Wow combat log not accurate?! And more

So as a very beginner player of this game, I started to analyze how this game is logging information. And there were a few questions and odd things I found that I can not answer through a simple google search. So here I am representing what I found and what I question…

First off, I found out that the combat logging timestamps are not accurate on the millisecond (the Logs/WoWCombatLog.txt). You can cast a DoT spell, and every DoT tick shows another timespan that is not for example once every 1.000 second. It sometimes varies like 100 milliseconds. I was hoping this was not the case, but sadly it is. More on the reason why in a future question.

With knowing the inaccuracy of the combat log, I was questioning if that is the only inaccuracy? So I started letting the player do some basic strikes on auto attack and find out how much time is between attacks. The result should have been once every 3.60 seconds with the weapon I had, and being completely naked. But this is what I got:

Measuring the SWING_DAMAGE of a paladin hitting a dummy automatically for over 30 minutes, I was able to find
21:20:26.574 - 20:50:32.446 = 1794.128 seconds timespan.
493 attacks.
1794.128 / (493-1) = 3.6466 seconds between attacks. So about 0.05 seconds slower…

Is the attacking the game delayed by calculations uncontrollable by the player? And if you unlucky or bad PC or busy server your attacks are slower? This can mean more than a % of dps difference… What am I missing about this? I am still just a beginner so my lack of knowledge of the game brings me to that point that I just can’t possibly know. That is why I ask veterans or devs about this?

Which brings me to the last thing. The whole reason I started this ‘research’. When using the ‘spells’ in the action bar, on any class, there is like a 0.5 second moment (guess on feeling) before the cooldown is ready again, when you can hit the button and the spell still casts. I wanted to find out if hitting a button slightly faster is casting a spell slightly faster or not. Sadly, with the combat log not being accurate on the millisecond I feel like no way to prove that. (…says that while realizing how to maybe test it for the future…). But I am asking this anyway. Because someone in the min-max world, or a dev that balances stuff might know about the nuances of this mechanic :sweat_smile:. And if it matter to hit the action bar spell on 0.4 second vs 0.1 second cooldown before ready would make a difference in dps?

2 Likes

I can’t help with your log attack timing inaccuracies, but I can tell you that at the very end you’re talking about global cooldown (gcd). I’m not sure what you mean by pressing a button earlier or not, but if you press a button during a gcd it usually just ques that spell up and uses it once your character can use it. The only way to “lose” damage here is if you haven’t pressed the button when gcd comes back. Most people just spam abilities a little bit, me included. Slightly bad form, but sometimes spells don’t go through on the first press.

Hopefully that answers at least something. You’ve delved into some pretty complex stuff as a new player.

2 Likes

I love this post! :heart:

Ten years ago, as a new player, I made a few posts like it, asking similar questions.

Despite my past interest in the subject, I can’t answer yours, though.

The devs change the detailed mechanics from expansion to expansion and from patch to patch, and I long ago stopped caring and keeping track.

For example, when you talk about hitting the button in advance of the GCD, I looked up what I remembered - SpellQueueWindow https://wow.gamepedia.com/CVar_SpellQueueWindow and found that this one has been removed.

This Console variable has been removed from the game.

  • It was removed in patch ?.

No doubt its function has been replaced by something else. This is what happens. They keep fiddling with these very basic building blocks.

Aso, the devs build in deliberate RNG in the results of attacks. If a player performs 3 minutes of attacks 20 times, even on a target dummy, and even perfectly every time, there will be a 10% or so variation in the results. This variation swamps any small differebces in the low-level mechanics.

I don’t want to discourage your curiosity, though, because an understanding of the fundamentals is a nice thing to have.

The general area you are entering into is called theorycrafting, and mostly it is practiced by fast simulations of characters on computers. The program Simulationcraft https://www.simulationcraft.org/ is where the theorycrafting community comes together. If you read code, most of the answers you are looking for will be in there somewhere, because that is the community’s best effort at describing exactly what happens in a WoW fight.

What peopletend to do is set up their characters’ gear and build, then use Simulationcraft to run 10,000 or so sims of a fight in a couple of minutes. Then they change their stats or build and try again. This tells them what works best. See also the site raidbots.com which implements Simulationcraft and shows its result on a web query.

I should mention that nobody bothers simulatiing anything below max level, so you won’t get anything sensible from simming your level 15 character.

2 Likes

Interesting info about the GCD though. What I meant, is that a few days ago I was able to press a spell that had like 0.5 seconds of cooldown left and it would still cast it. And apparently that doesn’t work anymore for me now… Edit: Found out is happened because I f’ed up the variable “spellqueuewindow”. It is set to be able to Que up a spell 400 milliseconds before you actually cast it. Thanks to the simulationcraft community if was found out. I had set that value to the string “setting” lol :man_facepalming:

Well thank you for both your responses. Lets see if I can learn and find more. Will look more into the simulationcraft too :slight_smile:

1 Like

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.