Do dodge and -ACC stack?

[Quickshooter]

i was wondering if it's viable to pair up minus ACC debuffs with dodge buffs
do they stack? or miss and dodge are two seperate things in DD?

from intial tests it feels like they are,or RNG was playing tricks on me?

[FancyFishie]

They stack. The chance to hit something is ACC - Dodge, and there is a hidden extra 10% accuracy, so if it says 90% chance to hit, you actually have a 100% chance to hit.

[Quickshooter]

Originally posted by FancyPerson:

They stack. The chance to hit something is ACC - Dodge, and there is a hidden extra 10% accuracy, so if it says 90% chance to hit, you actually have a 100% chance to hit.

so if i stack 3 vapors and two barrages of -ACC and the enemy still hits
it's just RNG right?

[FancyFishie]

Yes. Also, even if they have a large debuff, there's a flat 5% chance (can't remember if it's 5% or 10%) that they will have to hit no matter what.

[Quickshooter]

Originally posted by FancyPerson:

Yes. Also, even if they have a large debuff, there's a flat 5% chance (can't remember if it's 5% or 10%) that they will have to hit no matter what.

damn... means that nothing can be fully avoidable
bah, i guess stacking PROT is a more viable tactic...

[FancyFishie]

In some cases stacking PROT isn't the best thing to do. Dodge tanking with a Houndmaster is typically a better option, for there will be no crits or stuns hitting him as often.

[Quickshooter]

Originally posted by FancyPerson:

In some cases stacking PROT isn't the best thing to do. Dodge tanking with a Houndmaster is typically a better option, for there will be no crits or stuns hitting him as often.

still a gamble though, the enemy might just target someone else
though with the A.I being dumb as hell sometimes, it's a legit viable strat

[Manxome]

Originally posted by FancyPerson:

there is a hidden extra 10% accuracy, so if it says 90% chance to hit, you actually have a 100% chance to hit.

Source?

[FancyFishie]

None really, I just heard it somewhere and it seems true. Think of the times where you've had a 0% chance to hit and you hit the target, and I never seems to miss 90% hits so... let me see if I can find something that says this.

[Cure Seraph]

Originally posted by FancyPerson:

They stack. The chance to hit something is ACC - Dodge, and there is a hidden extra 10% accuracy, so if it says 90% chance to hit, you actually have a 100% chance to hit.

Doesn't that remaining 10% represent the usual chance to MISS every character has?
Which is what the Fated positive quirk turns into a reroll?

[Tailwind]

As far as I know, there's no proof for the "+10% extra accuracy" suspicion more than what FancyPerson pointed out. I made an experiment (possibly it's the source of this suspicion) and found out it's part-fixable, but still not really sure until the codes are revealed. And it doesn't explain differences between Dodge and Miss.

I have no idea how Fated works. Maybe similar experiments can help.

(edit:) For this thread's topic, -ACC Debuff's (and all other buffs/debuffs) duration counts down *before* the target's action, and it disappears right before it attacks. An ACC debuff only affects two of the target's actions, and a DODGE buff has somewhat shorter duration if the target acts right after getting buffed.

[Manxome]

Originally posted by Tailwind:

As far as I know, there's no proof for the "+10% extra accuracy" suspicion more than what FancyPerson pointed out. I made an experiment (possibly it's the source of this suspicion) and found out it's part-fixable, but still not really sure until the codes are revealed.

So, extrapolating from the post you linked, it sounds like a bug where the code is supposed to do this:

var R = random number in range [0, 1)
if (R > maxHitChance), the attack will MISS
else if (R < minHitChance), the attack will HIT
else if (R < regularHitChance), the attack will HIT
else the attack will MISS

...but it actually does this:

var R = random number in range [0, 1)
if (R > cutoffAlwaysHit), the attack will MISS HIT
else if (R < cutoffAlwaysMiss), the attack will HIT MISS
else if (R < regularHitChance), the attack will HIT
else the attack will MISS

Kind of seems like the programmer(s) chose confusing names for the gameplay constants and then couldn't keep straight what they were supposed to do. (Notice that, even ignoring the names of things, this implementation is still wrong because cutoffAlwaysHit is applying a bonus instead of a minimum.)

Originally posted by Tailwind:

For this thread's topic, -ACC Debuff's (and all other buffs/debuffs) duration counts down *before* the target's action, and it disappears right before it attacks. An ACC debuff only affects two of the target's actions, and a DODGE buff has somewhat shorter duration if the target acts right after getting buffed.

Wow. So, hypothetically an ACC debuff with a duration of 1 round would do absolutely nothing? That is ridiculously misleading.

[Tailwind]

Originally posted by Manxome:

Originally posted by Tailwind:

(...)

So, extrapolating from the post you linked, it sounds like a bug where the code is supposed to do this:

(...)

...but it actually does this:

(...)

Exactly, if my theory is actually to be believed. Modifying cutoffAlwaysMiss did make some of "90%" attacks miss, but actual, detailed results (which I have lost and forgot, to my shame) were in fact not close enough to the in-theory values, implying the issue can be even more complicated.

Originally posted by Manxome:

Originally posted by Tailwind:

(...)

Wow. So, hypothetically an ACC debuff with a duration of 1 round would do absolutely nothing? That is ridiculously misleading.

There are some exceptions though; Ripostes and "Bonus Initiatives" actions are not counted as actions and thus can be performed without removing the debuff. Watch the debuff values closely as the enemy acts.

[Manxome]

Originally posted by Tailwind:

Watch the debuff values closely as the enemy acts.

When playing with a gamepad, you cannot examine the debuff values while enemies are acting. When I get hit with an unfamiliar debuff, I have to wait until one of my characters gets a turn before I can check what it does.

[Sir Francis]

Is this kind of stuff modable?
Fixing the acc cap, making buffs/debuffs tick down AFTER the action of a character/enemy, not before and so on.... Are we able to fix this? Maybe that mod would then be called "The inofficial Patch" or smth.

[Tailwind]

ACC Cap is fixable (again if my theory, which I myself suspect is not too precise, is to be trusted), but with some limitations like these.
- The cap has to be "100% - <CtH Lower limit>", so the lower limit must be set to 10%.
- These values are defined in shared/rules.json, so it would be incompatible with other mods which affect it. The same file determines values for walking speed, base Virtue Chance, effects of Torchlight, etc.
- If an attack misses, the next attack (by the same combatant, presumably) gains extra ACC, but this effect cannot be told to the player.

I have no idea how Buff/Debuff's ticking can be fixed by a modder. Probably it's impossible.

... My mod did this sort of stuffs, until I started to "fix" non-bug inconveniences and work around unfixable bugs in game-balance-changing ways.

(I would like options for andmore types of buff durations - ticks right before initiatives are rolled (handy for buffs to SPD/PROT/DODGE/etc, and maybe Guards), ticks based on the distance the party walked (good for some camp/curio buffs), and maybe more.)