I totally mistyped DR, sorry for that :p

This is exactly my point. I can't really see how full-haste builds could be better than hybrid mastery>haste builds for "tendon fights". We don't have waiting time in those situations for we'll have HA always up anyways (considering you pick HA instead of the improved AW for your 75 talent for this fights), and the fights are so short that we probably won't have that "in the long run, this amount of haste will give you X more attacks because of the CD reduction". It really seems to be a situation in which having 15% more damage with mastery is more benefitial than having faster skills, since we'll be spamming HPgenerator+TV all the time with no waiting time (HA on) anyways.

That's my point. As someone who pick HA, I really think Mastery is way more benefitial to me on "tendon fights". I'll try and log my overall damage on heroic spine this week to benchmark my ideas properly.

Because min-maxing makes me wet in the pants!

Situation would be "I have so much mastery that the overall damage from a CS would surpass the Exo damage". But I'm not sure about that, since I have little notion of how crit works together with our mastery.


Got it! Thanks!

Yes, but I consider "tendon bursts" to be different than overall "burst phases". "Tendon bursts" revolves around that 20s and those 20s only. Having HA as the talent pick, we would have basically the same overall amount of attacks done (unless we have a really huge amount of haste). Stacking mastery seems to be a really good option along with HA for those situations - specially since it increases the damage from CS and HoW (which would be my two HPgenerator picks in the situation, considering I'd have AW up as well).



Thank you for all the answers!

Possibly yes. What it continues to boil down to is: it's based on the gear you are wearing at the moment. If there's some breakpoint for Haste (extremely likely), we don't hit it at L85 in final tier gear. This means it's unlikely to be achieved in MoP. Rough guess would be sufficient Haste to drop CS to (3 - Latency) seconds CD. That's a shade higher than 50% Haste strictly from gear/gems/enchants. Not exactly feasible.

There is no magical "hit X Haste, Y Crit, and Z Mastery and you have won the game." X, Y, and Z would change per person based on the exact gear, talents, gemming, professions, and other things that they have.

Also, stats slide in relation to one another. I've commented on this multiple times in the past. If you have absolutely no Crit or Mastery on your gear (impossible, but bear with me) and push Haste up to high values, then the value of Crit and Mastery will increase - sometimes to greater than Haste.

Real example:
If you did not have the 5% Crit buff in your raid it was entirely possible to get Crit > Mastery for some levels of gear in Cata. As soon as you added the buff, the value for Mastery jumped back ahead of Crit.

For these reasons I continually advise people - generate your own personal stat weights for your own personal gear. SimC and my spreadsheet both can handle this quite simply! Using some external master value is the equivalent of an 'off the rack' suit compared to a hand-tailored one.

I see your point. Well, I'll stick to mastery>haste + Holy Avenger for spine right now. It's so fun to be almost doubling everyone else's tendon damage!

I'll see if I can benchmark mastery>haste vs haste on my next run (if everyone else don't get too annoyed by reseting the fight :p), I'll stay away from trying to do a gimmick 20s benchmark as it would be too prone to errors. Sad thing is I'm currently playing on a 3g connection for the week, latency is a bitch.


Pretty good info, thanks for that!

Please don't bother. One comparison of (average kill time courtesy of WoL) 7.5 (10N) to 10.5 (25H) minutes doesn't provide any useful data. You can talk to the masters of statistics, but from what I understand, anything under 10,000 attacks has a very low signal to noise ratio. You need a very long test (or lots of short tests, like SimC uses) to reach usable averages.

Also "it worked at L85 in top tier gear" does not necessarily translate to "it works at L90 in starting gear".

I understand it'd need big numbers to be an accurate statistic, but a few can still be enough.

If I find a huge difference between mastery>haste and haste in a few tries, it should be enough to provide the data I want. It may not be perfectly accurate, but a "yeah mastery>haste gets a little bit stronger to me" or "lol haste all the way now" is enough for me!



Also, I have run a few SimC's and it looks like Haste is all the way better than strenght. I was able to get from 43.002 DPS to 43.098 by changing all 50 str's to 25str+25haste, all 25str+25haste to 50haste and the wrist enchantment from 50str to 65haste.

It may not be that much of a difference, but it's odd enough for me. I won't be changing fully into haste>str though, as latency can be a bitch and having str smooths everything out!

Yeah, but by doing a few rounds with each reforge, finding a big difference in results can be revealing (even though it may not be perfectly accurate). If, for example, I can reach ~6m dmg on a few tries by doing mastery>haste, but I can reach, say, ~7.5m or ~4.5m by doing haste - that might be noticeable enough for me to pick my favourite one!

If the difference is seemingly small, I shall just save the reforging money and stick with haste anyways as it's better for everyother fight.

I altered your quote. See what happens? Bad data might 'feel' good, but it's still bad data. A few is not enough.

An example:
You have Joe the raider in there with you on Spine.
Pull 1 Joe is eager to go and pushes himself to the max on Amalg. You get Amalg low very fast before you have sufficient blood and have to stop your DPS. Joe then hammers the tendon which pops in 10 seconds. Your DPS is a bit lower due to these limitations (less time able to hit targets).
Pull 4 Joe is bored that he's having to do the fight over and over again. He more or less autoattacks his way through the fight while watching cute kittens on Youtube. You are able to go full bore on Amalg and the tendons die in their last 10th of a second, allowing you full CD usage. Your DPS soars.
Pull 4 used reforge B instead of reforge A - so reforge B looks better. Was it better? Maybe, but you can't tell because there's too much mud in the water.

Another example:
You do a fight with Reforge A and again with Reforge B.
Reforge A does more overall damage/DPS. It wins, right?
However, you run World of Logs and look closely at it and find Reforge B did more damage to tendons, even though less overall.
Which is the winner, now? How does that carry across to any other fight that does not have identical timing/health? Would reforge B that did insane tendon damage in 18 seconds be as good if you had only 10 seconds on the target? How about 30 seconds?

We can appreciate the desire to 'Science!' Unfortunately, testing a single boss a few times does not qualify.

Edit:
I'll repeat something I've said in the past. I have literally seen World of Logs record from someone in crap gear that did more DPS than another log of someone in awesome gear. Why? Crap gear had a huge crit rate and a lot of procs (phenomenal landslide up-time, tons of DivPurp, etc), while awesome guy had a tiny crit rate and very few procs. On average the awesome gear wins by a mile, but comparing a very limited set you can get the wrong impression.

I'm not even sure what you're trying to say here. The purpose of SimC and the spreadsheet is to provide data that is as accurate as possible. Comparisons between two or more close builds are what they are designed for.

Anecdotal evidence and short trials can give you the impression that one build or another is good, but that may not be the case. Even if you are handy at math and can reduce some of the noise, the tools we already have at hand are faster and more accurate.

To the theorycrafters in this thread, short term data IS irrelevant.

The point of significant, thorough testing is that afterward, one can come here and make a statement that is very likely (>95% likelihood) to be true, such as "Haste > Str" or "This is the highest DPS build for Patchwerk fights". Statements like that are simply not possible with the type and amount of testing you're doing. It might be good enough for you, but not for a site and thread like this.

However, no one can stop you from doing tests that make you feel better, or provide you with a play style that you like better than the one outlined in the first post. What I do suggest, though, is that you keep those findings to yourself until you've proven them to the level recommended by those above.

------------------------------------

Regarding our level 75 talent options, I believe both tools currently show DP is our strongest single target talent. However, it appears to me that any time-off-target would favor HA and SW, even in a single target fight. Have there been any calculations to see how much time this would equate to?

Edit: I had old information. Thanks, Balhale.

No, SimCraft has both HA and SW above DP on patchwerk fights. Post mastery nerf anyway.

Edit: But yeah, any time off target during HA or SW cd will push them even higher above DP. Likewise fights with increased damage phases, where they can be timed to be used during that phase.

-Divine Purpose should only have one chance to activate once per finisher (which must land on at least one target).
-Source

From 5.0.5 change list.