His Aimed Shot crit rate sub-80% is fairly normal. Spamming AiS @ 100% crit for the first 20% of health skews the final number.

Probably because they found the exact haste amount needed to comfortably get in 2 AiS between each CS. After that's feasible haste doesn't really have anymore value.

Post hoc, ergo propter hoc: He most likely reforges haste to mastery because thats what the consensus is in regards to stat value - crit > mastery > haste for the 3 secondary stats, this is true for all 3 specs, so most hunters you can find (that read up and optimize) will reforge haste to crit (primarily) or mastery (secondarily) excepting only the plateus of haste that are being talked about - 1.67 second CoS casttime for SV, and whatever the math on the previous 2 pages ended up with for MM (since I don't play MM myself currently, I haven't kept up as well on MM as I have on SV).

If we're sticking to the basic rotation of SSx2 then CS or AiS depending on timer on SrS, then we actually don't need any haste from gear (or Pathing) to achieve that under ideal conditions. My opinion is that the amount of haste is strictly dependent on individual connection quality in order to compensate for latency.

For example, if your connection to the server is 200ms, and for some unfortunate reason you experience upstream choke so the client-server sync is messed up and you lose 100ms off every single gcd. In this case, your actual cast time for a 15 second rotation CS-SSx2-AiS-SSx2-AiS-SSx2 becomes

(2 x 6 + 2.9 x 2) / (1.1 x 1.15 x 1.02 x gear_haste) < 15 - 1 (CS gcd) - 0.1 x 9 (lag).

Solving for gear_haste equation results in a minimum haste value of 4.28% (2/3 Pathing) so that your SrS will not fall off.

I hope this will help people realize what haste really does. I really like Gavinas's approach to incorporation a Gaussian lag to the model. My example is just a specific instance of this problem.

I can't take credit for the approach, SimC incorporated that long before I started working on the hunter model to get it up to date for 4.0.6 a couple weeks ago. I agree it's a useful approach for simulating mechanics though.

Just because a player is in a top guild, or has a top parse, does not mean they are optimal, the best at their class, or even understand fully how their class functions. That being said, its quite possible (and very likely) he didn't fully reforge his gear that was his Surv gear 2 days ago to go MM...perhaps he wanted to test it out first, and/or get an entirely different set of gear.


Looking over his damage breakdown, we can do some quick napkin math...

First off, I'm comparing 128 haste to 128 mastery, or .715 mastery to 1% haste, or 1.3% chance to get a WQ to 1% haste.

To begin with, Wild Quiver procs on every ranged attack except itself. He had a total of 211 hits, 185 crits, and 6 blocks. Of those, 56+38+2 were WQ hits, so he had 402 total hits and 96 WQ hits, so 306 proccable hits. A 1.3% chance to proc WQ would result in .013*306~= 4 additional WQ hits. His average WQ was 737120/96=7678. Multiplying this by 4, 1% more mastery would result in 30712 additional damage.

Haste is a bit more complex, and I'm ignoring the existence of potential "plateaus" but we can begin with a few simple assumptions:

1% haste is +1% Steady, Auto, and MM-AiS dmg.
That's 1083k, 708k, and 665k. 1% of that sum is 24.5k damage.
That makes up 228 of 402 possible WQ procs, or an increase of .01*228/402=.567% increase in WQ dmg (.00567*737k=4.2k more damage). That puts us up to 28.7k damage, or roughly 2k damage short of what equivalent mastery added so far, with hard-casted Aimed Shots and pet damage left to analyze.
His pet did 623k damage, which afaik gets 100% of haste rating, adding another 6k on it's own (and putting 1 haste rating above 1 mastery rating).

Before moving onto hard-casted AiS and someone arguing about haste affecting MM-AiS, I'm going to go ahead and subtract off the extraneous SS DPS you lose. He presumably spent 14 globals on MM-AiS. 1% of this is .14 globals. Best case scenario for mastery, SS is already at global, so .14 globals is .14 of a single steady shot, or .14*1083378/(60+62+1)=1.2k damage. Haste is still ahead.

Looking now at his hard-casted AiS, he had 47 of them, meaning 1% haste would give him an amount of cast-time equal to roughly half the cast time of an AiS. Note this isn't an increase in actual hard-casted AiS, as focus is the restraining factor there, not time, but it would result in an extra ~3/4ths of a Steady (I'm ignoring haste effects slightly, which may drop it a bit, but meh). This is 6.6k for 75% of a steady, 4.4k for 50% of a steady.

This puts 1 haste at ~38-39k damage added, and an equal amount of mastery rating at <31k damage added on this particular fight. This puts haste some 25% ahead of mastery (5mastery = 4 haste roughly).

As for the whole haste plateau discussion...Chimera shot is making up 8.3% of our damage total, and time is -not- a restraint on AiS, so I really don't see it being something that matters. I also don't think there is a correct "best" rotation, and eyeballing the numbers to me actually implies that capping focus isn't even bad, and you should try to get your hardcasted AiS into 1 of 2 categories:

A. "Glitched" for 2x focus-free AiS (allow focus to build up higher, at 4stacks do SS>AiS>SS>AiS>SS>SS and accept the slight loss of ISS).
B. Aim to hit ~9x focus, and do back-to-back AiS, then go into Steady Spam + Chim refresh + more Steady. Basically the goal here is to let Auto-Shots fire uninterrupted for as long as possible before dumping focus into AiS again. If you do SS-SS-AiS-SS-SS-AiS, you're resetting your Auto timer much more often, and cheating yourself out of free shots. 1 extra Auto noncrit is worth more than 5 focus as a portion of AiS, so even if you hit 100, as long as you aren't wasting too much focus, it should be a net gain in DPS.

Thought I should share some info so there isn't a lot of misinformation on the community.
I am not 100% certain that my choices are the best so don't copy them by any means. (especially about cloak+gloves enchants and glyph)
But I can say that hardcasting aimed shots is for sure the best (obvious), and that above 80% you shouldn't bother putting serpent sting or casting chimera shot.
Now about a rotation, I think its very hard to actually keep a rotation so strictly as you guys might want to. Much harder than it was at survival at least. Sometimes I find myself casting 3 aimed shots in a row*.
In general i was doing either 2xsteadies or 4xsteadies and it didn't really matter if i casted aimed shot+chimera together or even 2xaimed+chimera (1proc), the ISS buff was still up. So trying to do the perfect rotation isn't as important at MM as long as you don't waste focus and keep ISS buff up and don't let your serpent fall off.
In general though, I indeed was refreshing serpent sting when it was below 3seconds left on it so I didn't prioritize it over aimed shot.
I haven't done any extensive theorycrafting though, it is just what felt to me as optimal after a couple of hours testing on dummy.
Also, unsure of the talent choices and reforging. Let's just say im with you at the journey of figuring out MM after 4.0.6
Judging from my chimaeron dps tho, my choices aren't probably far from the optimal.

*Not sure if everyone is aware of this issue but if you are casting an Aimed Shot while your !Aimed Shot procs, then your cast Aimed Shot doesn't consume any focus as well, so you do 2 aimed shots for 0 focus cost, thus letting you cast a third one right after.

During rapid fire you would probably be doing 4 Aimed Shots in a row too, and its fairly desirable to throw a few more steadies than you need to so you get 4 aimed shots back to back during the RF phase. I found you can peak at 40k DPS for a period.

Regarding the "rotation" I find MM a lot simpler to play than Survival, in that there are less decisions to make. You are either going to use chimera shot or aimed shot. If power auras are set up to throw a signal when you have 4 stacks of MM, you will just go: steady > aimed and this is the final decision you'd make, so about 3 things to keep track of.

About hotfixes:

Since the patch notes say Aimed shout be scaling at half the RAP (0.7) vs live's 1.4 I think it's safe to say this is unintended behaviour and that we will probably only be hardcasting Aimed shot during careful aim, heroism or rapid fire. For the rest of the time, if 0.7 is to be the scaling it looks as though arcane as filler will be desirable.

Or start with 1 AI, 2x SS, then AI. This way you wont waste any focus (neither do you), but you got ISS sooner, so you cast AI faster, not casting 2x AI without ISS haste, just 1. It is a little gain (around .2 sec), but it's worth something.

I just want to state that it was impossible to pull out such rotation (cs-ss-ss-ai-ss-ss-ai-ss-ss) with 9.5% haste and 3/3 pathing on target dummy. Srs always failed while I was shooting my chimera and also this rotation always messes up when u get aimed proc. I guess the priority system and just simple spam will do better than always being on the edge of losing your srs and/or constantly lack focus or being focus capped. Just for info I played with 150 ms alone on target dummy.

There are many ways to skin the DPS cat, and which is best depends on your current gear and raid buffs. Although with 4.0.6 it is generally true that stacking haste (to the proper point) over mastery and even crit can be a DPS gain as it allows you to be able to squeezae in an extra shot per CS cycle, it is not always true. I have already listed in the OP one such case - where you have the T11 4-set bonus but not high enough ilevel to reasonably allow stacking enough haste to squeeze in 5 to 6 SS per CS cycle (depending on whether you have an MMM AI proc that cycle) without sacrificing too much mastery and crit where your shots are doing much lower damage. In that situation, it is better to have enough haste to only do 4 to 5 SSs per cycle. The amount of haste required is only a few percent with some additional haste being useful to account for latency and reaction times.

In addition to what Lilyanna stated, when hasted with a glyphed RF, you can actually do the cs-ss-ss-ai-ss-ss-ai-ss-ss rotation in a 10s CS cycle with no haste on gear if desired, although waiting to cast CS to add in more a AIs and SSs before firing CS to extend SrS is better.