Let’s Not Be So Hasty
There seems to be a lot of hub and a lot of bub about Haste lately and how it pertains to our stat weights. Exemplar on the EJ Ret forums has apparently done some work and came to a few bold conclusions:
Having corrected autoattacks, Censure, and tweaked SoC this is what I am presently getting:
Mastery ~= Haste > Crit
I’ve tried various setups both in my present gear (4 T11), along with T12 and Heroic T12, with the same results.
Any severe imbalance in stats causes the ratio to shift. You want about 20% more Mastery than Haste (1.2:1 Mastery/Haste). Too much more than that and Haste becomes superior. The rate changes so slowly, though, that you could be as far as 1.4:1 before you would even get a few DPS variance. Thus it should be pretty safe to just reforge for Mastery without worry that you’re going too far.
Crit is a reasonably close third. In most reasonable gear/reforge setups their values are about 1.25 (M/H) to 1.1 (C).
In general if you swap Haste and Crit from our old reforging desires, you get what you want. Hit/Exp cap, then reforge Mastery. If an item already has Mastery, change some Crit to Haste.
To use myself as an example, my current Mastery:Haste ratio is 2.506:1 (1784:712). If I were to take his advice and change my Crit reforges to Haste, let’s say I would gain about 350 Haste rating without sacrificing Hit, Expertise, or Mastery (my own quick napkin calcs are around this number, if a bit lower), bumping the ratio up to 1.680:1. I can see as we get more T12 gear, this ratio will drop a bit to Exemplar’s predicted range.
Okay, we can get there. But how did Haste end up above Crit? Exemplar has more explanations:
- Correcting Censure increased the value of Haste, but being informed (and verifying) that it crits for 200% (not 150% as other spells) helped temporarily restore Crit’s balance.
- Correcting Autoattack then greatly increased the value of Haste. Haste obviously produces more Autoattacks (and seal/SoC Procs!), but this also impacts generation of Exo procs. This can ripple in a minor fashion to Mastery as Exo could proc DivPurp and produce a new TV.
- More Haste should increase everything equally (more attacks = more crits, more CS/TV = more HoL damage), but due to talents it appears to increase Mastery more than Crit. More Haste means shorter CS. Shorter CS means more HoL damage and more HP, therefore more TV for yet more HoL damage.
I’m not too sure what he’s talking about with the whole “correcting Censure” business. I know that the crit damage of Censure got bumped up to 200%, but perhaps Censure just wasn’t modelling correctly in his spreadsheet. Personally I don’t use his sheet, though now that Redcape is retired I may have to.
More SoT/SoC damage is a given, but I think Exemplar is overstating the production of AoW procs and subsequent Divine Purpose procs. Yes, I know he said “minor”, but let me show you how minor it really is.
This is a simple table translating Haste rating into weapon speed, using the following equation:
Weapon Speed = Base Weapon Speed / (Haste1 x Haste2 x Haste3… x HasteN)
In this equation, you have to calculate each Haste effect separately (Haste1, Haste2, etc.). For the purposes of this very general discussion, I am not including our raid buffs (5% from Mind Quickening/Wrath of Air Totem, 9% from Judgements of the Pure)
Why do we need weapon speed, you may ask? Because it is used to calculate your number of autoattacks:
So, over the course of a five minute fight (300 sec.), with a 3.6 weapon speed and no Haste, you would get 83.3 (repeating, of course) autoattacks. For those that may have missed it, that calculation is just a simple,
300 / 3.6 = 83.33
The other two columns aren’t so clear, so let me go over those.
Art of War, as the talent states, is a 20% chance on autoattacks to generate a free Exo. Therefore, it makes the most sense to model this as “procs on 20% of autoattacks.” Obviously, the longer the fight duration, the closer to 20% you’ll actually come, but for right now this is the best way to model it.
In case you were wondering, I also did 10-minute and 15-minute fight duration tables:
As for the “DP from AoW” column, this is where things get a bit wonky. Divine Purpose states that each attack listed (of which Exo is one of them) has a 15% chance on cast to proc DP, which is essentially a free TV. Now, if you’ve had enough time to Ret it up in raids recently, then you know that DP is a weird proc and it often likes to proc itself, so these numbers admittedly might be a little low. But let’s play it safe and not assume it’s going to proc seven additional DPs in a row, shall we?
Anyway, this all means that any additional AoW procs you get from more Haste each have a 15% chance to proc DP.
Taking all of this data, I can plot both the number of AoW procs and the number of DP-AoW procs against Haste Rating and see how it scales:
Anyway, I’m sure this looks like just a bunch of numbers to a lot of people, so let me summarize these graphs for you:
Click to embiggen.
What this table is saying is that, in order to go from X procs to X+1 procs, you need the specified amount of Haste Rating to do it. Most fights these days are between 5 and 10 minutes, so to get just one more Exo you need between 769 and 385 additional Haste Rating.
Clearly, fight length is a limiting factor. The longer the fight, the bigger role Haste is going to play. Of particular hilarity is the amount of Haste Rating need to get just one more DP proc from an AoW proc; that’s what happens when you have a low percent chance to proc something on a spell with another low percent chance.
I think Exemplar wrapped things up nicely here (emphasis added):
After slight cleanup I am still finding Hit > Expertise >>> Mastery > Haste >> Crit (quantity of greater than sign used to show relative superiority). Reforge to Hit/Exp cap. Then reforge Mastery. If an unforged item has Crit/Mastery, then reforge Crit->Haste. I understand and appreciate this is controversial and still under discussion. At this time the OP has not been altered from the previously believed Mastery > Crit > Haste.
I do not see a way to (dis)prove this definitively via testing or logs. RNG for DivPurp, AoW, and Crits, as well as Fight Length, optimum usage of CDs, just one more or less raid buff produces too much variance. Theory is only as good as our tools and we’ll keep working on them.
Of course, as I stated before, Seal damage is only going to go up in a linear fashion with more Haste, I’m not disputing that. What I am disputing, though, is the impact of Haste on Art of War and Divine Purpose. It definitely has an impact, but I feel that it’s not the ‘new stat to stack’ that everyone seems to believe it is.
Stacking Haste is essentially buffing the Random Number Generator. Sure, it may provide good results 15% of the time, but we can’t and shouldn’t base our gear choices around random chance.