Background
Over at the Enhancement Shaman thread, we have built dps simulators. Upon testing these simulators, we are finding that white hits are under-reported significantly. It seems obvious that Flurry being "messed up" is responsible for this. This throws a huge wrench in any calculations of Shaman or DPS warrior theory crafting.
Tooltip - What Flurry *should* do
|
Increases your attack speed by 30% for your next 3 swings after dealing a critical strike.
|
That seems almost straight-forward. Now, perhaps you could read it as "finish all swings that are in-flight, then hasten the next 3". Or, "hasten all attacks, including ones mid-flight, until you have swung 3 times". Either way, 3 is a pretty critical number.
The problem tested
1) Get two weapons of varying speeds, with as large a delay as possible.
- I don't want to have the synchronized hits causing any funny business
- The larger the delay, the less that ping time will affect the results
2) Go hit a blasted lands mob, turn on combat log, making sure no weapons/armor has procs that would affect the outcome (mongoose, drakefist hammer, abacus, etc)
The problem illustrated
In the attached screenshot, I think the problem is very clearly illustrated. A critical strike is affecting up to 6 hits with the flurry buff. Here's how to read the chart:
blue shading - in a flurry state
pink shading - the last crit before flurry fades. At this point, only 3 more attacks should be affected
yellow shading - the number of attacks before combat log reads "Flurry fades from you."
no shading - period of non-fluried attacks.
grey/white column shading for "average" - This shows how many rows I'm actually averaging.
In short, not only does the combat log read "You gain flurry" and "Flurry fades from you", but the average swing speed in those periods actually agrees with the combat log. That is, this does not seem to be server-client lag. This is why I have the average columns there - so you can verify that when the combat log says flurry IS NOT active, it is indeed about 2.6speed mainhand, and 2.8speed offhand. Also, when flurry IS active, the speeds correctly register at an average about 2.0 mainhand and 2.15 offhand.
Seeing as how two classes are highly flurry dependent for their damage, I do believe it warrants its own thread.
I'll be happy to provide the raw combat log if asked.