melee attacks that you dodge, if you haven't already riposted in this turn; if a monster (say, an eight headed hydra) gets multiple attacks per turn and misses you multiple times you can't get more than one riposte, but you can riposte multiple times in between your turn if something faster than you (say, a shrike) gets two turns and misses you in melee at least once on both. You can't riposte retaliation attacks, nor can you riposte ranged attacks or spells even if they were used in melee range. Riposte triggers your aux attacks so it's functionally the same as getting to tab something an extra time in most cases. Riposte does not chop hydra heads, but can trigger *rage and send you berserk. Because blocking is checked before dodging, riposte triggers less if you have SH.

B) Why the evasion melee build isn't good.
The amount of damage that you get from riposte is not as high as you'd think, but more importantly it does not scale very well with EV. To calculate the amount of damage riposte gives us relative to tabbing something, we can use the formula [chance the monster misses * 1/3 * (our weapon delay/monster's attack delay)]. You can find out how likely a monster is to hit you by running fsim. Let's use stone giants as an example, assuming that we're using a sword at mindelay of 0.7 and no shield:
With a low evasion (10EV) character, the stone giant hits 77% of the time. Riposte will activate 7.6% (0.23*0.33) of the time we get attacked, but since we are tabbing at 0.7 and the stone giant only attacks every 1.0 turns we are only getting an extra 5.3% (7.6*0.7/1.0) damage from riposte.
With a high evasion (30EV) character, the stone giant hits 34% of the time. By the same formula as before, riposte activates about 22% of the time we are attacked but we only get about 15% extra damage beyond what we get from tabbing things.
So the net effect of adding TWENTY EV to our character is a ~10% increase in damage against a stone giant. This is an incredibly lousy return. Note that 10 EV is quite low even for a character in plate by the time vaults comes around, and 30 EV is often going to be hard to reach, and the damage increase is even smaller if you compare 20 EV against 40. You aren't adding 20 EV by training just a bit more dodging or by putting your stats in dex, and you are probably compromising your AC or your skilling or both to get close to that much. The stuff you sacrifice to go "all in on evasion" or whatever is pretty much always better than the additional damage you get from riposte.
Considering a more realistic example, suppose I can go from 25 EV to 28 EV at minimal expense. Under the same assumptions as before, I am going to get about 1.7% more damage against a stone giant from my additional ripostes. This is less damage than you get from a single point of strength. Increasing my EV may still be worthwhile on other grounds, but the damage scaling from riposte is tiny in any halfway realistic scenario.

C) RIPOSTE SUCKS won the last dieselrobin and here's why.

• Riposte does not significantly change your strategy. The evasion melee build is bad. Realistically, you're choosing the weapon you have a good apt in or whatever powerful thing you find early and not basing your decision on your body armor.Longblades happen to outdamage maces for most reasonable values of EV against most enemies for most of the game, so they are often the better choice even on the "high AC, low EV" character. I am a little unsure about how much this damage difference is; my feeling is that it's not significant but actually calculating the exact longblade damage for non-trivial cases is a nightmare (more on this later).
• Riposte does not significantly change your tactics, and where it does it overlaps heavily with cleave, which has the same effect of letting you damage a bunch of stuff that's surrounding you but is better at doing it. There don't need to be two different cleaves in the game.
• There is almost no way to know that riposte damage scaling is bad unless you understand how evasion and monster accuracy work in crawl. Riposte misleads players into making very bad strategic decisions because they don't understand this, and I can't really blame them because crawl accuracy formulas are really opaque. By extension, this makes riposte damage really opaque; it's hard enough to tell how much damage you do with a mace, with a longblade it's just about impossible.
• To follow up on the previous point, riposte damage is affected by a whole lot of other mechanics. Most of these make little to no difference in practice, but they could matter if you hypothetically buffed the riposte rate and dropped the base damage to make EV scaling matter more. Against a single monster, riposte damage is affected by:
  • How fast the monster is (affects how many rounds of attacks it will get). This is visible and is fine.
  • How many attacks the monster has (you can't riposte more than one, but it has a greater chance to miss at least once and if you don't riposte its first miss you can still riposte a later one). This is visible and is fine.
  • Monster HD (affects accuracy). This is opaque.
  • Whether the monster has the fighter flag (affects accuracy). This is opaque.
  • How often the monster casts spells. This isn't even available on the learndb, as far as I know. For every monster that has spells and also has a melee attack, the chance that it will do something that you can riposte depends on numbers that you have to dive the code to get. If riposte damage scaling was big enough to matter, this would be incredibly bad.
  • The accuracy of the monster's weapon, if it has one. This is visible, and it's so small that I doubt it would ever matter in practice.
  This is probably not a complete list. Again, though, it makes it nearly impossible to know how much damage you're actually doing to an enemy with your longblade even though these effects are small. At least with other weapons you can just fsim a target with similar AC/EV; to accurately determine damage with a longblade you have another huge stupid formula on top of that.
• The flavor sucks. Why can't I riposte with a 27 skill rapier, but I can riposte just fine with a zero skill triple sword?

tl;dr: The melee evasion build is bad, and riposte is a bad mechanic.

e: longblades don't outdamage maces anymore after the nerfs.

[duvessa]

One correction: long blades do not currently outdamage maces in most situations where you are actively attacking. If you are fighting with 0 skill long blades are a lot better though. Everything else hellmonk says is accurate.

[Siegurt]

There is a situation in which a riposting weapon will out damage all other weapons... And that's when you never actually attack with it.

If you're doing damage with spells, at melee range, sticking a long blade in your hand is superior to most other options (bonus spellpower from enhacement staves might or might not be better)

Not to say that that's a good thing, it's just something that didn't get mentioned as either good or bad, and it's something relevant to riposte.

[Shard1697]

I think someone in another thread recently brought up the idea of merging short blades and long blades(thus giving currently existing long blades stab bonus instead of riposte, probably dropping a couple of weapons-eg, in this situation both short sword and falchion probably wouldn't need to exist). I am in favor of this, personally.

[Implojin]

I think someone in another thread recently brought up the idea of merging short blades and long blades(thus giving currently existing long blades stab bonus instead of riposte, probably dropping a couple of weapons-eg, in this situation both short sword and falchion probably wouldn't need to exist). I am in favor of this, personally.

As a bonus, this is already implemented in Hellcrawl, and could presumably be pulled upstream at will after any necessary cleanup.

[Hellmonk]

It's not live on the servers yet, but I'm getting there.

[Sprucery]

If you're doing damage with spells, at melee range, sticking a long blade in your hand is superior to most other options (bonus spellpower from enhacement staves might or might not be better)

This is in my opinion the most important thing about riposte (I don't even care about the other issues). It's like the old protection ego, wield a weapon just for a passive effect (and carry one around just for this reason).

If riposte is going to keep existing, it should definitely be affected by your weapon skill (0 weapon skill = no riposte ever). This is also better flavour.

[luckless]

I take it that the argument in the OP is basically that, as currently implemented, riposte doesn't actually scale meaningfully with evasion and is therefore just a pointlessly baroque and misleading way of marginally increasing damage. However, this is equally consistent with both

(a) riposte should be removed, and
(b) riposte should scale with evasion more.

Personally, I'm partial to (b). There are, I take it, lots of ways to do this. For example, rather than being a constant 1/3, your chance of getting a riposte on a miss could be a function of your Dexterity and weapon skill, penalized by your armor's encumbrance rating. I know there are arguments against (b) too, but I'd like to see them before I agree that riposte is just bad.

[Gigaslurp]

I know there are arguments against (b) too, but I'd like to see them before I agree that riposte is just bad.

[*]Riposte does not significantly change your tactics, and where it does it overlaps heavily with cleave, which has the same effect of letting you damage a bunch of stuff that's surrounding you but is better at doing it. There don't need to be two different cleaves in the game.
[*]To follow up on the previous point, riposte damage is affected by a whole lot of other mechanics. Most of these make little to no difference in practice, but they could matter if you hypothetically buffed the riposte rate and dropped the base damage to make EV scaling matter more. Against a single monster, riposte damage is affected by:

• How fast the monster is (affects how many rounds of attacks it will get). This is visible and is fine.
• How many attacks the monster has (you can't riposte more than one, but it has a greater chance to miss at least once and if you don't riposte its first miss you can still riposte a later one). This is visible and is fine.
• Monster HD (affects accuracy). This is opaque.
• Whether the monster has the fighter flag (affects accuracy). This is opaque.
• How often the monster casts spells. This isn't even available on the learndb, as far as I know. For every monster that has spells and also has a melee attack, the chance that it will do something that you can riposte depends on numbers that you have to dive the code to get. If riposte damage scaling was big enough to matter, this would be incredibly bad.
• The accuracy of the monster's weapon, if it has one. This is visible, and it's so small that I doubt it would ever matter in practice.

This is probably not a complete list. Again, though, it makes it nearly impossible to know how much damage you're actually doing to an enemy with your longblade even though these effects are small. At least with other weapons you can just fsim a target with similar AC/EV; to accurately determine damage with a longblade you have another huge stupid formula on top of that.

These both seem like pretty compelling arguments against.

EDIT: Also probably not good to buff casters w/ 0 LB skill just holding a long blade for bonus damage

[Rast]

EDIT: Also probably not good to buff casters w/ 0 LB skill just holding a long blade for bonus damage

What about casters holding a long blade with 5.0 skill?

[luckless]

These both seem like pretty compelling arguments against.

Not to me. They're arguments for the claim that riposte is hard to assess tactically. (Technically, I'm not even sure that's true. It's usually pretty transparent whether a monster casts spells often or whether it has the fighter tag--if it's wearing a robe or called a wizard or something, it casts spells; if it's holding a weapon and/or looks skillful, it probably has the fighter tag; if it looks like a big dumb brute, it probably has neither.) But if riposte is justifiable at all, it's not justifiable on the tactical level, i.e. with respect to decisions about how to approach a given encounter. It's justifiable on the strategic level, i.e. with respect to decisions about which items to use or which stats/skills to develop.

[Siegurt]

If you're doing damage with spells, at melee range, sticking a long blade in your hand is superior to most other options (bonus spellpower from enhacement staves might or might not be better)

This is in my opinion the most important thing about riposte (I don't even care about the other issues). It's like the old protection ego, wield a weapon just for a passive effect (and carry one around just for this reason).

If riposte is going to keep existing, it should definitely be affected by your weapon skill (0 weapon skill = no riposte ever). This is also better flavour.

A spell attacker should almost always have some weapon skill for popcorn, it is just that with riposte, long blades also give you bonus damage when you are using conjurations.

If you want to remove this interaction, basing it on skill isn't the right approach, you probably want "riposte will only trigger if you did a melee attack as your last action" or something like that.

[Gigaslurp]

But if riposte is justifiable at all, it's not justifiable on the tactical level, i.e. with respect to decisions about how to approach a given encounter. It's justifiable on the strategic level, i.e. with respect to decisions about which items to use or which stats/skills to develop.

You can't just handwave away the tactical implications like that, they're the only thing that makes the item/skilling choices interesting! Why would I ever take Long Blades over M&F if I'm going to get roughly the same performance from both except one decided it'd be cool to significantly obfuscate how much damage I'm actually doing?

[bel]

While it is a theoretical possibility, it's virtually never best to wield a zero-skill long blade while casting conjurations. The extra damage (only in melee range) is minimal; and you're typically much better off with some kind of enhancer staff or some artifact with resistances or stat bonuses. Even if I plan to train some weapon for melee, I would pretty much never take into account this minimal benefit when choosing a weapon type.

[njvack]

The amount of game where you haven't found a relevant enhancer or stat stick is not exactly small. It also coincides nicely with the part of the game where any additional damage is useful...

[mattlistener]

Riposte has saved my hide a few times by killing things I was trying to get away from.

[shping]

While it is a theoretical possibility, it's virtually never best to wield a zero-skill long blade while casting conjurations.

Actual field report - greatsword of protection has been very helpful on my EE, it just randomly kills chaff (and boosts ac by 7!) while I dump my mana on the bigger threats.

[bel]

It's hard to discuss these kinds of things purely on the basis of thought experiments, but without numbers. Additional damage is always "useful"; the question is whether it makes a difference or not.

If you're a caster-type character in the early game, you typically don't want monsters to be adjacent to you at all, certainly not anything dangerous. If a dangerous monster gets adjacent to you, you certainly don't want to rely on riposte triggering by chance and then hitting the monster, to kill the monster. Therefore, to my mind, the riposte damage is mostly meaningless. It's better than nothing, sure. But 0.01(I plucked that number out of a random bodily orifice) being greater than zero doesn't mean anything in practice.

While it is a theoretical possibility, it's virtually never best to wield a zero-skill long blade while casting conjurations.

Actual field report - greatsword of protection has been very helpful on my EE, it just randomly kills chaff (and boosts ac by 7!) while I dump my mana on the bigger threats.

To clarify, I was not talking about popcorn. I would call a greatsword of protection primarily a defensive weapon: similar to a buckler. Suppose you had +7 ac and no weapon; would you die to the popcorn monster? Seems unlikely to me. It's also a very special case; I was talking about long blades in general.

[Doesnt]

Dumb rework idea: allow riposte to proc on attacks even if they don't miss. Change riposte rate from being constant on all misses to (constant * PC's weapon skill / mindelay required for weapon). Don't allow ripostes if the player's last action was a melee attack or a throw.

This makes long blades' offensive power much less murky: it is just the base damage as written when tabbing. Its niche is more focused around flexibility, as long blades users can reposition while maintaining partial damage output and can mini-cleave by waiting.

It also eliminates the 0 skill long blades "problem". Still some weirdness with regard to monster cast rates and shooting though, and I think the devs wanted EV to be relevant so I'm sure someone wouldn't be happy about that going away.

[duvessa]

Dumb rework idea: allow riposte to proc on attacks even if they don't miss.

*chorus of angels descends from the sky* minmay propoooosed that in ##crawl-dev literally yesterday and you were there

  Code:

14:32:31 <minmay> I think if anything it would be better if riposte didn't depend on EV at all; just have the same chance to riposte whether you dodge, block, or get hit (ofc you need to change the name from "riposte")
14:32:52 <Pleasingfungus> what would that riposte mechanic add to the game?
14:32:54 <minmay> then the damage output from it would just depend on how often the monster attacks, which is pretty perceptible to users
14:33:02 <minmay> whereas how often you dodge isn't at all
14:33:03 <Pleasingfungus> like, what decisions would the player make based on it?
14:33:30 <minmay> The same that current riposte adds: you're better against monsters that make many melee attacks at the expense of being worse against monsters that don't
14:33:47 <Pleasingfungus> well, that's not how riposte works right now.
14:33:57 <advil> I sort of agree with the flavor point about rapiers, but I think if anything the issue is that rapiers are misnamed for what they generally do in crawl
14:34:05 <minmay> and, theoretically, would decide on weapons based on whether you want to be better against the first group of monsters or the second
14:34:09 <minmay> How is it not how riposte works right now?
14:34:26 <minmay> It's certainly how I treat the decision
14:34:26 <hellmonk> yeah, right now you're better against monsters with low hd and no fighter flag and no spells and worse against those that have high hd or the fighter flag or cast a lot
14:34:28 <Pleasingfungus> hydras and trolls trigger riposte at most once per turn; executioners trigger it several times per turn.
14:34:45 <minmay> Oh
14:34:54 <Pleasingfungus> this is based on minotaur headbutt code. it's a change that elliptic made some years ago.
14:35:05 <minmay> I thought you were talking about EV or something
14:35:09 <Pleasingfungus> no.
14:36:05 <Pleasingfungus> anyway, i've got to go to lunch. good talking with you all!
14:36:14 <Pleasingfungus> have plenty of sexy elf dreams :)
14:36:22 <minmay> So, sure, if you want to be pedantic about it, replace "you're better against monsters that make many melee attacks" as "you're better against monsters that make many melee attack actions"
14:36:24 <minmay> ...
14:36:41 <hellmonk> something I learned today while running fsim is that weapon accuracy sometimes matters for low hd monsters against high ev players, as in I'm getting 20% in one fsim and 15% in another for an orc against 28 ev
14:38:14 <minmay> hellmonk: going from 80% to 85% dodge rate is a 25% decrease in the damage you take, yes, but it's only a 6.25% increase in the number of ripostes
14:38:30 <hellmonk> I wasnt talking about this being significant in the context of riposte
14:38:36 <hellmonk> just in the context of getting fsim numbers
14:38:54 <hellmonk> but I didn't make that clear, so mb
14:38:56 <minmay> hellmonk: well, I'm also responding to your hd and fighter flag comment because those are pretty small effects on riposte damage
14:39:05 <hellmonk> yeah I'm aware
14:39:28 <minmay> imagine if other weapon classes' damage were this impossible to evaluate lol
14:39:38 <hellmonk> but if riposte is going to exist it probably should not depend at all on those things, since they are extremely opaque
14:39:56 <hellmonk> it also depends on how often the monster casts spells, since you can't riposte a non-melee action
14:40:22 <rumflump> so wait, a hydra is making one attack that hits multiple times, but executioners actually make multiple individual attacks?
14:40:35 <hellmonk> hope you memorized the frost giant cast rate
14:41:37 <minmay> rumflump: how PF phrased it is misleading
14:41:38 <MarvinPA> rumflump: executioner was a slightly misleading example, a better comparison would be something like "hydra vs hasted thing that just has one attack"
14:41:41 <hellmonk> I believe it steps through a whole monster turn, however many attacks it gets, and you can retaliate once, then if it gets another turn you can retaliate again?
14:41:48 <minmay> yes
14:41:55 <hellmonk> so you can multi-riposte executioners bc theyre fast
14:41:57 <minmay> also, difference between bad mechanic and flawed mechanic
14:41:59 <rumflump> ahhh
14:42:11 <minmay> flawed mechanic: https://i.redd.it/n57v70kcjbkx.jpg
14:42:15 <hellmonk> but hydras are doing one melee action and hitting you with 4-8 attacks in that one action
14:42:19 <minmay> bad mechanic: http://i.imgur.com/OZ7o86N.jpg
14:42:33 <hellmonk> executioners get like 3 attacks or sth too
14:42:44 <hellmonk> in their turn
14:43:22 <Cheibriados> 03MarvinPA02 07* 0.20-a0-1086-g4bd2f5b: Fill in gaps in chequers_guarded_unrand_mask_of_the_dragon (#11043) 10(69 seconds ago, 1 file, 4+ 4-) 13https://github.com/crawl/crawl/commit/4bd2f5be180f
14:54:49 <minmay> thing that's no fun to implement but someone should probably do it: convert death channel to a fineff
14:55:47 <minmay> you can kill something with fireball or any bolt and the fireball/bolt hits the spectral thing too
14:55:51 <MarvinPA> i think i looked into doing that when i did it for infestation but it's harder because it has to actually keep track of what the dead monster originally was
14:55:56 <MarvinPA> so i gave up and didn't

(if this is a clever way of calling the idea dumb i 100% approve though)

[Doesnt]

I hadn't seen that, sorry! Just because I'm on the userlist doesn't mean I'm actually present or paying attention, I just leave it on all the time and maybe look at backlog. It's a good idea, I just called it bad because I'm in the bad habit of putting myself down a lot, and it also makes the name riposte unfitting since that implies a parry.

[mattlistener]

Any analysis of the state of riposte needs to take account of the fact that passive riposte damage occurs when you are not taking weapon-swinging actions. (Running away, using consumables, casting spells, invoking, evoking, etc.)

I assume that the benefit of the above would need to be balanced by long blades being a bit behind other base weapons in the "I am actively trying to kill you with this weapon" department.

[scott9027]

I would say the crux is right there -- the fact that riposte does damage when not actively attacking a monster in melee is the whole issue. This will always be superior than other weapon types (all other things being equal) if your character uses conjurations as offense.

Even on a Berserker using long blades, it basically ends up feeling and looking too much like cleave. I personally think merging long and short blades is by far the most elegant solution.

Off-Topic: Then bump the base damage of all M&F by 1 because I want to use them sometimes.

[Elitist]

I would say the crux is right there -- the fact that riposte does damage when not actively attacking a monster in melee is the whole issue. This will always be superior than other weapon types (all other things being equal) if your character uses conjurations as offense.

Even on a Berserker using long blades, it basically ends up feeling and looking too much like cleave. I personally think merging long and short blades is by far the most elegant solution.

Off-Topic: Then bump the base damage of all M&F by 1 because I want to use them sometimes.

Careful, most of the M&Fs already do really good damage, and people have won harder games with the Morningstar.

[Shtopit]

So riposte is like spines, but it activates if you dodge instead of when you are hit?

[damerell]

If Riposte granted a weapon delay bonus when attacking a monster whose last action was to miss you, it would still be an opaque source of extra damage (bad) but would eliminate the conjurer's passive riposting and the poor-man's-cleaving effects.

[Shtopit]

Is it bad that mages get free attacks, though? I think it's actually a good reason to move riposte to staves.

[Jeremiah]

My suggestion for a long blades special ability to replace riposte: feint -

Whenever you attack an enemy and miss, you get a (skill-dependent?) chance to do a second attack against the same enemy.
If the extra attack also misses, you don't get any more extra attacks.

This should have the effect of making long blades more effective compared to other weapons against monsters that are hard to hit, but not much more effective agains low EV opponents.

It would also make longblade-wielding monsters more of a threat to high EV/low AC player characters.

[Rast]

Jeremiah that's like a variable +hit bonus, best when you are hitting 50% of the time.

[prozacelf]

It would also make longblade-wielding monsters more of a threat to high EV/low AC player characters.

There are already way more serious threats to high EV/low AC player characters than there are to low EV/high AC characters. This is compounded by the fact that a good number of high EV/low AC characters also tend to be races that are low Str and/or have excellent dodging apts and poor armour apts.

EDIT: Also low HP!

[Paperbell]

My thoughts on this:

Even if the bonus from riposte is small, that doesn't mean that it is meaningless. the Doublesword has 1 less damage than the Eveningstar (14 against 15), and the Triplesword has 1 less base damage than the Executioner's Axe and Bardiche (17 against 18). If a one-handed sword wielder wants to outdamage a one-handed mace wielder, he/she just has to go above 7.14% extra damage, meaning he/she would need to end up dodging a stone giant's attacks little more than the OP's low evasion scenario. This gets a bit harder when a shield is taken into account, and since blocking comes before dodging, this can mean that in many cases, maces would come out on top. However, this only applies in corridors. A shield will fall in effectiveness as more attacks occur during the time between the shield-bearer's action turns. As such, a person that is just blocking the attacks of one monster for, say, 8 turns, will block more attacks than if 8 of the monster attacked simultaneously for 1 turn. This means that the character will end up needing to dodge much more when facing a group, meaning that the riposte result in a higher damage capability for swords than maces when facing multiple enemies.

When one is using a Triplesword, the character just has to do 5.88% more damage than Axe and Polearm users, at least for one on one. That means that even highly accurate foes will still take more damage as long as the character has passable EV, like around the 20s. This is much simpler since use of a Triplesword is very often not accompanied with use of a shield.

The Riposte encourages characters to build EV more, which, despite definitely not encompassing everything, is a defensive stat, so that a player that found a super nice two-handed sword is encouraged to not pour everything into it and die later because he/she couldn't dodge that bolt of whatever it is or that Two-headed ogre's second giant thing of death. Even if new players begin to notice that they are dying more often due to not putting time into AC, it is just another step to getting better at the game.

It also gives characters that wear lighter armor a means of fighting against faster opponents, like executioners. Where heavily armored juggernauts would yawn and calmly dismantle the unfortunate demon as it scratched at nigh impenetrable armor, a more dexterous warrior, though sustaining more damage than that of an AC-focused character, would be able to defeat the enemy more quickly due to the fact that more attacks = more ripostes (unless you are fighting a hydra, whose extra heads function as auxiliary attacks. Run. Or use your magic that your lighter armor permits you to use).

There is nothing wrong with riposte functioning more like a cleave that only works against other melee fighters. It boosts one on one combat with extra damage, and gives a character the option -- just the option, not the full out ability -- to fight multiple enemies at once. The lower skill requirement of swords is not there for nothing, either. The extra bit of xp will allow you to increase your dodge skill a small bit more, or even allow you to attain magic that will enable you to one-up enemies that don't use normal melee attacks.

It is just like the trope found in most rpg's: the sword is versatile. You can advance on a spear-wielder to attack them if you are able to dodge, you can go toe to toe with very powerful enemies and still outdamage them, and if the situation calls for it, you can face off with an entire crowd. I have fond memories of having a good portion of my hp missing and tearing through a pack of yaktaurs to regain full health, and cutting down pandemonium lords with surprising ease. While there are some things I would change, like the 1/3 chance or the fact that even bucklers lower the chances that a riposte would happen, I don't see much reason to remove it entirely. It doesn't matter if it is unrealistic; I have found that it adds to my gameplay experience as a whole.

[Shard1697]

The Riposte encourages characters to build EV more,

As has been explained thoroughly in this thread, it doesn't really actually incentivize this since the returns are so small.

[watertreatmentRL]

The hellcrawl approach of making long blades stab well instead of riposting and merging with long blades has a lot going for it, particularly with its elimination of a melee weapon skill. A while ago I tried doing some "EV melee characters" using long blades in DCSS. I did three of them and they were reasonably good. The DgGl I did was particularly good, though a lot of that is item drops. This is not to contest what hellmonk and others are saying about "EV melee," which of course is pretty bad generally.

My take on the whole thing is that the riposte concept would work a lot better as an interaction with SH than EV. The long blade could give a nontrivial contribution to SH and riposte on rolls where the SH bonus makes the decisive contribution (however you decide to hack that idea into the hit roll code). Whether the damage makes a big difference or not, having more SH definitely does make a difference. You also eliminate the oddity that SH stops ripostes.

[PlatinumSpider]

The Riposte encourages characters to build EV more,

As has been explained thoroughly in this thread, it doesn't really actually incentivize this since the returns are so small.

Except it does for the majority of players who don't go in depth analysis on the odds, they just see that when they dodge they sometimes get an extra hit and so train dodge more.

[PlatinumSpider]

My take on the whole thing is that the riposte concept would work a lot better as an interaction with SH than EV. The long blade could give a nontrivial contribution to SH and riposte on rolls where the SH bonus makes the decisive contribution (however you decide to hack that idea into the hit roll code). Whether the damage makes a big difference or not, having more SH definitely does make a difference. You also eliminate the oddity that SH stops ripostes.

Couldn't we just have EV check before SH, no need to block something you can move out of the way of and riposte would not get borked by shields.

[Hellmonk]

The Riposte encourages characters to build EV more,

As has been explained thoroughly in this thread, it doesn't really actually incentivize this since the returns are so small.

Except it does for the majority of players who don't go in depth analysis on the odds, they just see that when they dodge they sometimes get an extra hit and so train dodge more.

Yes, riposte misleads players into making bad strategic decisions. That's a problem with riposte, not a benefit; "this game mechanic encourages people to skill poorly unless they are spoiled" is a bad thing.

[Shard1697]

The Riposte encourages characters to build EV more,

As has been explained thoroughly in this thread, it doesn't really actually incentivize this since the returns are so small.

Except it does for the majority of players who don't go in depth analysis on the odds, they just see that when they dodge they sometimes get an extra hit and so train dodge more.

So the game is convincing them into thinking that pumping EV at the expense of other things is worth it with LB, which they would see is actually not true if they went out of their way to find obscure information... and this is a good thing?

[duvessa]

My take on the whole thing is that the riposte concept would work a lot better as an interaction with SH than EV. The long blade could give a nontrivial contribution to SH and riposte on rolls where the SH bonus makes the decisive contribution (however you decide to hack that idea into the hit roll code). Whether the damage makes a big difference or not, having more SH definitely does make a difference. You also eliminate the oddity that SH stops ripostes.

Couldn't we just have EV check before SH, no need to block something you can move out of the way of and riposte would not get borked by shields.

Then you would have the same oddity except instead of SH stopping ripostes, it'd be EV stopping reflections. (The change would also be a small buff to SH but I think that'd be fine.)

I'd think the easiest solution would be to riposte on both blocks and dodges, instead of just one.

[damerell]

My take on the whole thing is that the riposte concept would work a lot better as an interaction with SH than EV.

Then it won't work with two-handers at all, to a first approximation. [1] Mind you, the conceptual division of swords into Short Blades including rapiers that don't riposte and Long Blades including claymores [2] that do riposte is a bit of a puzzler, to be honest.

Perhaps Short Blades should become Blades and acquire the long sword, demon blade, scimitar, eudemon blade, and falchion. Blades would all stab (but the smaller ones would be better at it in some user-transparent way) and all riposte (but the larger ones better at it in some user-transparent way). Riposte with SH _and_ EV. If that's too good, drop the overall chance of a riposte.

Long Blades is now Great Blades, with the bastard^W double sword, great sword, and claymore. The difficulty now is finding a role for Great Blades which isn't doing a lot of damage like Maces or cleaving like Axes. Boost your EV while attacking (it's hard to approach someone who's swinging two metres of blade about)? But is that even going to be worth having / too much like protection?

This is not hellcrawl's merging of the two altogether, because that still gives you people running around stabbing with enormous two-handers...

[1] Yes, formicids, amulets of reflection, etc.
[2] Yes, I know, but it's a stupid name.

[Shard1697]

What's wrong with stabbing with 2handers?

[luckless]

Yeah, the more I think about it, the more I think just giving long blades stabbing is the way to go. (I wouldn't go full hellcrawl and just make rapiers the worst long blades though.)

[sooheon]

One other thought, add parry. Riposte should follow a parry, after all. It's a + to your dodge depending on long blades skill and maybe encumbrance.

[duvessa]

The reason small weapons are more suited for assassination is that they're easier to carry and conceal, and quiet to use (because you want to be able to get away afterwards). Now if this were a game about scaling the castle walls and climbing into the king's bedroom and killing him while he sleeps, that'd be one thing, but Crawl doesn't have that kind of assassination at all. Just monsters that are incapacitated, completely unaware of you, or distracted. In the former two situations, from a realistic perspective, killing someone with a decently heavy club is not really any harder than killing them with a dagger, and killing them with a huge axe or sword or maul definitely isn't harder. A distracted or confused person might be harder to kill with a more cumbersome weapon because they'd have time to react, but that's already covered by the fact that in these cases, you get more stab attempts with faster weapons.

The gameplay reason for stabbing being mostly exclusive to short blades is to make it an investment. Since short blades are clearly inferior to other weapons at everything that isn't stabbing, by training short blades you are spending some xp for the ability to stab better, instead of spending that xp on being better at regular melee.

So if you merge long blades and short blades, there is no reason to make the smaller blades specifically better at stabbing. It's not realistic and it no longer serves a gameplay purpose, since the same skill is used for both the small and large blades. The only thing it accomplishes is making players carry around two weapons, one for stabbing and one for everything else, which isn't a good accomplishment.

[prozacelf]

All weapons should be equally useless. We got rid of club stabs already, and they were fun and useless flavor. Flavor is bad. All weapon types should be the same, because if they weren't, some misguided person might choose one for reasons that have nothing to do with mathhammering the game to find out which is best.

[Shard1697]

Do you think riposte is good flavor? I don't really, considering how arbitrary the weapons that are and are not allowed to riposte are(it would make sense to be able to with almost anything, but in particular rapiers are the one weapon it would make the most sense to, and they can't!) and the weird ways it functions wrt player attack speed and multiple enemies/multiple enemy attacks/attacks that have multiple hits.
This mechanic hasn't been around super long and isn't super great, and I don't know why you would equate removal of it to removal of all flavor(or of flavor in general), or flattening all weapons out into one singular mush.

Also, do you not think it will be fun to stab with big weapons? Having played a number of games with boots of the assassin I can assure you it is.

[damerell]

What's wrong with stabbing with 2handers?

Well, I think of it as sneakily inserting a blade into a vital spot [1], which is probably quite hard to do if the point is two metres from your hands. If stabbing with claymores, why not stabbing with great maces or tridents? (You do, after all, still get the chance to avoid the target's EV and some bonus damage with another weapon, so it's not like sneak attacks don't exist at all without a Short Blade).

Game mechanically, as duvessa says, what's wrong with it is that two-handers are good already; you get it on short blades because otherwise they're not very good - hence in my proposal you get more of it the smaller your Blade is.

[1] On a jelly? I know, I know. But felids can eviscerate jellies and IDK how that works either.

[luckless]

Dumb rework idea: allow riposte to proc on attacks even if they don't miss.

I'm surprised there hasn't been more discussion of this idea, which has the following very strong considerations going for it:

--) it's continuous with an existing, well-established mechanic (the minotaur headbutt)
--) it's by far the least-fussy, easiest-to-balance way of realizing the closest thing riposte has to a tactical niche (i.e. making long blades better against melee foes, giving you a modicum of offense while casting/repositioning/using items)

[Doesnt]

  Code:

You riposte. You impale the shock serpent!! You burn the shock serpent!
 The shock serpent is severely wounded.
 The shock serpent's electric aura discharges, shocking you!

i just wanted to zap polymorph at it but apparently riposte can be countered so i took more damage from the shock serpent than i would've if i was using a different weapon

[luckless]

further considerations in favor of riposte where you get hit:

--) riposte where you get hit but don't take damage due to AC: https://youtu.be/WRYM6B7CTs8?t=42s
--) riposte where you get hit and totally take damage but whatever: https://youtu.be/doSjHWwHkvI?t=2m28s

eta: #2 is basically me vs. my character's first halberd gnoll like all the time

[genericpseudonym]

The only thing it accomplishes is making players carry around two weapons, one for stabbing and one for everything else, which isn't a good accomplishment.

This already happens. Daggers get a bigger damage bonus when stabbing than other short blades, but are the worst for 'honest' melee combat. Rapiers are the best for tabbing, but get lower stab damage than daggers. Brands like flaming or freezing are better for stabbing since they multiply the stab damage, but brands like elec or pain are better for not-stabbing since short blades have low base damage.

Stabbers are incentivised to carry around a rapier and a dagger, or an elec/poison dagger for melee-fighting and a freezing dagger for higher stab damage, etc.

Short swords and rapiers should just be buffed to be on the same tier as daggers for stab damage.