I respectfully disagree. Your objection is purely academical and hypothetical. As a rule, the higher the threat the sooner you want it on cooldown, so you can use it again in the same fight. You say, by using HT when it's up, it will continually clash with other high threat abilities, but the fact is, it won't. Your example doesn't translate to the reality of tanking choices as they are presented to us. You're much more likely to lose a long cooldown during a fight than a short cooldown. Contrary to your statement, the short cooldowns are much more flexible than the long cooldowns, and will "fit in" on an ability timeline much easier than the long cooldowns.

You don't measure the worth of an ability by its average threat per cooldown, you measure it by its average threat per gcd.

You'll never get a lower tps by inserting a HT in your tanking "cycle" (using the values in Yomon's post, your personal threat values are probably different), and the sooner you get it off, the sooner you're able to get another HT in. Now, there are other reasons for not using HT. Utility, mainly, and perhaps from a threat-per-rage perspective, but clashing with a Shield Slam cooldown is a spurious reason. Fact is, that because of S&B it is meaningless to work with Shield Slam as the basis of a 4 gcd rotation.

Think about it, what ability could HT/SW clash with that would be better to use? Every other combination would lower your average tps. The point is that if we go by Yomon's values, even the optimum S&B cycle of SS>Rev>SS>Dev ad nauseam isn't giving you the highest tps, and that the relative gain of a S&B proc must be weighted against what you postpone by hoping for procs. Revenge and Devastate only gives meaning if we think of them as two gcd moves that must be read in combination with the move following it. That's what I did in the post above. We can all agree, I hope, that using a Devastate just before SS comes off cooldown gives no added tps-benefit. What I did was to calculate the average benefit of S&B in other situations. If you gamble, and go for the S&B everytime, you will lose tps on average. Nowhere is that calculation more clear than when you have abilities that out-tps Shield Slam (and, btw, that's probably not the case for most of us, I just used the numbers provided).

Sometimes, you don't want to maximise tps, of course, sometimes you want to keep up sunders and demo shout and thunder clap more than pushing out the last possible tps, but if you do want to maximise, the numbers indicate that using the "SS > Rev > Dev > ??" isn't always the best bet, even if it is the most manageable. If the numbers change, for example by making Shield Slam much higher in threat this could change.

I can't see how you can possibly reach this conclusion. I delay Shield Slam by 1 GCD, for whatever reason. I've lost 1/4th of a Shield Slam, assuming no further collisions. I delay Heroic Throw by 1 GCD, for whatever reason. I've lost 1/40th of a Heroic Throw, assuming no further collisions. This is neither academic or hypothetical: this is the actual cost of delaying those abilities by 1 GCD should they collide in reality, which they or other abilities certainly will. The correct choice is very clearly to delay longer cooldowns first and even the most rudimentary simulations and tests will show this.

The question is, if we want to discuss it furthermore as it is becoming more and more complex. In the end the discussion is only about a very small TPS gain - the good thing is that if we get the exact value for all specials you can make a priority list within a specific cycle as a "rule" and thee who follows this easy rule will do maximum TPS.
It wound be a easy rule for any tank that seeks advice to do good TPS.

As said, the math becomes more and more complex and maybe it is time, as mentioned above, to go from theory to simulation.

I could program a special move simulator assuming a infinite rage scenario, a encounter with 300 second length with a function to do one, ten or hundred simulations at once and average out the TPS value, showing what is the best rotation.

I could also program, bosses, rage gain, heroic strike, deep wounds, parry hasted attacks and other things; that wound also clarify with what stats you gain how much TPS if you swap from a slow to a fast weapon, and how they affect parry hasted attacks. But I wound only program such simulator if it is really needed and if many people wound use it - it wound be about four to five times more time consuming than the first program.

We do very good TPS, often it is more than needed. So do we want to do, but maximum TPS? Or is very good TPS enough?

I finished my expertise cap testing. The results can be seen at Wow Web Stats .

Summary
Testing Series 0 Starts with 13% Expertise up until 15:13'28.501
-We can see a parry occurs at 15:12'54.000
Testing Series 0 Continues with 13.75% Expertise up until 15:20'45.858
-We can see a parry occurs at 15:20'23.029
Testing Series 0 Finishes out with 14% Expertise up until approx 16:00'38.295 when my elixir of expertise runs out.
-We can see that no parries occur prior to 16:00'38.295

Testing Series 1-4 are all sitting at exactly 410 Expertise Rating which equates to 56 expertise with talents(rounded down)
Series 1 - 1hr 10min
-4309 Total Attacks
-2970 Regular
-289 Crit
-1050 Miss
-0 Parry

Series 2 - 27min 45sec
-1696 Total Attacks
-1186 Regular
-118 Crit
-392 Miss
-0 Parry

Series 3 - 45min
-2574 Total Attacks
-1817 Regular
-175 Crit
-582 Miss
-0 Parry

Series 4 - 3hr 30min
-12864 Total Attacks
-9031 Regular
-915 Crit
-2918 Miss
-0 Parry

Total From Series 1-4
-21443 Total Attacks
-15004 Regular
-1497 Crit
-4942 Miss
-0 Parry

All testing was done dual wielding MH Broken Promise/OH Last Laugh both at max weapon skill. From this testing, unless I screwed something up, we can assume that the parry cap for a level 83 boss mob is between 13.75% and 14% or 56 Expertise.

Don't know how this will change your TPS but hopefully 20k is a large enough sample

I do not know how to test the global cooldown of an ability, but I think that the GCD that Concussion Blow invokes is shorter than normal abilities - dare I say 1sec maybe? Can someone who knows how to test this find out?

If this is the case, concussion blow may be a lot more powerful threat/GCD-wise than we are modelling it as for the moment. Otherwise, just ignore me

Thats exactly what I've been noticing, whenever I use Concussion Blow, the GCD it invokes is definitely not 1.5 second.

Hey there everyone. Xav pointed me to this thread, and I'd thought I'd see if I could lend a hand.

The primary issue I saw with determining Rend's utility was that you're pushing Dev down lower in your rotation, which will in turn reduce the amount of Shield Slams you get due to S&B procs. Actually trying to model this with a mathematical formula is pretty difficult (at least, I couldn't figure out a way to do it that didn't involve lots of iterating), so I built a simple rotation simulator and pulled average ability usage from it. I've made a post at Tankspot with the methodology and the results for a standard tanking rotation. The next step was to find out how putting Rend into the equation changes thing, which is done here.

Apparently you guys were looking at doing the same thing, so I hope I saved somebody a bit of programming

I believe that should be enough information to get a fairly accurate idea of how well Rend does when thrown into a tanking rotation.

Indeed it is. I'm curious if anyone has tried to use this to gain a tps advantage over the course of a fight. I never have, as the half second does not seem to be worth delaying a shield slam cooldown for, but I was hoping someone had done some math to verify one way or the other. I usually just pop concussion blow early and hope for a sword and board proc.

With the current situation now it looks like in the not high health situation rend is only going to be an increase if it is used on the GCD immediately before Shield Slam refreshes as you eliminate any potential threat gain from a S&B proc then. On that same premise your Conc Blow and Shockwave will both have to be down or else you are better off using them.

Seems tedious to me like others have said. You gain a very tiny amount of TPS with the current best possible gear assuming you can perform the rotation flawlessly every time and you don't get strings of S&B procs that would push rend out of your rotation completely since you will have other specials to use first.

Its definitely something to watch though. Rend scales faster then devastate, there should be no doubt at all about that. The glyph pushes rend well ahead of a single devastate and it begins to approach the devastate + S&B proc in a non high health situation. In a couple of Tiers we could be looking at a significant increase, especially if our shield slams do not scale up very well which would relieve some of the tediousness of using rend. There is also no comparable 2.5 speed weapon to last laugh which might start to yield better results if we end up with a situation in the future where they do give us slow tanking weapons.

A 2.5 speed version of Last Laugh at max dps would cause glyphed rend's threat to push up above shockwave and conc blow and get awfully close to revenge + S&B. A combination of a high dps but slow weapon with deep wounds, glyphed rend, and devastate could start pushing against the TPS gains of faster heroic strikes if they itemize the end bosses with one.

It'd be nice if they fixed the improved rend bug too. It is very tempting to use at it's current values glyphed above 75% but the bug changes rend to above 90% unless it is corrected on test. Its not like my HS glyph is doing me much good in some of these fights now.

I can also second that. Here is a WWS log from the raid night with me having 56 exp (=14%) and haven't seen a parry.
Wow Web Stats

Ok so someone please summon up the updated shot rotation or priority list please? Do most of you pull and then charge for rasge or are you letting hunters in the groups pull? I like to pull unless there is a good hunter in the group that has misdirection up and ready. Opinions?

What do you think of this Build:

"Normal" High Threat Build

Arms Tank Build

You can use Revenge about 60% of the time and SS every 6 Sec MS if Revenge is on CD.
Tanking Weapon is Last Laugh.

What do i loose?

Shockwave
Sword and Board
12% Crit on Shield Slam
Dmg Shield
Devastate
2%Stamina/Strength
2 Expertise
2% Crit on Heroic
Critical Block(only Avoidance loose)
Focused Rage
Warbringer
2 Points in AttT

What do i win?

Revenge on a lot of GCDs (~70% Block/Dodge/Parry)
5% higher Crit DMG
5% Crit on normal Attacks including Revenge (which will be your one button spell)
MS if SS/RV are on CD

Focused Rage you dont need you get tons of free Heroics
Warbringer is nice for almost no Boss Encounter but nice to generate Rage fast, if your not yet tanking.
Dmg Shield is a low Part of my Dmg 4% (not yet having Deep Wounds)
Devastate is an inferior Aggro Spell (ok to build up Sunders it is nice) but could do another warrior too :P
Shockwave has a 20 Sec Timer and no use on a lot of Big Encounters (Sapphiron, Malygos ...)
Sword and Board is not as great as it sounds. 50-75% Reduction on SS CD (with no Delay and I often miss A SnB Proc because of the Serverlag)
Critical Block is around 7-8% of hits of which i will cut 1,1-1,3k less dmg.
the loose on the SS Crit and the 2 Points in AttT needs to be looked at.

Also you lack in AoE threat but in a 25 man you got druids who are even with Shockwave highly Superior.

I'll test this Build next Naxx ID and Report some findings.

It looks like complete shit. You've got 3/3 2h specialization and redundant bleed-improvements. You're dismissing critical block as bad when it's one of the best talents in the entire prot tree. There are just too many synergies you're missing by skipping deep prot (especially for ae tanking which is 1/2 of every instance) to justify spammable revenge and mortal strike. Post a wws and some anecdotal information from your next naxx run with that spec, maybe it will end up being fantastic.

I always like when someone thinks out of the box. I look forward to hearing about your experience with it. I agree with the incredibly rude Angeron that speccing for 2h is wasted in tanking, and those points could probably be better spent on improved rend.

My initial thoughts on it were similar to Angeron, but upon closer inspection you give up surprisingly little by skipping the end of the tree. However losing critical block is a fairly hefty hit to both mitigation and threat/dps, and I suspect that the extra damage gained through more frequent revenge won't overcome that. It's a shame that you need to take so many useless filler talents to get to that point in Arms.

Shockwave - As has been stated by this thread is a very high aggro tool that should be used on bosses as well as trash
Sword and Board - If used in a proper rotation a 1st or 2nd GCD S&B proc can greatly increase your threat.
12% Crit on SS, this is huge being as SS is your highest threat move as a prot warrior
Dmg Shield - Icing on the cake for threat and more so as your SBV goes up
Devastate - Due to S&B Dev > Sunder by a long shot
2%Stam/Str - Actually 6% Stam/Strength, given 30k unbuffed health at top end gear for T7.5 you would receive 1800 more health. 3 Talent points to essentially add more stamina than any single piece of gear will give you. Str will equate to attack power which further increases the damage done by nearly all of your threat moves.
2 Expertise - Actually 6 Expertise and we all know that for both avoidance and threat the reduction in parries is awesome. In fact without excessive stacking its difficult for a protection warrior to get Expertise capped without this talent.
Focused Rage - Even on the closest thing to near infinite rage fights (patchwerk) I still manage to drop to 0 rage quite often, without this one talent it would happen a lot more. The loss of 3pts on all of your abilities will stack up and become very noticable.
Warbringer - Not such a big deal on boss fights as a MT.
2 Points in AttT - As gear scales up this ability becomes even more effective for threat. This will quite easily outscale Cruelty as well as the gear gets better.


All in all, the high end prot talents are not worth missing out on for multiple reasons. These talents do provide a large amount of your threat as well quite a bit of your avoidance levels.

You also forgot to mention that your talent option misses out on Imp TC, and Incite. You could argue the spec would be good for an Arms/OT warrior. But why wouldn't you just play as arms and go respec to prot when needed.

Sorry, I hadn't personally considered it so it was new to me.

At the moment going full prot is just too good, shockwave is an extremely good talent. Pre-Wrath if you were geared and spec'd mostly arms you could tank most things with really good threat generation. Now there really is no need, unless someone finds out a way to beat shockwave/damage shield/critical block etc.

His spec looks like poor man's version of Failure's spec, which looks like this. More about it straight from the horse's mouth: Aphmau's UR Prot Spec Q&A.

He specced into Strength of Arms, mind you - so it's only 2% Stam/Str and 2 Expertise loss, as he stated.

The unrelenting assault prot spec still sacrifices far too much, for a marginal gain, if any. I've yet to see a Unrelenting Assault prot spec put out any decent numbers, but I haven't tried it myself (as I consider it a useless spec). If anyone's got decent parses of it, share!

I tested it one week during Naxx and pulled a little over 3k DPS. I did it when we were pushing for high parses -- the following two weeks I pullled similar numbers with 15/5/51 (two differant versions -- one with 5/5 cruelty and one with 3/3 AttT and 2/5 curelty).

"Failure spec"
Wow Web Stats

Note: I did not sunder -- instead had a HaT rogue do it.


I guess there are some advantages, as well as many (massive) disadvantages, to the spec. 1) You will pretty much never have problems with rage because of the revenge glyph. Of course, if you can't revenge, it's pretty bad (Gothik comes to mind). 2) You could probably call it "easier" than a normal spec.

But honestly, I really felt the loss of Warbringer on trash. Critical block is our defining mitigation talent, and that's gone. AE tanking really blows with the spec. I could probably not worry too much about that in a 25 man for the most part, but doing 10 mans I would probably have to respec.

All in all, I don't see too much of a point in the spec. The main advantage that Failure commented on was extra DPS, and I can do pretty similar DPS with a standard spec. On the bright side, it's what got me to start speccing for Deep Wounds.

Back on topic a bit

Yomon did you update your program to account for the 14% expertise cap as well as the 1sec GCD on Concussion Blow? I realize we could discuss TPS forever but I, at least, feel that these two factors can have a fairly significant impact on our TPS.