nice post on haste and dot mechs:

Explanation of the new dot mechanics and haste | StratFu

It might be worth mentioning Improved Cone of Cold and Shatter for AoE intensive fights. Dropping NP and some point in fire (probably one in Fire Power), you can get Shatter and Improved Cone of Cold. The downsides are many, starting with it not working on anything immune to roots. Additionally, CoC costs a huge chunk of mana and does crap damage for fire specs. On top of that, the loss of NP requires you to find 398 more haste rating to get to the soft cap.

The benefits on AoE fights might make up for it, though. First off, with shatter, you only need a 33.33% crit rate to guarantee crits on frozen targets, which CoC would provide. The duration of the freeze is 4 seconds, so you can CoC -> Scorch -> Fireblast, which should give you an HS proc. Alternatively (or next), you can Living Bomb -> CoC -> Pyroblast! -> Combust -> Impact Fireblast to spread a huge set of DoT's to everything you are AoE'ing.

I haven't extensively tested whether it can break on damage, but from my tests it wasn't breaking. I did test for diminishing returns, and there don't appear to be any.

Anyway, I feel it might be worth mentioning in the Subspec Talents section, for an AoE spec or for encounters where you have time and mobs to gain a HS proc while the boss is immune to damage or something. There is also potential to drop 3 more points in fire and grab Ice Floes, giving the frozen effect a 50% uptime (4 second effect, 8 second CD), which would allow two such specced mages to lock down an entire group of mobs while dealing tons of AoE damage to them with their 100% crit rate.

So if i understand this correctly with the Pyromaniac bug then we're allready over the (raidbuffed) hastecap just by skilling 1/2 Pyomaniac and 3/3 NP, am I right?

Calculation would be:

1.05*1.05*1.03=1.135575

So a raidbuffed hastevalue (without any haste from gear) of ~13.56% wich would result in beeing over the hastecap.

Is that the fact or am I off with the calculations here?



PS: this is my first time posting in here, so please forgive me if I made mistakes.

Here's a chart for the dot duration with increasing haste.


Also, I noticed an error in the OP. In the "Haste and DOTs" section, it is mentioned that Flame Orb gains extra ticks through haste, while in fact it is not affected by haste in any way and always ticks 15 times.

I've posted some very similar calculations in the affliction warlock thread regarding haste and DoTs, so I will do the same here.

TLDR: the 12.5% "soft cap" is being over-emphasized as a gearing threshold that significantly alters the relative value of haste above and below. There is no compelling reason to treat haste different above and below this amount of haste.

I'll start by saying that I have used SimC to test this theory out. I started with the BiS 372 fire mage and then knocked haste off the gear until I was at 12.49% haste. Then I added a little and ran a simulation at 12.51% haste. As the OP notes, SimC shows a jump in the stat weight for haste from 12.49% to 12.51% haste. But, the actual DPS does not change between these two simulations. The gain you get is within the margin of error for a 300 second fight.

So, lets take a look at Living Bomb:
If you have 515 haste on gear, your haste is:
NWP (3%) + 5% buff + gear (515/128.05701) = 12.4994% haste

If you have 516haste, your haste is:
NWP (3%) + 5% buff + gear (516/128.05701) = 12.5079% haste

The interval of LB is the base, 3 seconds, divided by haste:
515Haste Rating: 3/1.124994= 2.666681 seconds
516Haste Rating: 3/1.125079= 2.66648 seconds

In a 300 second fight you will get:
515Haste Rating: 300/2.666681 = 112.5 LB ticks, or, 112
516Haste Rating: 300/2.66648 = 112.5 LB ticks, or, 112

You get an extra tick on LB at 516 haste rating, but the damage (ticks/fight) you gain from LB doesn't jump.

The only real "jump" is that the duration of LB at:
515 haste = 10.67 seconds = cast LB 28 times
516 haste = 13.33 seconds = cast LB 23 times

The GCD at 12.5% haste is 1.33 seconds, so, you gain a total of 6.67 seconds over the course of a 300 second fight. What can you do in that time if you are perfect? Not quite 3 fireballs.

Now, what do you gain from the extra tick on pyroblast in a perfect, ideal world? With BiS gear you will use pyroblast 16 times in a 300 second fight about 18 seconds apart. I took that info from SimC. You would gain 1 extra tick each time, so, 16 extra ticks over the course of a fight.

Lets add all this up. Reaching the threshold will, in a perfect world, give you 3 more fireballs (25k average damage) and 16 more pyro ticks (3419 average damage). So, the total would be about 125,000 damage. Now, we all know that hot streaks won't come perfectly spaced out - it is reasonable to assume that we will clip that extra tick half the time. Lets realistically knock this down to 100k damage. Now, lets say we aren't robots, that means we'll probably only squeeze in two extra fireballs. 75k damage sounds like a realistic total number for the gain for reaching the threshold.

75k damage in a 300 second fight is less than a 1% damage increase.

You raise an interesting point. I would note, however, that its hard to evaluate the relative value of different stats at the moment. Different tools disagree somewhat wildly on the relative values. Magegraf, for instance, tells me that each point of Spell Power is worth roughly 2.5x as much as one point of mastery -- making the decision much more obvious.

In fact, it shows that the value of haste plummets significantly after the soft cap, it still remains better than Mastery. I'm not sure what to attribute these vastly different numbers to, but there are some mechanics that are not immediately obvious that may not be factored into calculations for some programs. An example that jumps to mind is the fact that haste has a positive impact on mana regeneration during scorch phases. This is due to a somewhat unusual feature of Master of Elements to return mana on both Scorch and Pyroblast! despite their lack of mana cost. Until you reach the point where scorch no longer benefits from haste, haste increases your total scorch casts and consequently increases your mana regeneration (while increasing fireball damage but not DPM on the flipside).

I don't disagree with your math, but I think it misses the point to claim that the 12.5% soft cap is overrated because it only causes 1% damage increase. To get a 1% increase from a single point of haste is actually incredible. I think you do yourself a disservice to devalue small stat increases because of their minor impact on a holistic fight. It reminds me of the old joke that there are no big numbers because every number is just one bigger than the last. The fact remains that a point of haste is worth more below 12.5% than it is worth above.

You say that 3 fireballs is not a big deal at all. Looking at Lhivera's basic SimC output, there are 83 fireballs cast during a 300sec fight. Imagine if those 3 missing fireballs were misses. Thats almost 3.5% miss rate, for which you would need roughly 359 hit rating. I doubt you would suggest we could drop hundreds of points of hit rating because it would only lose us 1% dps.

Have you also taken into account the mana saved on less Living bomb casts?

EDIT: The mana saved means less scorches and more Fireballs.

That is true but at the same time you could argue that reforging/geming into crit rather than haste could increase your crit chance - thus giving you more master of elements procs as well as free pyros

I like the way you put this.

I think it is notable that when you run a simulation just above and just below this "step"... you don't actually see the expected 0.75% damage increase. Even in a perfect simulated environment we can't realize this gain!

I find it safe to ignore small hiccups in mechanics like this because the number of trials we will actually have in our gaming never even approaches enough to realize a gain from something like this. You are just as likely to do more damage from a string of lucky crits as you are from tweaking your gear up over this threshold.

All that being said, it is very interesting. Bane of Agony is even more interesting because of the way the tick damage is calculated.

We still don't know if haste is better or worse than mastery... I hesitate to say that maybe fire mages have a rare chance here to just customize their character to their preference with regard to those two stats! That would be crazy, though.

I have to agree with revulva that this soft-cap is over emphasized. A lot of people are confused by why gear optimization programs are telling them to reforge into more haste or pick up some haste gear even after 12.5% haste in a raid because of it (and partially the way the original post describes haste and haste rating's relative value after this soft-cap).

The thought that "Oh I've got the soft cap, haste is kind of lame now since I can't reach the next cap without a ton of haste gear, so I should spend more time getting other ratings" is something that keeps popping up on this forum and many others. If you ignored the soft cap entirely and stacked the stats as suggested by the values given after the cap, you could VERY easily get to the cap with half raid epics. This is compounded by bugged Pyromaniac making it even easier.

Breaching this cap is however definitely worth noting. As Kavan pointed out this huge of an impact for such an almost insignificant change in stats is a huge deal, because this tiny change can make a huge difference in damage done over the course of a longer fight. Even if the fight or test period is small, the difference is quite noticeable when you are faced with a wipe where the 6 minute boss was just 100-200k from death but hit the enrage timer. That could have been a kill by virtue of a Mage in the raid, opting to get just one additional point of haste.