About Attack Duration buffs

[auximenes]

Thanks for the discussion Vastin. I love the under-the-hood information.

[Kaeneth]

By a rather surprising coincidence, I was going over much the same issue a couple weeks ago, doing some game archaeology into the skill execution timing system - though in my case I was approaching it from the actual code and data side, in addition to a fair bit of testing. It looks like you have better testing tools than I do.

TL;DR your conclusions appear to be largely accurate.

Skills are defined by an Animation Duration, and a Skill Duration, and they generally use the longer of the two.
- Skill Duration can be modified by attack speed buffs by +/- 40%.
- Animation Duration can be modified by only +/-10% however.
- So if the duration are similar - or animation duration is longer, then the skill will respond very poorly to attack speed buffs.
- Fast attacks always use Animation Duration, so they never respond well to attack speed buffs, ironically.

The system was originally designed with the idea that Skill Duration would always be substantially longer than Animation Duration, but over the years that's simply often not the case any more. With many skills they are the same, or skill duration is a bit shorter than animation duration - so these skills can't adjust much.

There are also some flat out problems with how the math is being done in one or two places, though with the tight clamp ranges currently in place that particular bug is hardly noticeable - but it would be if I loosened the clamps without us fixing that first.

Weapon Type *shouldn't* have any appreciable effect on attack duration these days unless the weapon is giving you a specific buff to Attack Speed, but I'll double check that and make sure.

We do have plans to go in and dust the cobwebs off of all of this and get it modernized and working nicely again, though I don't have a timetable on that yet.

In any case, nice sleuthing!

I'm not terribly savvy on the technical side, but when you guys go over this is there a chance that all races and male/female avatars will end up having the same animation speeds?

[Askelin]

I'm not terribly savvy on the technical side, but when you guys go over this is there a chance that all races and male/female avatars will end up having the same animation speeds?

I believe the race and sex attack speed discrepancies are due to animation differences and not attack duration. Same might be said about 2H weapons as swords/clubs seem to push animations of many skills faster than a hammer or axe. Apparently it's not an easy fix.

A dwarf with a 2H hammer using wild attack is visibly slower than a high elf female with a 2h sword, considerably slower in fact. Beornings also have this problem with man form skills, though it's less considerable as all their skills are 'fast'.

[Krindel]

I wonder why there are skills with different implementation of animation, which are avoiding all these rules. For example Dance of War, it has animation which cannot be cut down by using immediate skill and it's duration is unaffeceted by any attack duration buffs/debuffs.

[Stever1388]

So is it worth slotting -AD%, and is there an amount where it becomes less useful? Or should we be aiming for other stats? I remember back when LIs first came out, the -3% attack duration relic that you could put on your LI was one of the best relics for a champ because the -3% AD was basically a +3% increase in damage, and it stacked additively so there wasn't any diminishing returns. But that doesn't sound like that's really the case anymore, because not all skills benefit from it, or they don't benefit from it as much anyway?

[Gabli]

So is it worth slotting -AD%, and is there an amount where it becomes less useful? Or should we be aiming for other stats? I remember back when LIs first came out, the -3% attack duration relic that you could put on your LI was one of the best relics for a champ because the -3% AD was basically a +3% increase in damage, and it stacked additively so there wasn't any diminishing returns. But that doesn't sound like that's really the case anymore, because not all skills benefit from it, or they don't benefit from it as much anyway?

What you say is absolutely not true, when new LIs came out the speed system was exactly as it is now so -3% AD does not translate to +3% DPS at all. The time when that was true is a few years back, when most skills had a noticeable PSD.
As for the amount of AD after which it becomes useless, it is mentioned and then confirmed in this very thread to be -40%.

[Gabli]

About previous posts here, are you people certain that is there an actual discrepancy in animation or skill duration based on character race or gender ?

I'm really asking because the last posts are claiming so without bringing any sort of data to support the claims while as I wrote in this very thread, it does not look to be the case. That preliminary test was comparing Wild Attack's ACT (or "Animation Duration) without any sort of Attack Duration buff with a 2-handed sword with a Dwarf, a Man (male) and a High-Elf (male).

I have pushed this test on Wild Attack with two-handed swords for Dwarf, Man & High-Elf males with AD buffs ranging from 0 to -10% AD and the results here indicate that race has no effect on animation duration for this skill with this weapon type :




If anyone wants to try another race/gender/skill, ideally they would make the same measurement and then post the findings.

Just remember that to allow meaningful comparison of durations, measurement uncertainty is pretty much required. As a reminder :

• Measure the duration, animation or skill depending on what you're testing, N times. N from 15 to 25 is usually enough to get close to a few miliseconds accuracy.
• Discard measurements that are wildly off (usually more than 10ms away from the most common values, may depend on internet connexion).
• Compute the average "t" of the results.
• Compute the uncertainty "u(t)" of the measurement from the standard deviation (1 deduced degree of freedom) and the sample size :

For example, here is the spreadsheet with the data for Wild Attack's animation duration for a male High-Elf with a two-handed sword :

[Gabli]

I wonder why there are skills with different implementation of animation, which are avoiding all these rules. For example Dance of War, it has animation which cannot be cut down by using immediate skill and it's duration is unaffeceted by any attack duration buffs/debuffs.

Not sure what you mean, I just logged my old Warden and had no problem to cut Dance of War's duration with either Steadfast or Shield-Slam (the only 'Immediate' skills I could find in a one minute search, I basically know jack about this class).

[Stever1388]

What you say is absolutely not true, when new LIs came out the speed system was exactly as it is now so -3% AD does not translate to +3% DPS at all. The time when that was true is a few years back, when most skills had a noticeable PSD.
As for the amount of AD after which it becomes useless, it is mentioned and then confirmed in this very thread to be -40%.

I didn't say "new LIs", I said "when LIs first came out", which was back when Moria came out. There use to be a post on the forums that went over AD in as much depth as this one (if not even moreso somehow), and in that post it basically argued that -AD was the best stat for champs because of the additive damage increase it gave (of course I think back then you could only really get the 2 relics that gave -3% AD so it wasn't like you could go bonkers with it). I think the post got lost in the forum revamps, since it was posted back in the Moria days (some ~14 years ago? yikes!).

I did see that the post mentions it becomes useless at ~40%, but is it worth even trying to get to 40%? Is 5% the "ideal" amount before switching to using other stats (like crit, mastery, etc)? This is probably not an easy question to answer, because there are probably a lot of variables, but for example, some of the new traceries have -AD as a set bonus on them, but to get that you have to give up +crit or some other bonuses from other sets. Is it worth getting the -AD set vs getting the +crit set? What if you have -AD on other items already? And of course champs get a bunch of -AD from skills/traits so maybe it's not worth getting any more from traceries or other sources.

[Gabli]

I didn't say "new LIs", I said "when LIs first came out", which was back when Moria came out. There use to be a post on the forums that went over AD in as much depth as this one (if not even moreso somehow), and in that post it basically argued that -AD was the best stat for champs because of the additive damage increase it gave (of course I think back then you could only really get the 2 relics that gave -3% AD so it wasn't like you could go bonkers with it). I think the post got lost in the forum revamps, since it was posted back in the Moria days (some ~14 years ago? yikes!).

I did see that the post mentions it becomes useless at ~40%, but is it worth even trying to get to 40%? Is 5% the "ideal" amount before switching to using other stats (like crit, mastery, etc)? This is probably not an easy question to answer, because there are probably a lot of variables, but for example, some of the new traceries have -AD as a set bonus on them, but to get that you have to give up +crit or some other bonuses from other sets. Is it worth getting the -AD set vs getting the +crit set? What if you have -AD on other items already? And of course champs get a bunch of -AD from skills/traits so maybe it's not worth getting any more from traceries or other sources.

The old runes were -2.5% IIRC, hence I thought you spoke of Heroic Alacrity when you wrote -3%.
As for the value of AD in the current system, if probably varies a lot from one class to another as some skills might get full value, but most don't (only about 21-22% of what would be expected).

[Krindel]

Not sure what you mean, I just logged my old Warden and had no problem to cut Dance of War's duration with either Steadfast or Shield-Slam (the only 'Immediate' skills I could find in a one minute search, I basically know jack about this class).

That is strange, I can't cut Dance of War animation with Steadfast and Shield-Slam, try to use next skill after immediate skill, you will see that DoW animation is still there. I have female elf warden.

[Gabli]

From the very first post in this thread :

According to moebius92 a third skill can’t start activating until the last of both skills’ GCD have ended: this is not something I tried to confirm but there is no reason to assume it works any other way now.

This is exactly what is happening. When you 'cut' Dance of War with Shield-Slam, DoW immediately goes on cooldown and S-S starts activating. However any third queued skill won't start before both previous skills are done. Since DoW has such long durations (animation and skill), it always finishes after S-S.

I've done a few tests to confirm this, though I did not do the measurements more than twice so the accuracy won't be better than several milliseconds hence I only wrote timestamps up to 0.01s.
I used [Thrust - DoW - Thrust] and [Thrust - DoW/S-S - Thrust] sequences after having gathered values for the relevant skills durations (including the two stances counterparts for the builder skill) :

For every entry, it reads "ACT | GCD", ACT being the "animation duration" and GCD being the larger of the two durations (animation versus skill) since according to Vastin, and what follows is consistent with this unless immediate skills are used, the game always uses the longest of both before you can fire another skill.



Note : the 10% values were tested both with Assailment + Recklessness and with In the Fray + Determination. Orange entries are those for which "animation duration" >= "skill duration", green ones those for which "skill duration" > "animation duration"


Then I checked the times at which the skills went on cooldown when executing both chains. When using the immediate to cut DoW, I tried both to use it about as fast as possible after DoW started activating and waiting a bit before doing so.
The idea is that using an immediate skill "kills" the previous' skill's animation but not its "skill duration". Hence, the skill flow depending on whether or not the immediate is used is as follow :





From there we understand that the time that separates both times Thrust/Toss goes on cooldown is the sum of DoW's GCD and one Thrust/Toss (ACT+(GCD-ACT)), hence its GCD. The tricky part is that using an immediate can modify the cut skill's GCD because it kills its "animation duration" but not its "skill duration" :

• Thrust/Toss is not cut, so the duration is indeed the previously measured "GCD" : the max of animation and skill duration.
• If DoW is cut by an immediate skill, we must use its skill duration, which is fully reduced by Attack Duration. If it is not cut, we must use the "GCD".

As such, the timelaps between both times the builder skill goes on cooldown is as follow because without AD buff, DoW's GCD is equal to its skill duration:

• If Shield-slam is used : GCD(Builder, buffed AD%) + GCD(DoW, 0% AD) * (100-AD%)/100
• If Shield-slam isn't used : GCD(Builder, buffed AD%) + GCD(DoW, buffed AD%)

We can then expect the following times :

• With Shield-Slam to cut DoW :
  
  • 0% AD (Assailment-Deter.): 4.15s (1.35 + 2.80)
  • 10% AD (Assailment-Reck.): 3.84s (1.32 + 2.80*0.90)
  • 10% AD (In the Fray-Deter.): 3.41s (0.89 + 2.80*0.90)
  • 20% AD (In the Fray-Reck.) : 3.11s (0.87 + 2.80*0.80)
  
  
• Without Shield-Slam to cut DoW :
  
  • 0% AD (Assailment-Deter.): 4.15s (1.35 + 2.80)
  • 10% AD (Assailment-Reck.): 3.92s (1.32 + 2.60)
  • 10% AD (In the Fray-Deter.): 3.49s (0.89 + 2.60)
  • 20% AD (In the Fray-Reck.) : 3.41s (0.87 + 2.54)

Gathered data : times at which skills go on cooldown (initial Builder skill is origin of time, meaning it goes on cooldown at 0.0s)

• 0% AD (Assail-Deter.):
  
  • No Immediate
    
    • DoW : 2.67
    • Builder : 4.15
    
    
  • With Immediate
    
    • DoW : 0.31 or 0.75 (depending on how long I waited to cut)
    • S-S : 1.22 or 1.66
    • Builder : 4.15

• 10% AD (Assail-Reck.):
  
  • No Immediate
    
    • DoW : 2.60
    • Builder : 3.92
    
    
  • With Immediate
    
    • DoW : 0.48 or 0.72
    • S-S : 1.37 or 1.61
    • Builder : 3.84

• 10% AD (ItF-Deter.):
  
  • No Immediate
    
    • DoW : 2.60
    • Builder : 3.49
    
    
  • With Immediate
    
    • DoW : 0.42 or 0.70
    • S-S : 1.31 or 1.59
    • Builder : 3.41

• 20% AD (ItF-Reck.):
  
  • No Immediate
    
    • DoW : 2.54
    • Builder : 3.41
    
    
  • With Immediate
    
    • DoW : 0.36 or 0.65
    • S-S : 1.23 or 1.53
    • Builder : 3.11

As you can see, it is a decent fit for expected rotation durations for every tested case.


Edit : the test was done with a male Elf warden using a sword.

[Krindel]

Thanks for clarification, I would never figured it out by myself, I appreciate your work. Good to know that using immediate skill after DoW can save some time and also that attack duration buffs works too. Lot of defensive gambits have long skill duration animations and determination trait line does not have +10% attack duration buff, so gameplay in blue feels much more clunky than in red or yellow, also using stun break skill as animation cutter in tank spec can be dangerous. Sorry for off topic, discussion about clunky blue warden belongs to warden forum section.

[Gabli]

You're welcome.
It's not that big of a derailing, and a good occasion to sump up a few points related to gameplay concerning Fast and Immediate skills :

• Fast skills are most useful when used after skills for which "skill duration" >> "animation duration" because they start immediately after the previous one finished animating. Other ways to word this is that fast skills always force the previous skill to use "animation duration" rather than "skill duration" to decide when the next skill can start activating, or that fast skills negate the previous skill's PSD if it has one.
• Immediate skills are most useful when used after skills for which "skill duration" << "animation duration" because they "kill" the animation of the previous skill but don't bypass its "skill duration". Hence an immediate does not speed up rotation if used after a skill that has a PSD*.
• Regardless of rotation duration, immediate skills allow the cut skill to go on cooldown sooner than it would have otherwise and as such allow to use those more frequently.

The last point is particularly important for champions for example, where Ferocious/Brutal have close to 1.5s "animation duration" and so using immediate skills to cut them can easily shave about 1s off of their effective cooldown.
For Brutal Strikes, if you don't cut it you can use it about every 5.7s on average. If you could cut it every time, it would likely be close to every 4.5s.
In the most extreme case (Brutal has a negligible "skill duration" compared to "animation duration" and you manage to perfectly time the immediate), the DPS of a [Brutal - Fillers - Brutal] sequence where you cut either of the Brutal Strikes would shoot up by about 25%.





*This might not hold when using the "double-tap" technique. I don't use this and have not tested it so far either so no input from me on that front !