This mod is the same performance as FreeMe. However, this mod has a lot of junk removed, so that is nice.

No, crit doesn't have diminishing returns. There *is* a softcap, but even wearing the BIS set with some agi pieces, you aren't in danger of hitting it. The softcap is essentially 100% minus your chance to miss or get a glancing blow -- so the more expertise or hit you have before their caps, for example, the higher your crit cap is. Realistically for a ret paladin, the crit cap is I believe about 75 or 76% (glancings are 24%?).

Also, for the most part retribution paladins don't trust simcraft. Rawr seems to be the standard when it comes to modelling (though Exemplar and Redcape's spreadsheets are also quite popular and accurate, just not as easy to use), and in my experience the last few builds are actually quite accurate. I haven't look lately, but last I checked crit was still better than haste (by a small margin) in Rawr.

There is a perception that the current modelling for haste is not entirely accurate and reliable while the modelling for crit is completely solid. This perception is quite accurate, since all of the good sims take haste into account when finding 2t10 proc rate as well as reduction of the GCD for Consecrate and Exorcism but these calculations are necessarily approximations and don't reflect the way a real combat unfolds. It is nearly certain that in some fights haste comes out noticeably above crit and in some fights it is the reverse but actually nailing down a 'firm' value for haste is effectively impossible.

Depending on the gear, sim chosen and settings on that sim haste comes out slightly above or slightly below crit. We know that haste and crit are close and we know the value of crit quite precisely. Given those things most top players gem for crit over haste. They are hedging their bets by doing so and I think that is most likely the right thing to do.

Not that any sane Ret will hit it, but the white attack crit cap for Ret is 24 glancing + 4.8 crit depression, for a 71.2% cap. Specials continue to crit past the softcap.

Simcraft likes haste too much, but Rawr shows a Crit 1.92 versus Haste of 1.9 for my gear. Due to movement and TAJ mechanics, crit is always better.

Crit depression works *against* the crit cap. As such white crit cap (when expertise and hit capped) would be 100 - 24 + 4,8 = 80,8.

Haste has the 'nasty' tendency to not translate to a smooth dps curve. More haste doesn't necessarily mean more dps. Yes, you'll get faster swings, and as a result more white damage, but more haste may also mean that the FCFS system results in a less efficient cycle of abilities being used.

It's very well possible you have tested the simcraft case when you were indeed just at an edge of a bad efficiency cycle, and even a small amount of more haste getting you into a more efficient ability cycle. This can cause haste to have a greatly increased relative value over crit. But once you add in the extra haste the relative values between the two could be more reasonable.

This doesn' t mean simcraft is wrong, or right. It just means that haste is a difficult stat to valuate. This is why 'good rets' gem str+crit over str+haste, because it's a more reliable increase in DPS. Whereas haste is a 'could be good, could be bad' type stat. I'm not saying haste is bad, having haste is definately a good thing. It's just you can't always blindly trust the numbers of a sim or model when it comes to haste.

While this is true from a modeling standpoint, when it comes to actual play this isn't really the case. A human's reaction time + latency ensures that we won't be able to react as perfectly as a simulator, and these peaks and valleys tend to even out.

The other wierd thing about haste is how it interacts with percentage effects. The value of haste rating increases with effects like windfury and heroism/bloodlust. This is why it's always recommended to use your haste pot during a blood lust (although it is more important to pair it with AW, but you should be dropping all of your cooldowns at once in a bloodlust).

Actually, you're right. The accepted theory as of late '09 on white attack crit suppression was wrong. I hadn't checked the crit supression thread in a few months, but it appears Aldriana, Landsoul, and company have changed their tune on "4.8% forced crit to hit conversion."

I performed a test of 287 attacks against a level 83 target dummy. I had 83.8% crit from buffs and debuffs and was expertise/hit capped. After 287 melee attacks I had 68 glancing blows (23.7%), 219 crits (76.3%), and 0 hits (0.0%). If the '09 theory was correct, then I should have observed on average 4.8% hits. Basically it boils down to crit capping being even less of a concern for Retribution given that the real cap is 80.8% rather than 71.2%.

Maybe I'm being dense here, but how does that prove that the real crit cap is 80.8% instead of just 76%? It seems to me (especially based on your results) that the "crit suppression" was simply removed.

From the data set given we can't tell if crit suppression simply isn't in effect or if his crit is suppressed from 83.8 to 79.0. If it is the second then having a 24%/76% split would just indicate that glancing is > crit on the combat table. To know for sure we would want a sample with a toon with 76% crit including all buffs/debuffs and a couple hundred swings. That would tell us exactly what is going on.

I hope you don't mind, but I updated your awesome little addon to include an in-game slash command for easy toggling of the chaos bane canceling. This should make testing a bit more convenient.

You can check out the updated version here: cbb_with_slashcommand.zip - ChaosBaneBuddy

Usage is: /cbb [toggle|on|off]

edit: updated link