Following discussion in both the Hunter class forums and the Rogue class forums, there appears to be some form of a trend in crit reduction on physical attacks.

The perceived reduction appears to apply to physical attacks, and shows itself in the form of the actual amount of crits being upto 5% below what our paperdolls imply. So far there has been testing with larger samples in both the Hunter and Rogue class forums, but my suspicion is that this also applies to other physical specs.

Existing topics in the class forums:
Rogue Class forums
Hunter Class forums

External information on the topic:
WoW Forums -> # [Bug?/Theorycrafting] Feral crit chance


As a search on these forums did not provide any other direct connections to other classes, I will edit these in if they exist.


Now the purpose of this thread is to get a broader view of this phenomena. Other classes might be affected by this as well, but this somehow showed up in testing done. My questions for this thread would be:

Does this affect other physical dps classes and specs?
Where does this deviation come from?

Tests against level 80 dummies show the right critrate... so far.

Correct, this issue was tested in one of the various threads touching on this and found it's not a tooltip error, but a result of attacking boss-level targets. You won't see a crit reduction against a level 80 dummy but you will against the skull.

There are a few possibilities I can toss out. The two most likely ones are a flat level-based reduction on the order of 4.5%, and that when attacking a level 83 boss you're actually using the level 83, rather than 80, rating:crit conversion ratio. Other options include a multiplicative level-based thingy (ie your crit rate gets cut by 1/5), using level 83 base crit, and various permutations thereof. If someone could get (very very large) samples of crit rating and zero crit rating that could help debunk the theory of level 83 rating conversions. Seeing the reduction at zero crit rating means it's not crit rating.

Vulajin's testing (here) seems to torpedo any multiplicative theory (as a multiplicative factor can't totally remove your chance to crit), and strikes some pretty decisive blows to the crit-rating conversion theory. I think, in absence of evidence to the contrary, we have to assume that it's a flat crit reduction of 4.8%.

There's still questions of what classes of attacks this applies to; my preliminary testing indicates that rogue yellow attacks are subject to the same 4.8% reduction, while poisons (and, presumably, all spells) suffer a lesser reduction of around 3%.

The speculative theory in the rogue threads at the moment is that there's a global 3% crit reduction on boss level mobs, affecting all incoming attacks of all sorts; however, physical attacks are subject to an additional .6% per level penalty based on boss defense. However, as this is largely speculative, additional testing would be required to confirm it.

I tried to explore PSGarak's crit rating conversion theory, since Whitetooth posted rating conversions for all levels up to 100.

What I came up with was 46 crit rating per 1% crit at level 80, and 58 crit rating per 1% crit at level 83.

The missing parts of this analysis are two-fold:

1. It is entirely possible that the 80-90 range of rating conversions uses a different formula, as they do with 1-60, 60-70 and 70-80.

2. Even if #1 were false, I don't have the math skills to extrapolate if a 12 rating/crit difference is enough to explain the 3-4% crit reduction.

Maybe someone with more experience at this sort of thing can try tackling the next phase of the problem?

I do not see definitive proof that passive haste is reducing miss. I see one person testing with very small sample sizes who has theorized this could be the case. Please PM me further large-scale tests if I haven't spotted them - this interests me.

Haste reducing crit also would not correlate with Vulajin's tests. 4.83% crit chance modified by haste would not produce 1 crit in 11156 attacks. Reference Picture 1

He used [Omen of Ruin] - note 1.5 attack speed. His paper doll had 1.49 attack speed. That's no more than 1% haste. 1% haste would not reduce crit by 4.8% even if it affected the calculations.

60/1.5 = 40 attack per minute (times two, as dual wield = 80).
60/1.49 = 40.27 attack per minute (doubled to 80.52).
This tiny change would not cause you to go from 4.83% chance to crit to approximately 0.03%.

We can rule out haste/PPM affecting crit. Whether it's based strictly on level, defense, or other factors is still debatable.

If you're refer to my test, linked by Aldriana above, you'll see that my chance to crit was nearly eliminated if I was only barely above 4.80%. For any difference in crit rating/agility conversion to achieve this effect, my total crit from agility and crit rating would have to have been reduced to 0.295% (since rogues have -0.295% chance to crit before agility and crit rating are applied).

In other words, the crit rating/agility conversion theory is pretty much completely implausible based on the data.