Remove confusing touch

[bel]

Confusing touch is an extremely weird spell. It's not in any of the starting books and it's not clear why it even exists and why it works the way it does.

It is melee range which limits its usage (this is good). But also keep in mind that a melee attack is faster than a spell cast. For instance, if you swing a weapon with 0.5 delay, you can get two confusing touches every 10 auts. So the chance to confuse a golden dragon per 10 auts is 2 * 0.8 * (24 - 18) / 24 = 40%.

To debilitate an endgame enemy with this consistency is absurd for a lvl 1 spell.

If it does exist, it should work through MR, like confuse (why doesn't it do so already?).

[braveplatypus]

God forbid any hex do anything past lair

What kind of hooligan would run up and slap a panlord anyway? No decent self respecting off screen fire storming theoretically optimal player that's for sure. Slapping a pan lord to make them dizzy, who thinks of this stuff?

[Sorcerous]

It is melee range which limits its usage (this is good). But also keep in mind that a melee attack is faster than a spell cast. For instance, if you swing a weapon with 0.5 delay, you can get two confusing touches every 10 auts. So the chance to confuse a golden dragon per 10 auts is 2 * 0.8 * (24 - 18) / 24 = 40%.

To debilitate an endgame enemy with this consistency is absurd for a lvl 1 spell.

An endgame enemy. Singular.

When have you ever played through Z: and been all hex-happy with 1v1 engagements? The place is a battlefield. You can't smack one random moth there without casting Summon enemies on loop!

[petercordia]

I think confusing touch is a well-designed spell.
Having a way to hex a gold dragon is fun.
Though, in my opinion at present Confusing Touch is too cheap, because finding that it is always optimal to memorise Confusing Touch is unfun. Increasing the spell level would fix that problem.

[duvessa]

HD/XL-resisted stuff is awful, it doesn't even use one consistent formula so meph and confusing touch have slightly different chances of working, for example.

[TheMeInTeam]

Right now, this is the only way for a stabber to put ANYTHING aside "distraction" that allows stabbing on a fairly large group of enemies. And "distraction" means we're looking at a summoner, not a stabber...and a stronger build with or without CT.

If it's removed, I would like to see that reworked so that stabbing isn't significantly worse, because right now stabbing builds are already considered broadly weaker than alternatives.

I would agree that the utility you get from this spell is extremely strong for a level 1...if you're stabbing. For axe/most melee builds, ranged builds, blaster builds, and similar using confusing touch is a waste of time where you'll make many weapon swings that do nothing, then inflict confusion (which lasts less time than other sources of confusion) on a single monster. By the time we're confusing "things a lvl 1 spell shouldn't work against", quite a few alternative options could have already killed the monster, which is objectively superior.

I'd say push the level requirement on this up a bit, but I don't think that changes the fundamental issue with it and how it interacts with stabbing setups as a whole right now. Even at level 4 it's still a no brainer if you're using short blades in depths/zot/extended and still a waste of time in scenarios that matter if you're using literally anything else.

HD/XL-resisted stuff is awful, it doesn't even use one consistent formula so meph and confusing touch have slightly different chances of working, for example.

I recall in the past you were advocating for more attacks that ignore MR rather than fewer, in one of the discussions about a potential new enemy in snake. Have you changed your mind since then?

What kind of hooligan would run up and slap a panlord anyway? No decent self respecting off screen fire storming theoretically optimal player that's for sure. Slapping a pan lord to make them dizzy, who thinks of this stuff?

You'd have to ask the bot or wiki to know if it's even possible.

Dispater has 16 HD, so you'd have a reasonable chance of (briefly!) confusing him and getting stabs in. Cerebov has 21 HD, which means his spells/attacks can kill you several times over on average before you'd confuse him one time.

This is a bad idea regardless for obvious reasons, but short blade users can still use CT on various enemies who don't have 100's of HP and the ability to trivially kill them in the "average confuse time" window PRIOR to these fights right now.

It just seems odd to advocate for the complete removal of something that is propping up what is, per several poster's estimations, one of the weakest approaches to killing monsters in general (stabbing). A rework of how the stuff interacts sure, but outright nerfing one of the already slowest general methods of killing monsters? In what way does that make crawl better?

[bel]

God forbid any hex do anything past lair

What kind of hooligan would run up and slap a panlord anyway? No decent self respecting off screen fire storming theoretically optimal player that's for sure. Slapping a pan lord to make them dizzy, who thinks of this stuff?

I cannot really parse this post but I think it is trying to say that Hexes are weak past Lair. They aren't.

Hexes are good all game. Confuse/Cause Fear remain useful for the entire game, plus there are plenty of higher level Hexes with situational uses, like Darkness, Invisibility and Discord.

Even if it were true that Hexes are underpowered, adding a ridiculous lvl 1 hex is not the way to fix their power.

[petercordia]

I think braveplatypus's satire makes a point about magic-enabled stabbing, not the hex school of magic. He also makes fun of people calling CT 'ridiculous'.

Cause Fear, Darkness, Discord, and to a minor extend Confuse & Invisibility, make investing in Hex skill worthwile, but they don't let a short-blade specialist take out, say, a herd of Death Yaks. (Except Invisibility, but that wouldn't be castable by end-lair.)

[Hellmonk]

The bad parts of playing "stabbers" are having to train a shitty weapon and a bunch of stealth, so you could probably buff those things, for example by merging long and short blades. Melee range and untargeted hexes are good concepts imo but I think that hd dependence is pretty bad if hd isn't displayed anywhere. Would be nice to see these concepts reimagined going forward.

[Sprucery]

Just make it MR-dependent and be done with the issue.

[braveplatypus]

I really like confusing touches ability to confuse certain "MR: Immune" monsters, because i think the number of monsters with "MR: Immune" is way way too high. The idea of removing or "nerfing" a spell that can confuse a high MR monster for a few turns because it is too powerful is ridiculous. Getting into melee range of a dangerous monster to use it will be dangerous, you are now in melee range of a confused dangerous monster that can and will hit you with it's dangerous attack because that's something confused monsters in melee with you do. And assuming it's not a one on one scenario, the short duration of the confusion means the monster is far from "dealt with", as it possibly moves away from you towards other monsters, possibly throws an attack at you, or possibly snaps out of it's status effect. If spells where listed in order by their ability to eliminate danger to the player confusing touch wouldn't make top 20. Also the "it uses min delay" goes both ways, and if you're not a character yet trained to swing at things be prepared for a slower "cast" and possible miss.

Having said all of that crap, Hellmonk is right, a monsters resistance to the spell being dependent on entirely hidden from plain sight numbers and mechanics is just objectively bad, the power level of the spell aside. And just because i think MR: immune is overused doesn't change that.

[TheMeInTeam]

Just make it MR-dependent and be done with the issue.

Doing this is very close to the same thing as removing it and trashes stabbers quite a bit. Is there a reason stabbers are too strong right now, relative to alternatives, even considering CT? I haven't seen evidence of numbers supporting that conclusion.

Having said all of that crap, Hellmonk is right, a monsters resistance to the spell being dependent on entirely hidden from plain sight numbers and mechanics is just objectively bad, the power level of the spell aside. And just because i think MR: immune is overused doesn't change that.

With confusing touch active you actually can xv monsters in game and see confuse odds from each attack. That you can't do this in advance is a problem. Unfortunately this problem is far from unique to confusing touch. Still, compared to needing to fsim weapons to actually know which is better w/o extensive trial and error it's substantially more convenient.

[bel]

It seems to me that many of the posts here are arguing backwards. They are giving me reasons why confusing touch is useful, but not why it should exist in the first place.

Here's a clarifying thought experiment: Suppose it didn't exist, and someone proposed a lvl 1 Hex which ignored MR, almost doesn't care about spellpower, confused yaks with 80% chance per 10 auts in melee, would anyone accept it?

For comparison, corona requires training Hexes, checks MR, checks spellpower, doesn't confuse a monster and has a lower chance to land (per 10 auts). It is not melee range, but you after casting corona you typically melee the monster anyway.

Also, the commit which changed confusing touch expresses the concern that the spell might be too powerful. But it simply removed the check for unarmed (which typically means a lowering of delay) without any compensation -- which makes the spell even more broken.

Yes broken things can be fun, but that's not a reason for the things to exist.

[petercordia]

Corona should be removed or buffed.

If someone proposed CT as a way to allow stabbers to deal with high-MR enemies, I would be in favour. As I said earlier, I think it's a great spell.
It is too powerful, and I think increasing its spell level would be sufficient to fix this.
If it cost 3 spell levels to memorise, it wouldn't be worth it for most characters (I think).

[TheMeInTeam]

It seems to me that many of the posts here are arguing backwards. They are giving me reasons why confusing touch is useful, but not why it should exist in the first place.

I have stated such reasoning multiple times.

Yes broken things can be fun, but that's not a reason for the things to exist.

You've failed to establish that the spell is broken, or so far to even address counter-arguments. Some case can be made for pushing it above level 1. I'd be okay with it at level 3 for example, even though this wouldn't change how it is used or who uses it much.

The confuse duration is short and using confusing touch is strictly dominated by alternative spell or non-spell choices on the vast majority of builds even at level 1.

The same fully general reasoning you're giving can be used to claim air strike is too good "because it can kill almost anything and is only level 4". Yes, it can. But there's a reason you don't usually see players trying to kill Cerebov using exclusively airstrike. Same goes for CT on most builds.

[bel]

This is again backwards reasoning.

The spell is called "confusing touch". If it was meant for short-blade specialists, why wasn't it called "confusing hit" and why did it only trigger on unarmed attacks (prior to 0.23)? I highly doubt that whoever introduced this spell was thinking "man, stabbers are short-changed against high MR enemies, let me create a spell to help them out".

What happened was a change in 0.23 which made the spell roughly twice as effective for short blade users and much less annoying to actually use, so people started using it with short blade specialists more. That doesn't mean that "improving stabbing playstyle" is the spell's purpose; it just means that a broken spell's mechanics work out that way.

This is a very old spell. I am not sure when it was introduced but I think it may have been present in 0.8. DCSS had similar kinds of "brand" spells. Confusing touch used to be a lvl 2 spell which temporarily brands your fists with confusion, similar to lvl 2 spell "Fire Brand" (which temporarily branded your weapon with fire). In 0.15 most of these spells were removed (though Excruciating Wounds still exists for some reason).

[TheMeInTeam]

The spell is called "confusing touch". If it was meant for short-blade specialists, why wasn't it called "confusing hit" and why did it only trigger on unarmed attacks (prior to 0.23)? I highly doubt that whoever introduced this spell was thinking "man, stabbers are short-changed against high MR enemies, let me create a spell to help them out".

Regardless, stabbing was and is its only real use case after the very early game. Removing it creates a hole in an offensive style to no apparent purpose.

it just means that a broken spell's mechanics work out that way.

You've still failed to establish that the spell is "broken" in any way whatsoever, conveniently not even addressing that multiple people pointed out that in many scenarios (including nearly all scenarios for a majority of builds) casting this spell is a mistake/strictly dominated strategy.

For other builds, it is useful.

The reasoning stated is that that it actually does serve a function right now, and there has yet to be a refutation of that...or any clear reasoning for why removing CT makes crawl better. All it does is remove a *situationally* useful spell and nerf one of the offensive styles broadly considered weak.

Or are you going to try to make a serious case that having a 40% chance to confuse a gold dragon in > 1 aut while wielding say a broad axe or demon whip is a superior move to simply hitting the dragon with your broad axe or demon whip? For 2h weapon users and blaster/summoner/elemental spellcasters it's even more of an obvious failure move to cast CT.

I don't care what CT did or looked like in 0.8. Your proposal is to arbitrarily remove it now, and right now its main practical use case is to allow stabbers a (relatively weak) means to deal with most MR/poison immune enemies in 1v1 scenarios.

OP reasoning is "this spell is absurd". Applying that standard:

- Draconians are "absurd".
- Orb of destruction is "absurd".
- Air strike is "absurd".
- Bolt of Magma is "absurd". It doesn't even completely check rF! Why not?

So per stated logic all of these should just be removed. It's not clear why Crawl is better for doing so, but we could even use the "backwards reasoning" refutation here and suggest that "bolt of magma" and minotaurs wouldn't be approved if proposed today. Too similar to Hill Orc, all it does is add an Aux attack and cast a little worse!

[Siegurt]

This is again backwards reasoning.

The spell is called "confusing touch". If it was meant for short-blade specialists, why wasn't it called "confusing hit" and why did it only trigger on unarmed attacks (prior to 0.23)? I highly doubt that whoever introduced this spell was thinking "man, stabbers are short-changed against high MR enemies, let me create a spell to help them out".

What happened was a change in 0.23 which made the spell roughly twice as effective for short blade users and much less annoying to actually use, so people started using it with short blade specialists more. That doesn't mean that "improving stabbing playstyle" is the spell's purpose; it just means that a broken spell's mechanics work out that way.

This is a very old spell. I am not sure when it was introduced but I think it may have been present in 0.8. DCSS had similar kinds of "brand" spells. Confusing touch used to be a lvl 2 spell which temporarily brands your fists with confusion, similar to lvl 2 spell "Fire Brand" (which temporarily branded your weapon with fire). In 0.15 most of these spells were removed (though Excruciating Wounds still exists for some reason).

I would say that arguing about what a spell was "meant for" us purposeless in the extreme, who cares *why* it was created a decade or more ago (much less what they chose to name it), the question is how it fits in now, and if it's a good thing to have in DCSS, and if so what's broken about it, and can we fix it if so.

I'm quite certain that the reason it's called "Confusing touch" is because there was a spell in D&D by the same name and the creator was trying to make something inspired by that, it's not more complicated, as with many of the old features there wasn't so much "game design" as "let's stick this in because I think it sounds cool" some of those features work well, and some of them are awful, that's really the only question we should be concerned with.

The change in .23 was, on the other hand, clearly to make the spell less annoying, and has had some side effects that were probably unintended (I recently had a game with fencer's gloves on a crossbow user, so I would cast confusing touch every time someone came close so my ripostes would confuse, since awkward riposte bashes with my crossbow were useless anyway, and firing bolts flat out ignores confusing touch)

[bel]

[This was written prior to Siegurt's post]

By "broken", I mean "too powerful". It seems to me that you agree that it's too powerful, because you don't have any problems with tripling its spell level.

As for whether it should exist, I generally work on the basis that the burden of proof should be on the person who wants to introduce a feature, not remove a feature. I tried to give arguments that nobody knows why the spell exists and why it works the way it does. That, to my mind, is sufficient justification to remove it.

Incidentally, I have played hundreds of "stealthy stabbers" and I generally manage to kill (or avoid) high MR enemies just fine. God abiities are the easiest way. Branching out into necromancy or summons also works.

[TheMeInTeam]

As for whether it should exist, I generally work on the basis that the burden of proof should be on the person who wants to introduce a feature, not remove a feature. I tried to give arguments that nobody knows why the spell exists and why it works the way it does. That, to my mind, is sufficient justification to remove it.

You're still simultaneously arguing to remove minotaurs and bolt of magma.

Incidentally, I have played hundreds of "stealthy stabbers" and I generally manage to kill (or avoid) high MR enemies just fine. God abiities are the easiest way. Branching out into necromancy or summons also works.

Yet you're still not establishing basis for nerfing stabbers in particular, or refuting that (in most cases) stabbers are the only build to meaningfully benefit from the spell.

The fencer's glove interaction is odd. Making the spell level 3 will barely change its usage, but does put it on level with other common sources of stabbing (confuse, dazzling spray, mephitic cloud) while keeping its niche. Of these options mephitic cloud is still the strongest confuse source vs non-rPois enemies by a wide margin, because unlike CT you can actually damage the enemy in the rounds after casting it and have massively higher per-aut confuse odds the moment the clouds touch more than one monster that is susceptible. Enough that it dominates CT even while CT is level 1 in most cases. Similarly, for anything without huge MR or immunity confuse > confusing touch, because you can cast it without moving adjacent to the monster or waiting for it to move adjacent to you.

But CT level 3 is still fine, it's not like that will put serious mana pressure on stabbers or damage its niche, so if others agree the spell is too strong that's not a bad spot for it in my mind.

[bel]

Bolt of Magma sucks. I can't remember the last time I learnt it, so I wouldn't even notice if it was removed.

If you had compared it to Fire Storm, it would make sense. Fine, let's make confusing touch lvl 9!

Anyway, joking aside, about the stabber stuff, it seems to me that you consider stabbers weak and this change in 0.23 (which had absolutely nothing to do with stabbing per se -- it was supposed to make the spell less annoying to use) buffed stabbers.

I do not think this is the correct way to analyze mechanics. Any mechanic at all can be justified on the basis of "X is too weak, this spell helps X". The solution to fix the power level of X is to fix the power level of X directly. It should not be done with an accidental by-product of an unrelated change.

[TheMeInTeam]

I don't see how you can claim it's "unrelated" when it's one of the only build setups to use the spell at all.

But okay, we can focus on the spell itself. Again, you're not refuting that in most cases, using confusing touch is a *strictly* dominated option. However, to make a serious case that the spell is "too strong*, you MUST either refute that assertion or demonstrate that the outcomes in the narrow range of uses where it isn't dominated are not in line with normal expected outcomes in crawl. No post in this thread has done either of those things yet.

Absent such justification, there really is no clear difference between saying "confusing touch is too strong" and "bolt of magma is too strong".

[bel]

I used confusing touch on my transmuters back when it was unarmed only. So, it is not true that "only stabbers benefit". Anyone can benefit from it. It was just much more annoying to use back then, so few people used it consistently. Just as Beastly Appendage is very powerful (especially with slaying) but very annoying to use.

As for demonstrating that confusing touch isn't "strictly dominated", lay out some criteria of what you consider "broken" and I'll attempt it. To my mind, an 80% chance to confuse a yak for a lvl 1 spell is "broken".

[TheMeInTeam]

"I used X" does not imply X actually helped. The problem with using CT on a transmuter, then and now, is that transmuters have > 40 damage when they hit something late game, and still > 30 the moment they have blade hands (usually available by lair). If you're swinging fast enough to have 2 CT attempts/aut, you're also swinging fast enough to hit the target with blade hands or dragon form twice/aut...and to hit very hard when you do so.

In what scenario is a 40% chance of confusion better than a much higher percentage chance of killing the enemy outright in the same timeframe?

Some factors for considering a spell/option too strong:

- It's a "no brainer" (similar to the dev reasoning --> any time you have this option it is always beneficial to select it)
- Similar to above, except that it's just too inexpensive/has too little opportunity cost
- It trivializes/centralizes the game --> alternatives you might otherwise have to use are unimportant, because this option will suffice in the same scenarios and more generally

"I can confuse a yak at 80%" is meaningless by itself. It doesn't address how long the confusion lasts, what you can/can't do to the yak after it's confused, or the fact that you just spent, on average, > full turn doing no damage for the chance to (more briefly than with other sources) confuse something that is immediately adjacent to you and possibly not alone. Even in THIS case, mephitic cloud can strictly dominate CT depending on the specifics of the encounter, while CT is a better option against the rPois enemies (though a pretty bad idea vs hydras!).

Again, which is better:

- Hit the yak 1v1 with a +3 dire flail at < 1 aut per swing
- Cast CT and try to confuse the yak with a +3 dire flail, then once it's confused keep hitting it as it wanders around a bit or sometimes attacks you anyway

You're saying the level 1 spell is too good here. I'm not convinced it's actually better than the level 0 spell of "cast nothing and just kill the yak".

If you're swinging a dagger of speed instead, then CT is useful against a yak if you don't have mephitic cloud or confuse. CT is not *generally* useful, even now, even with a dagger of speed.

[bel]

So you think that I should "git gud" and was simply wrong to use confusing touch? As it happens, your calculations are wrong. It is easy to find situations where confusing touch isn't "strictly dominated".

An average swing (taking into account chance to miss etc.) with a blade hands character in say, Spider, would be something like 20-25 damage (easy to check with fsim). A confusing touch chance of 40% would require about 2.5 hits to land (in expectation) -- confusing touch has a +to-hit bonus and I'm too lazy to figure out how much bonus, but it usually makes contact most of the time. Since an enemy is almost incapacitated when confused, if an enemy takes more than 2.5 swings to go down, it is always optimal to confuse it before taking it down. Add in some fudge factors (like a confused enemy still attacking you, or confusion lasting not long enough before you kill the monster), if the enemy takes more than 3-4 swings to go down (which translates to something like 60-80 HP), it is not a "strictly dominating" strategy to simply hit it before confusing it.

Also, in your calculation, you disregarded a very important case of confusion: spellcasters. Confused monsters can't cast spells or use wands.

Several minor things are also wrong or under-rated. Blade hands is a lvl 5 spell, confusing touch is a lvl 1 spell. Also, confusion stabs, monsters hitting themselves, monsters drowning etc.

[braveplatypus]

The Reason to keep confusing touch: It's a balanced way for stabbers to confuse not top-end Hit Dice MR - Immune monsters, a subset of monsters that has no right to be as populated or as totally hexproof as they are.

The Reason to get rid of or rework confusing touch - monster HD is never shown to the player, the mechanics of what a monsters HD does is never explained to the player, so having a spell that interacts with monster HD is bad, even if after casting the spell you can see your odds of an individual monster, you have no idea why the odds are what they are for said monster without closing the game and looking elsewhere.

As far as "arguments" go bel's taking the wrong angle, the spell is not too powerful at all, the chance of it working is not too high, it's not too safe to use, or any of that. But it's poorly designed, at least partially.

I'm not convinced either that showing and explaining monster HD to new players for these few things that use it is worth it, i mean if they are going to stay in the game it definitely is, but i'm not sure they should.

[TheMeInTeam]

So you think that I should "git gud" and was simply wrong to use confusing touch? As it happens, your calculations are wrong. It is easy to find situations where confusing touch isn't "strictly dominated".

If I wanted to say that I'd have said it. How good either of us are at the game doesn't matter. What matters is whether the move is actually the best one you can make or not.

if an enemy takes more than 2.5 swings to go down, it is always optimal to confuse it before taking it down.

What basis are you using to claim "optimal". You are spending more turns, and without heavy investment in stealth you're probably not adding damage that can compensate the extra time spent even with "stabs". A tier 3 stab on non-SB is +30% damage if your UC/stealth average is 15. You need an unrealistic average of 24 just to get 50% more damage from a confusion stab with UC, and it isn't a guaranteed proc.

A wolf spider (decent median monster for spider) has 36 to 74 hp. If we're doing 20-25 damage, we can expect to kill it in 2-4 hits. You're claiming that CT is worthwhile if the enemy takes more than 2.5 swings to go down, but this isn't realistic, because with CT you still have to do another ~2-4 hits to kill it. You're spending 4.5-6.5 to confuse and kill a monster that you stood a realistic chance to kill in those same 2.5 hits. Maybe more, if it moves away.

Yes, drowning, spellcasting, and other considerations can factor into whether CT is dominated. In practice, it's usually dominated. Spellcasters don't tend to have a ton of HP, and killing them also prevents spellcasting...with less odds of them rolling lucky to break out of confusion quickly and get extra casts or having a bad streak of not getting confuse.

This gets increasingly worse as you add more enemies in sight, since turns you spend confusing target are turns you're not killing something, but everything else can act both while you attempt the confusion and as you follow up on it. CT already looks bad 1v1 vs a wolf spider, but it starts looking terrible when you have 2 wolf spiders, 3 redbacks, and a tarentalla in view. It's also bad against ghost moths and spark wasps with UC, but you can make a case when doing 1v1 with emperor scorpions.

Several minor things are also wrong or under-rated. Blade hands is a lvl 5 spell

I mentioned it since you said you were a transmuter, and blade hands is a level 5 spell that is routinely castable by lair for transmuter starts, making it relevant to fighting yaks. If you are fighting 5 yaks, you'd have to cast CT 5 times to confuse all 5 of them (maybe more since it's short duration and UC stabbing will do little extra damage).

It's also not like you're factoring every contingency. Going for confusion with blade hands active could force you to cast blade hands again, since you're burning a lot of extra turns trying to confuse each yak. Suddenly you're burning more food (usually trivial), more MP (not always trivial) and more time (also not always trivial) for an outcome that's at best equal.

In most cases, casting CT with blade hands is a waste of time. Not all cases, but most. In those cases, where it is strictly dominated, it's difficult to make a case that the dominated option is "too strong". This continues to hold even when not using UC though. Not just for dire flail example, but for other weapons too. Not to mention it does absolutely nothing for launchers (absent the artifact interaction) or spellcasting.

~~~

You get good value with short blades because short blades are already at ~75% extra per at 14 skill average, and add up to 30 damage before even applying that multiplier. This is an enormous damage spike compared to swinging a short blade without stabbing. The added output for other weapons is minimal, and CT might cause you to take more damage rather than less with standard weapons. It's a situational spell.

[petercordia]

So you think that I should "git gud" and was simply wrong to use confusing touch? As it happens, your calculations are wrong. It is easy to find situations where confusing touch isn't "strictly dominated".

An average swing (taking into account chance to miss etc.) with a blade hands character in say, Spider, would be something like 20-25 damage (easy to check with fsim). A confusing touch chance of 40% would require about 2.5 hits to land (in expectation) -- confusing touch has a +to-hit bonus and I'm too lazy to figure out how much bonus, but it usually makes contact most of the time. Since an enemy is almost incapacitated when confused, if an enemy takes more than 2.5 swings to go down, it is always optimal to confuse it before taking it down. Add in some fudge factors (like a confused enemy still attacking you, or confusion lasting not long enough before you kill the monster), if the enemy takes more than 3-4 swings to go down (which translates to something like 60-80 HP), it is not a "strictly dominating" strategy to simply hit it before confusing it.

You seem to describe a 1-on-1 fight in spider, with a transmuter. Unless there was a really buff unique in there, you shouldn't have had any trouble with 1-on-1 fights, so that memorising a spell to make 1-on-1 fights even easier seems moslty pointless.
That said, if it only costs 1 spell level and it increases the odds of surviving Asterion I'd also do it. Because 1 spell level is almost nothing.

[petercordia]

I think Monster HD is usually intuitive. Monsters get tougher as you go deeper. Tougher monsters take longer to confuse (by CT or Mephitic Cloud) and hit harder. I'd only need to know more if I wanted to achieve hyothetical optimal play.

[Sprucery]

This is a very old spell. I am not sure when it was introduced but I think it may have been present in 0.8.

Heh, it was already in Linley's Dungeon Crawl, for sure in 4.00 alpha 24, and I'm pretty sure in 3.30 and 3.20 as well (too lazy to check).

Now that I think of it, it should be changed back to only working with unarmed attacks. It was like that for 20 years and I think it was better that way.

[duvessa]

I prefer the current version, swapping is annoying.

[Sprucery]

Ok so return it as it was and add a restriction: if you wield a weapon, the confusion is immediately cleared.

"As you wield the morningstar, the yak regains its senses."

[gameguard]

The spell is absurdly overtuned. Anyone who says its not overpowered never played around with it. You can pretty much kill every single thing except OOF, Alich and tentacle monstrousity in a 3rune game. And no, there isn't much danger when you funnel monsters around a corner or through a choke point and have passable defenses.

With that said, this is the only thing that makes playing a stabber even bearable. Increasing spell lvls won't do anything. Decreasing the hit die cap slightly might be reasonable.

[stormdragon]

edit - accidental double post

[stormdragon]

bel, you propose to remove Confusing Touch, and your first sentence is

Confusing touch is an extremely weird spell.

This is a reason to KEEP Confusing Touch. We want spells to be unique and interesting; if it wasn't weird at all and was easily comparable to some other spell, that would be a reason to remove it (as should be done with Bolt of Cold or Bolt of Fire, for example).

Here's a clarifying thought experiment: Suppose it didn't exist, and someone proposed a lvl 1 Hex which ignored MR, almost doesn't care about spellpower, confused yaks with 80% chance per 10 auts in melee, would anyone accept it?

Not worded that way, but if someone proposed a buff that turned your melee attacks into hexes, yeah that sounds like it might make it. It might be simplified into a range-1 hex spell, which would definitely be interesting, and I wouldn't mind that as the future of Confusing Touch.

Your argument seems to be based on how overpowered it is, but as you should know, poor balance does not equate to poor design. The only suggestion you have for improving the spell's design is to change it from HD-based to MR-based, which I agree would be an improvement. This sounds relatively simple to implement and I bet a pull request to do this would have a high chance of acceptance. Balance can be re-evaluated after that.

[TheMeInTeam]

Unread post Today, 10:17
The spell is absurdly overtuned. Anyone who says its not overpowered never played around with it. You can pretty much kill every single thing except OOF, Alich and tentacle monstrousity in a 3rune game. And no, there isn't much danger when you funnel monsters around a corner or through a choke point and have passable defenses.

So nearly every weapon and spell in the game is overpowered? You can kill enemies in zot with a hand axe or the searing ray spell if you're literally always capable of doing this.

Your argument seems to be based on how overpowered it is, but as you should know, poor balance does not equate to poor design.

Assertions that it is overpowered in this thread have repeatedly ignored refutations of that. We're still at a stage where the standards being used to assert something is "overpowered" logically imply removing basic species and spells like stone arrow, searing ray, or similar "because they can kill gold dragons".

Ok so return it as it was and add a restriction: if you wield a weapon, the confusion is immediately cleared.

"As you wield the morningstar, the yak regains its senses."

The opening post already suggested removing it from the game, which is basically the same thing.

[bel]

Assertions that it is overpowered in this thread have repeatedly ignored refutations of that. We're still at a stage where the standards being used to assert something is "overpowered" logically imply removing basic species and spells like stone arrow, searing ray, or similar "because they can kill gold dragons".

Actually I made quite a bit more involved argument than that: I compared two strategies and tried to show that confusing touch is the superior strategy. You obviously didn't buy my argument, but to say that nobody made the argument is disingenuous (to say the least). Please don't do that.

My argument above was abbreviated and skipped quite a few steps, so it's possible that it's not convincing. I am willing to try again with a more elaborate setup, using fsim to get hard numbers. But before I do that could you answer a few questions of mine?

• Before 0.23 did you play Crawl and did you use confusing touch (on stabbers or anything else)?
• Have you used Confusing Touch on any Transmuters (before or after 0.23)?

[svendre]

I'm not going to make any eloquent arguments about this because a lot has already been said. I'll just add that it's been my experience that when trying to play *awful* combinations that leave you scratching your head how you can possibly survive, when you find confusing touch you have a viable path. There is no question in my mind that it is an extremely powerful spell, possibly one of the best crutches the game has to offer. Whether or not it should be changed/nerfed, I don't care to get into that much, but probably.

[TheMeInTeam]

Before 0.23 did you play Crawl and did you use confusing touch (on stabbers or anything else)?

I started playing in August 2018 and stayed on trunk, so I have (almost) no pre-.23 experience. I say almost because for a cosplay challenge or so I have dipped into earlier versions, but this is not nearly enough experience to be meaningful and I don't recall what I used (I wound up submitting on current version for that one anyway).

Have you used Confusing Touch on any Transmuters (before or after 0.23)?

Yes. I am unimpressed with confusing touch on transmuters in practice. The stab damage is unimpressive, CT often takes a few swings to apply, and the monster moves around, sometimes out of reach or in a way that it's obstructed by another monster, wasting all the time establishing CT and confusing the monster.

I have used CT a lot, actually. When I used it on non-short blade builds I stop using it, because it consistently wastes time in practice. When I used it on short blade builds, I keep using it because the increase in damage from stabbing is much greater relative to simply hitting the monster so I will tolerate the above contingencies. I don't even bother with it on blasters or launchers, for obvious reasons.

It's in an odd place because when it's useful at all, it's a very strong benefit. However it's less consistently useful than other level 1-2 spells on average across build types. Animate skeleton, apportation, summon butterflies, blink, passwall, and swiftness are all more "no brainer" memorize options over CT because they help a larger percentage of different builds. Many of the arguments regarding CT actually apply in greater force to animate skeleton, while the others are more "just" utility spells that are consistently useful for any character that's not worshipping Trog.

[duvessa]

The existence of other broken level 1 spells isn't a reason to leave confusing touch alone. It's a reason to fix those spells too.

[TheMeInTeam]

The existence of other broken level 1 spells isn't a reason to leave confusing touch alone. It's a reason to fix those spells too.

But what standard are we using to claim something is "broken"? Some consistency in this evaluation would be useful.

"Sometimes this is better than alternatives" isn't what I'd define as "broken". To me the obvious candidates are no-brainers w/o significant opportunity costs, not magic that is straight up mildly harmful compared to casting nothing in many cases.

And I have serious doubts that fsim will conclude CT is useful for non-stabbers in any scenario where more than 1 monster is on the screen + confused monster can move away. Maybe I'm wrong, hard data would be better than anecdotes for stuff like this.

[Aean]

But what standard are we using to claim something is "broken"? Some consistency in this evaluation would be useful.

"Sometimes this is better than alternatives" isn't what I'd define as "broken". To me the obvious candidates are no-brainers w/o significant opportunity costs, not magic that is straight up mildly harmful compared to casting nothing in many cases.

I'd say it'd be reasonable to call something broken if it has notable long-term utility for a wide range of character types at negligible cost, without a meaningful limiter, when it is not intended for such a purpose. Thus, potions, whose purpose is just that, and whose use is limited, would not be broken. In contrast, a spell that could easily be cast by a dumb melee brute in heavy armor, and would be valuable throughout the entire game, would be. Or to put it in other terms, things that are "no-brainers" are usually broken in one respect or another.

As you said, animate skeleton, apportation, summon butterflies, blink, etc. are also good examples; they are good for every character type, valuable through all 15 runes (though admittedly animate skeleton drops off a bit in some extended endgame areas) and a very minor cost compared to the benefits they provide.

As Duvessa points out, just because these spells could probably also use some revisiting, doesn't necessarily mean that confusing touch is where it should be.

[petercordia]

I think another definition of 'broken' would be more useful. Blink is a good spell. Almost everyone can cast it and it is useful for everyone, but the devs want everyone to be able to blink. That's why there are items with evocable blink. Being able to blink makes the game more fun. Not being able to blink is a challenge you must work around, which can also be fun.
Summon butterflies, apportation, and passwall are also good examples of spells which are powerful for their level, but not broken.

I would considder a spell 'broken' if a significant number of players find DCSS less fun because of the existence of the spell. I find Crawl less fun if I always end up playing similarly to make use of particular overpowered spells - for example, because I almost always invest in Passage of Golumbira (PoG) if I find it. Some players have this same isue with Animate Skeleton (I don't, because I can't be bothered to use AS, and ignore its existence).

I think CT is currently broken (following this definition) - because a number of players find DCSS less fun because they think it's almost always optimal to memorise CT. From that point on, you're either making all your games more similar, or delibarately playing suboptimally, neither of which is fun.
To be clear: I am one of these players. I don't want to memorise CT on a DEFE, but it feels wrong not to in the current balance.

Fixing CT shouldn't be too hard. I love it's quirkiness

[TheMeInTeam]

But what standard are we using to claim something is "broken"? Some consistency in this evaluation would be useful.

"Sometimes this is better than alternatives" isn't what I'd define as "broken". To me the obvious candidates are no-brainers w/o significant opportunity costs, not magic that is straight up mildly harmful compared to casting nothing in many cases.

I'd say it'd be reasonable to call something broken if it has notable long-term utility for a wide range of character types at negligible cost, without a meaningful limiter, when it is not intended for such a purpose. Thus, potions, whose purpose is just that, and whose use is limited, would not be broken. In contrast, a spell that could easily be cast by a dumb melee brute in heavy armor, and would be valuable throughout the entire game, would be. Or to put it in other terms, things that are "no-brainers" are usually broken in one respect or another.

As you said, animate skeleton, apportation, summon butterflies, blink, etc. are also good examples; they are good for every character type, valuable through all 15 runes (though admittedly animate skeleton drops off a bit in some extended endgame areas) and a very minor cost compared to the benefits they provide.

As Duvessa points out, just because these spells could probably also use some revisiting, doesn't necessarily mean that confusing touch is where it should be.

I continue to dispute that CT is actually a good idea on "melee brutes in heavy armor". These guys aren't usually swinging daggers of speed.

I think CT is currently broken (following this definition) - because a number of players find DCSS less fun because they think it's almost always optimal to memorise CT. From that point on, you're either making all your games more similar, or delibarately playing suboptimally, neither of which is fun.

Thinking something is optimal vs something being optimal is not the same thing. A blaster can kill most monsters in 2-5 turns with direct damage spells. A blaster cannot kill most monsters in 2-5 turns with confusing touch. They will often enough still have a full health, unconfused monster in the same timeframe they could have killed it.

The assertion is that CT is always useful does not seem to bear out to me. If you're not getting bonus stabbing damage it seems like a waste of time. Do we have data of any variety supporting that this spell is actually too good? If not, we're STILL in the realm of claiming that magma bolt should be removed because it's too strong.

I'd say it'd be reasonable to call something broken if it has notable long-term utility for a wide range of character types at negligible cost, without a meaningful limiter, when it is not intended for such a purpose.

I hold that melee range and doing no damage while attempting to apply the spell are "meaningful limiters". Meaningful, as in they are both significant enough drawbacks that using the spell is often strictly dominated for a majority of builds. When using the spell makes you worse off, how is that not a "meaningful limiter"?

[petercordia]

The problem is that CT is so cheap to memorize. Only 1 spell level. And if you just do happen to meet a unique who is all alone, and too tanky to kill normally, and you can't run away normally, then it will be a useful spell. It rarely happens, but for every character a scenario is imaginable where CT would be useful.

[TheMeInTeam]

The problem is that CT is so cheap to memorize. Only 1 spell level. And if you just do happen to meet a unique who is all alone, and too tanky to kill normally, and you can't run away normally, then it will be a useful spell. It rarely happens, but for every character a scenario is imaginable where CT would be useful.

"Every" character is a bit of a reach considering that Trog is a god in DCSS and berserker is one of the stronger backgrounds.

Even with Trog aside, how does this differentiate it from a wide swath of items and spells that are supposedly fine? What you're describing puts it in "wand of flame" territory, not even "summon butterflies" territory. That it has situational use cases at level 1 doesn't indict it in terms of utility, especially because it takes at least SOME XP investment to use it with a heavy weapon + armor at better than ~70% or more fail odds, just to (maybe) do something that a potion or common wand could do better/with less training/more useful alternatives alongside it.

I'm not seeing the no-brainer for CT. For animate skeleton, yes. Anybody worshiping a non-good/fedhas god can simply cast that rather than butchering and have almost no detriment with a significant potential upside. Using it virtually never hurts you. CT just isn't like that. Using it thoughtlessly will cost you a lot more than 1 MP and a spell slot.

[svendre]

The problem is that CT is so cheap to memorize. Only 1 spell level. And if you just do happen to meet a unique who is all alone, and too tanky to kill normally, and you can't run away normally, then it will be a useful spell. It rarely happens, but for every character a scenario is imaginable where CT would be useful.

"Every" character is a bit of a reach considering that Trog is a god in DCSS and berserker is one of the stronger backgrounds.

Even with Trog aside, how does this differentiate it from a wide swath of items and spells that are supposedly fine? What you're describing puts it in "wand of flame" territory, not even "summon butterflies" territory. That it has situational use cases at level 1 doesn't indict it in terms of utility, especially because it takes at least SOME XP investment to use it with a heavy weapon + armor at better than ~70% or more fail odds, just to (maybe) do something that a potion or common wand could do better/with less training/more useful alternatives alongside it.

I'm not seeing the no-brainer for CT. For animate skeleton, yes. Anybody worshiping a non-good/fedhas god can simply cast that rather than butchering and have almost no detriment with a significant potential upside. Using it virtually never hurts you. CT just isn't like that. Using it thoughtlessly will cost you a lot more than 1 MP and a spell slot.

Kind of funny we see eye to eye on a lot of stuff, but I consider animate skeleton pretty crappy but CT very powerful.

[braveplatypus]

Kind of funny we see eye to eye on a lot of stuff, but I consider animate skeleton pretty crappy but CT very powerful.

Oh well that makes everything clear, we're talking about different video games.

[TheMeInTeam]

It'd be more surprising if any two people agreed on everything lol.

Animate skeleton is nice and can mostly carry lair/orc/D:15. The skeletons hit other monsters pretty hard and this scales decently until you enter areas like spider (can't raise many enemies) or elf (skeletons do no damage for most elves). They're also a bit weaker in other S branches (questionable corpses in swamp, ranged spam just killing skeletons in the other two).

I'd be willing to change my mind regarding CT if someone could post properly-controlled wizmode data on it. IE factoring that confused monsters wonder, measuring typical times to kill using weapons like great mace, battleaxe, executioner's axe, broad axe, demon whip type weapons. I have my doubts that this helps kill monsters faster for non-short blades, and that's a substantial part of the argument from others that CT is too strong. I think we're mostly in agreement that for blasters and launchers it's not helpful, and that it's consistently good for stabbers.

As a bit of an extra anecdote, I basically never see the good youtubers/streamers that I've watched using it on non-short blades, and unlike lightning spire the explanation for that isn't because it's too strong/always optimal :p.

[Siegurt]

It'd be more surprising if any two people agreed on everything lol.

Animate skeleton is nice and can mostly carry lair/orc/D:15. The skeletons hit other monsters pretty hard and this scales decently until you enter areas like spider (can't raise many enemies) or elf (skeletons do no damage for most elves). They're also a bit weaker in other S branches (questionable corpses in swamp, ranged spam just killing skeletons in the other two).

I'd be willing to change my mind regarding CT if someone could post properly-controlled wizmode data on it. IE factoring that confused monsters wonder, measuring typical times to kill using weapons like great mace, battleaxe, executioner's axe, broad axe, demon whip type weapons. I have my doubts that this helps kill monsters faster for non-short blades, and that's a substantial part of the argument from others that CT is too strong. I think we're mostly in agreement that for blasters and launchers it's not helpful, and that it's consistently good for stabbers.

As a bit of an extra anecdote, I basically never see the good youtubers/streamers that I've watched using it on non-short blades, and unlike lightning spire the explanation for that isn't because it's too strong/always optimal :p.

Actually it's dumb, but frequently optimal, on a blaster to cast it while wielding a long blade and continue to blast as normal, sometimes you will confuse-riposte stuff for no additional turns, and no additional XP investment, and a miniscule 1 MP cost. The only time this isn't a good idea is if your weapon slot is occupied by something more useful than something that will occasionally confuse things for you for free.

For Launcher users it's obviously only useful if you have the aforementioned weird fencer's glove interaction.

Overall, it'd be hard to generate reasonable wizmode data with no explicit parameters given, as "too strong for it's level" when you don't have any idea what level we're talking about, or what kind of skill training, or what kind of equipment we're even comparing here. On the one side of the coin we have short blade users, where it's clearly optimal, and on the other we have the "has a maxed out very powerful two handed weapon that 2-3 shots most things" where it's clearly not, but there's a very large space in the middle where it's "kinda maybe a good idea some or a lot of the time" I've seen it reiterated several times that the duration is too short to make it useful, but the duration of the confusion isn't really all that *short* either, 3-5 turns of confusion when it takes me 6-8 hits to kill something (Say I'm a middling-trained one-handed weapon user) particularly when the something in question can do significant damage to me in that time period, and can't or is much less likely to when confused, is still pretty good. It's highly situational and the situations in which it has any use get less frequent the later you are and the more powerful you are, the question is "is the slope at which it decreases utility sufficient" Which has a lot to do with how powerful of a character you're playing is.

So for example, I whipped up a Wizmode human with a +9 giant mace of flaming, and checked confusing touch against a stone giant, 26% to confuse, confuses for 3-5 turns, average damage has me killing the stone giant in around 3 turns (this is with min delay on the giant mace)

It took an average of 3-4 attacks to confuse it (it should be a little higher than that, I rolled well I suppose) and the average number of turns it was confused was 3 or so. Even if I rolled very well on the confusion action table (the stone giant never attacked me while confused, and never moved out of melee range) It still took me an average of 3 turns of incoming damage to kill a stone giant whether using confusing touch or not. In the non-optimal "usual" case the confusing touch is going to lose. (There's also the silly, but also possible case where the confused stone giant does nothing but attack *itself* each turn, which while rare, would of course be even more advantageous for CT)

A counter example, using +9 demon whip of flaming (again with min delay) against a stone giant with a shield, it still takes the same amount of attacks to confuse, but that same number of attacks only takes 2 turns, further, my damage takes about 4 turns to kill the stone giant rather than 3. So on average it takes 2 turns of damage of damage in the optimal case using confuse touch, vs 4 turns without it, probably around 3 turns of damage in the "usual" case.

There is such a large range of variance that even these two specific examples could be run to opposite conclusions with minimal tweaking of variables, or even just running them several times, so take even these specific examples with a grain of salt.

Both of these disregard the additional "stab" damage you get sometimes because it only happens sometimes, and isn't of meaningful impact when it does, so ignoring it probably doesn't change much in the bottom line, given the ridiculous variance we're dealing with here.

The only general conclusion I can make is "if you already have a really really top-notch offense, CT is almost certainly worse than just attacking, but if you have a sub-par offense it might be better, even later in the game when you might not expect it to be"


For reference the build I used:

  Code:

Human, 20 str, 10 int/dex, 20 fighting, platemail, 15 shields, and 20 M&f for the great mace, and 12 M&f for the demon whip, 6 spellcasting, 8 hexes, ring of wizardry


Personally, my opinion is that I like confusing touch in the later game on creatures that it has a >33% chance of success with, when the critter both has a lot of resilience relative to my offense *and* has abilities that are annoying, but shut down by CT, this isn't a very large number of things, but it's nonzero. I feel like the number of times that it's useful to cast a level 1 spell should approach 0 sometime around the end of the lair/orc, and CT can generate situations where it's beneficial to use it into the vaults/depths, which feels too late for a level 1 spell to me. A good example of something that I feel like it's beneficial to use CT on, when it's probably too late in the game for it to be appropriate might be a cacodemon (36% to CT, more HP than a stone giant) presuming I don't have some way to kill it at range and am using enemies to force it to get next to me. Another might be spriggan druid (43%), while they have far less HP than a stone giant, their EV and Healing means I'm going to swing a lot of times if I just use normal attacks, but if I land a confuse on them, they can no longer evade, and will kill them before the confusion wears off (in addition to shutting down their annoying spells)