Well, Purifying Brew has a diminishing effect the more you use it. Shuffle required low uptimes because we simply didn't have enough chi to use it on top of PB.

With the latest energy change, optimal Shuffle uptime is going to be pushed up to around 90-95% because PB can be used so often.

I tanked Tsulong and Garalon last night on the Beta.

Here is the log for our "kill" of Garalon: World of Logs - Real Time Raid Analysis

I say "kill" because we called for a wipe at 10% so we could practice another attempt, but then Blizzard ended the testing and killed us.

My shuffle uptime was just under 80% and I wasn't using guard enough (I hadn't played Brewmaster in a couple weeks so I was a bit rusty.) Damage between myself and the other tank (Arij, warrior) was comparable (I took ~14% more than him), and may have dropped lower than his had I had better Uptimes on Shuffle and Guard.


Here is the log for our Tsulong kill: World of Logs - Real Time Raid Analysis

Uptimes for Shuffle are low as there is a quite a bit of downtime what with tank swaps and nothing to strike for a while. I took ~23% more damage than the other tank in this encounter. This boss hits like a truck; my stagger was getting to moderate and heavy in just a few hits and I felt very squishy. Seems on these heavy physical encounters that our lower armor will really start to take its toll with our relatively low avoidance and mastery at this stage.

You should consider, while looking at world of logs, that damage taken is calculated without absorbs. When you stagger damage you absorb hits and then you take that absorbed damage as a dot, that means that if you wouldn't use purifying brew your stagger damage would double dip into damage taken.
If anyone can post on US beta forums can they mention to add combat log event for staggered damage, so world of logs could for example count absorbed damage as healing done by the brewmaster (similar way that blood shield accounts for death knight's healing), or maybe this can be accomplished by worldoflogs itself by tricky calculations of refreshing stagger dot


/edit:

Also Venyasure i don't understand why do you use Summon Black Ox Statue in your calculations since the tooltip specifies that

I do get 1 stack sometimes with a 2H ; it's working as described.

I think it depends on what we call "optimal". If your aim is to never take a full melee swing from the boss (that's the kind of thing you might want to do on Blackhorn for example), obviously you need a 100% Shuffle for the +20% stagger bonus.

But as far as overall damage taken is concerned, i've made a few calculations after Astrylian and Disargeria told me a 100% uptime wasn't optimal. Well, what i did was rewrite my program with a varying Shuffle uptime (and debugging it ; i've checked many times the validity of my calculations, and I'm beginning to be pretty confident in the absence of errors). And I found they were quite right. I even get uptimes inferiour to Disargeria's 90%.

The uptime depends on gear of course. It also depends slightly on boss dps ; that's because, even if Purifying Brew, Shuffle, Stagger and avoidance reduce a percentage of the boss damage (which means damage reduction should be proportional to boss damage, making uptime independent of damage taken), there's still a part in damage reduction that is not proportional to the boss damage : the healing from Expel Harm, Guard and the statue. Note that a higher uptime increases dps, it therefore increases the rate at which the shields of the statue are activated. I observe that the harder a boss hits, the lower uptime should be (but it doesn't fall very low, even in extreme conditions ; it has been nearly impossible in my calculations to get an optimal uptime lower than 55%).

The results are quite different from what i had earlier with a 100% uptime. Haste has gone down a lot ; probably because haste increases the number of Purifying Brew, and having a lower Shuffe uptime gives us a "base" number of Purifying Brew that is already high, making it difficult for haste to increase it significantly.

Here are the results for :

- a boss hitting once every two seconds
- a boss doing 100000k dps (meaning it hits for 200k every two seconds)
- 16k armour
- 1k dodge
- 1k parry
- a weapon with 3285 dps (ilvl 463, ie MoP heroic dungeon)
- weapon speed 3.2 (as i said earlier, it has no impact)
- 2k haste
- 1k5 expertise
- 1k5 hit
- 8k agi
- 2k crit
- 2k mastery
- 18k stamina




The curve shows the damage reduction depending on Shuffle uptime. Note that the 78% uptime depends on gear, and a bit on boss damage. The optimal uptime tends to increase when I increase parry, dodge, haste, hit, expertise and haste, but it decreases when I increase mastery. Which is quite logical. The general idea is that the better your gear is, the higher your uptime should be.

I did take that into account, don't worry ^^

Edit :
The statue is not a direct increase of the Monk survival. But it increases the survival of the raid group. Therefore i considered that it had to be taken into account. A bit like if a DK's Glyph of Rune Tap scaled on AP. It's a minor increase of survival, but i prefer to be thorough.

I think that's how we're going to be approaching the issue in practice, but it's nice to know where we stand from a completely theoretical point of view.

Thanks for the analysis so far, Venyasure. I'm curious about the seemingly very low crit priority. Is Elusive Brew simply not worth it enough to reforge away from Haste and into Crit?

Source: MMO-Champion - Faction Capital Changes, Lorewalker Story, World Boss Preview, Blue Posts, MMO Report

Yes, as i told you earlier, I think that this is indeed the idea. You can (and should, I think) also priorize Keg Smash over Blackout Kick when KS's cooldown finishes. In other words, fit something at the end of some Shuffles, from time to time ; a Purifying Brew would certainly be a good fit if you have a high Stagger. And obviously, you can keep in mind the moments when you're going to need parry+stagger, and the moments when it's ok to forget about the buff for a couple of seconds.
Elusive Brew is quite huge. But apparently, not huge enough for crit to compete with the many advantages of haste (remember that haste influences Elusive Brew too ; but not as much as crit though).


Also, I've read a couple of websites quoting the results I've posted here. For instance, they wrote that 78% was the optimal uptime for Shuffle, and one of them copied the stat priorities my program had given. But they forgot that these numbers depend on gear and boss damage. They do vary if we modify these parameters. The results I've given were for a rather low item level. By increasing some stats, especially expertise, hit, agility, and haste, we get these results :




Obviously, hit and expertise are less valuable in this case, since they have already been increased significantly. Note that the value for hit should be 0, since the Monk is hit capped in this calculation. After these modifications, the 78% uptime is not valid any more.

I am currently working on an Excel version of the program, so that other people than me can use it.

Edit : for the sake of curiosity, here is what you get for the same Monk, but for a boss hitting twice harder (200k dps ; 400k hits every two seconds, that guy is tough !) :

Oh, I'm at fault for spreading the 78% uptime. I didn't know the extent of how much gear influenced it.

I do wonder about the optimal use of Elusive Brew. It doesn't have a cooldown, so instead its frequency of use is tied to its charges. That means that we can use it as we get it.

So, instead of saving these charges for a long cooldown like Evasion, wouldn't it be more beneficial to use these charges right before individual melee attacks, sort of like Druid's Savage Defense?

Regarding EB, my guess so far was to use it at 10 charges or before the charges vanish (if it happens) but that was very empiric, not based on any calculation. It seems to stack quickly enough to justify a use on CD if there's no reason (due to the encounter) to save it. Again, it's very empiric ! (i prefer ton stress that :p)

Also, it's worth nothing that it might be very powerful on fight where tanks swap often. (Imagine that on Blackhorn HM : free 15 sec CD each time you taunt the boss :o )

For overall damage reduction.
Haste has many benefits : it increases Chi generation, therefore increasing the possibilities to use Shuffle or Purifying Brew, it increases the number of white hits proccing Elusive Brew, and it increases dps and therefore the efficiency of the statue (but this is a really minor gain). Hit and expertise are great because they increase Chi generation too : a missed Jab costs 8 energy, and a missed Keg Smash is horrible.

About that, i'd like to comment on why the uptime moved from 78% to 88ish%. I have significantly increased Chi generation by raising haste, hit and expertise. This gives us much more uses of either Purifying Brew or Shuffle. The Shuffle efficiency seems to have increased more that the PB efficiency resulting in a higher uptime, the probable explanation being :
Which means that an abuse of PB is not what is more profitable after the gain in Chi generation.

Mastery is a poor damage reduction stat because one point of mastery does not increase Stagger very much. This has been explained in the blue posts in a paragraph comparing Brewmaster mastery and Druid mastery : mastery does not only decreases damage taken over a fight, it has a double utility. It also diminishes the power of each individual hits, delaying the damage. This virtually increases the damage you can take before you die, like stamina. As it has an effect on both overall damage reduction and risks of sudden deaths, it had to be pretty low for damage reduction, otherwise it would be too good.

That's something I didn't really think about. I simply calculated the average number of stacks gained per second to calculate the average uptime of the buff.
As far as overall damage reduction is concerned, I guess the optimal use is one which includes as many boss melee swings as possible while the dodge buff is active. For a boss hitting more slowly than once every second, it would indeed be interesting to consider using one stack just before a hit, so that each stack is useful on one attack. Unfortunately, it's not possible to choose how many stacks you use.

Did anyone test if Keg Smash is only affected by hit chance? I think that it cannot be parried or dodged, but I will definitely check this after I get back home today.

That would value hit much higher than expertise in your calculations.