Merciless Gladiator's War Edge still missing from the latest version.

On a side note... even Kungen is only 7150 / 21000. Based on 1000 rep for Archimonde and 400 for each other boss in Mt Hyjal even he would need another 5 id's and an Anetheron kill

If i didnt miscalculate, i should be able to help you out sometime mid october if i dont miss a single boss. drop me a PM around oct. 10th :P

That's strange; I'm 4150 (or something around that - roughly one full Hyjal clear away from the number you stated)/21000 and I've missed out on quite a few Hyjal boss kills; I'm also fairly certain that Nihilum was in Hyjal and killing Archimonde far earlier than I was. I'd wager there're some 10000~15000+ people out there, perhaps among the humans?

There's a human warrior at 17885/21000 so expect some testing in short, i'll bash it out of him.

Enough about reputation thou, get back to discussing relevant things :P

Ok, can someone explain to me why Relentless Earthstorm is worth more than Thundering Skyfire? Every time I go through the numbers on these two gems Thundering always comes out on top. Unless I'm missing something here I dont see how the increased crit damage can outweigh 22.82% haste (15.28% in 2.2) for a sword rogue. Even in 2.2 with all the nerfs, Thundering still comes out better.

The problem with TSD is that it simply doesn't proc all that much between the lowish ppm and internal cooldown, and also it has no constant effect like the 12 agi from RED.

Then you're doing your numbers wrong.

How about you provide me with the math to prove its superiority instead of wasting space with a post like that.

The math to prove its superiority is in the spreadsheet itself. Thundering Skyfire procs at 1 PPM and has a 40 second cooldown, to the best of our knowledge. Its uptime generally comes out to around 10%. This apparently ends up not being as good as Relentless Earthstorm. If you don't trust the calculations in the spreadsheet (which accommodate a lot more factors than any napkin math you're going to see or do yourself), that's fine, but don't try to just bully people into reproducing an extremely complicated calculation for your benefit when it's already in the spreadsheet.

I'm not bullying anyone, I asked for an explanation as to why Relentless is worth more than Thundering for a sword rogue and I dont need someone coming in and telling me I'm wrong without providing any answers or explanations as to why I am wrong.

You yourself say that its "apparently" better. I can take a look at the spreadsheet, but just looking at the formulas in there without any accompanying explanations isnt very helpful.

I cant see the benefit of using Relentless with the low critrate sword spec has, so that was my reasoning for asking why its given higher points than Thundering when you are looking at the optimal gear for a sword spec.

Giske, what ppm and internal cooldown are you using for your Thundering calcs?

There was a pretty heavy thread on this at:
Thundering Skyfire Diamond

But I think the core of the dispute here may be the numbers that people are using for the ppm and cooldowns on the Thundering.

Also from the thread, the gem choice is highly gear dependent. Some people were reporting that the Thundering WAS a better choice for certain builds and with a lot of raid buffs.

Okay, I'll take a shot. Here's a very very very very very rough comparison. First, assume you have 30% crit (raid buffed, unbuffed, doesn't really matter) and capped hit. Here's your attack table for a white attack:

0% miss
5% dodge
25% glance
30% crit
40% hit

Your average white attack will deal (0.75 * 0.25 + 2.0 * 0.30 + 0.40) = 1.1875, or 118.75% of your white damage range. Now, if you add Relentless Earthstorm Diamond, your crit multiplier increases to 2.06, and therefore your average attack damage increases to (0.75 * 0.25 + 2.06 * 0.30 + 0.40) = 1.2055, or 120.55%. That's an increase of (1.2055 - 1.1875) / 1.1875 = 1.5% white damage, roughly.

Now, assume that Thundering Skyfire Diamond has 1 PPM and a 40 second global cooldown. With a 2.7 speed main hand, a 1.5 speed offhand, and assuming Sinister Strikes once every 4 seconds exactly, you swing with your main hand on average 0.62 times per second, or once every 1.61 seconds. Since it's 1 PPM, that's an average of 0.045 procs per second. Thus it takes on average 1 / 0.045 = 22.43 seconds to proc it. Total uptime is 6 / (40 + 22.43) = 9.61%. Thus the proc averages out to 23.07 haste rating, or 1.46% haste, or approximately 1.46% increased white damage.

Now for the added effects: TSD provides you with 1.46% additional Combat Potency energy. Assuming you have Slice and Dice up 100% of the time with the same 1.5 speed weapon and capped hit, you're already getting 2.47 energy per second from CP. Thundering Skyfire will increase that by 0.04 energy per second. On the other hand, Relentless Earthstorm will additionally impact your Sinister Strikes. Here's a combat table:

0% miss
5% dodge
28.5% crit
66.5% hit

Prior to RED, you'd do (2.3 * 0.285 + 0.665) = 1.3205 or 132.05% of Sinister Strike damage per swing. With RED, that increases to (2.369 * 0.285 + 0.665) = 1.340165, or 134.02%. The increase is (1.340165 - 1.3205) / 1.3205 = 1.49%.

Thus, RED provides not just 12 agility, but also boosts to both your white and yellow damage comparable to what a TSD would provide on only your white. Both effects scale well with your gear, but I'd anticipate that RED would continue to be superior.

More on-topic for the thread, Aldriana, I wanted to do a quick hack in your sheet to figure out exactly what effect the hidden cooldown on DST would have. Based on your formulas for uptime for other procs with internal cooldowns, I think this is correct -- I changed the formulas for "DST Uptime" to:

=[DST Procs]*10/([DST Procs]*20+1)

Where [DST Procs] references the cell immediately above the "DST Uptime" cell. Could you tell me whether I'm on the right track or not?

Thanks Vulajin thats exactly what I was looking for.

Alternately, grabbing the same numbers out of the spreadsheet for my current gear (Sword spec, T5 level gear):

My white damage multiplier is 1.1911. With the Relentless Earthstorm Diamond crit modifier, it increases to 1.2106, an increase of 1.64%.

Meanwhile, TSD has a computed uptime of 10.95%, so provides benefit comperable to 26.3 passive haste rating; normally, this would increase my number of white attacks by 1.67%; however, since I already have some haste rating (from Dragonspine Trophy) which stacks additively rather than multiplicatively with the TSD haste, the true increase in white attacks is only 1.61%

So, RED increases my raw white damage more just from the 3% crit effect. Now lets consider yellow attacks. The crit bonus from RED provides increases my yellow damage multiplier from 1.4240 to 1.4465, an increase of 1.58%. Meanwhile, the 1.61% haste effect from TSD increases my CP procs by 1.61% - grinding the numbers on this, we find that this works out to .34% more energy regen, hence .34% more damage from yellow attacks.

So, what we conclude is: the crit multiplier, by itself, on RED generates both comperable white damage and superior yellow damage to TSD, *and* provides 12 agility above and beyond that.

Note that without the haste nerf, they were far more comperable, although RED still scores a bit ahead, particularly since it has better scaling properties in interrupted fights and pvp.

Looks right. For those of you wondering where this formula comes from...

Each proc lasts 10 seconds. The time between procs is 20 seconds of cooldown plus however long it takes for another proc to go off; DST procs computes the chance of a proc occurring in any given second. Thus, the additional time for the proc to occur is 1/[DST Procs]. Hence, this formula would normally be written as 10/(20 + 1/[DST Procs]) - the uptime per proc divided by the time between procs. However, since with DST unequipped, [DST Procs] is 0, this would cause a divide-by-zero issue; hence, the solution is to multiply numerator and denominator by [DST Procs] to obtain 10*[DST Procs]/(20*[DST Procs] + 1)

As an additional comment: make sure you change the DST proc rate all 18 places in which it occurs (3 times on each of 6 calculation pages).

Or, if you're just looking for a ballpark number, I made these changes myself and posted a summary of my findings in post 602.

My off-the-cuff estimate would be that there's about 20-25 rows per sheet that would need to be changed/modified/replaced/whatever per sheet to support simple Evis cycles (like 3s5e or whatever); Implementing something like 2s/5r/2e would require writing a full additional calculation sheet. So, it would be a decent amount of work, but something that could be done with sufficient demand.

The thing of it is, it's pretty straightfoward to show that no Eviscerate cycle can be optimal. Consider:

A 5 point rupture does 1000 + .24*AP damage; with my current raid-buffed stats entered in the sheet, I have, on average, roughly 3k AP; hence a 5 point rupture will do about 1720 damage - more if Mangle is available, which it is with some regularity; but lets stick with that 1720 number for the moment.

A 5 point eviscerate does 985 + .15*AP raw damage; with the aforementioned 3k AP, this works out to 1435 damage. 2/5 will raise this to 1635. With 3/3 Imp Eviscerate and 3/3 Aggression (which is sort of a rare talent spec, but we'll go with it), this is increased by a further 21% to 1978 damage. Based on the crit rate in the sheet, for me, 32.6% of those eviscerates will crit; hence, total damage from the eviscerate is 2623 damage. This is, however, mitigated by armor, and hence will only do .6956 times this much damage, or 1825 damage.

So, in the best possible situation for Eviscerate (wherein you have 2/5 plus all possible talents to boost it), and the worst possible situation for Rupture (no Mangle or Serrated Blades), a single Eviscerate does about 100 more damage than a single Rupture.

But...

Eviscerate costs an extra 10 energy relative to Rupture. Thus, to be fair, we should add 1/4 of the mitigated damage of a sinister strike to the damage of rupture for a fair comparison. Grinding a few numbers gives a base mitigated Sinister Strike damage for me of 1083. 25% of that is 271 damage, which more than covers the damage gap.

So, even in the most favorable possible situation for Eviscerate, Rupture still does more damage. Hence, the rule of thumb would be: if it can be ruptured, it should be ruptured. And since non-rupture-able bosses are the exception rather than the rule (you named all of them except Doomwalker, basically, and I even question the inclusion of Leo on that list), I decided it wasn't initially a priority; you're not going to focus your gear selections around non-rupturable bosses, for the most part. Which is not to say I'll never implement it - it's just a ways down on the priority list.

If i get some extra time today i will test 4/5 t5 for you, i would need to use two different speed weapons correct(single wielding each? or dual wielding two of the same speed?)? with no haste or specials?

For each test, use either 2 weapons of the same speed, or only 1 weapon. For the moment I would do so without haste or specials until some resolution is reached in the Proc Per Minute Mechanics & Haste thread, as there is reason to believe that haste might effect the proc rate (although I'm dubious, until a final conclusion is reached, it pays to be careful).

To totally pin this down, I would like to see decent-sized combatlogs at two different speeds of weapon.

Ok i'll try to do it soon, but no guarantee. I run my guilds website and recruitment so im usually pretty busy.

I don't disagree with anything that you posted. My curiosity is more about the situation in which you have too much energy for a Rupture cycle, i.e. where, by the time you're ready to Rupture again, your previous Rupture hasn't yet ticked off. I'm interested to know whether it would be better, in such a situation, to simply increase the CP used on Slice and Dice, or to switch to an Eviscerate cycle.

However, I think such a question may be beyond the scope of your sheet, which I understand is primarily intended to aid gear choices, not cycle selection (although it does a pretty bang-up job of the latter regardless).

Does this happen often? I'll often have 5cps and 2-3 seconds left on my last rupture but I'm also usually drained on energy. In those cases I just let my energy tick up and rupture when the last one falls off. The only time I can think of that it would make sense to evisc instead is if you have 5 cps, several seconds left on rupture and full (or nearly so) energy. Perhaps my gear is lacking but I've never run into that situation.