In fact, beyond the haste cap, crit gives you more haste bang for your buck (on average) due to Nature's Grace, even ignoring the other benefits of crit. People probably intuitively prefer haste because:

(a) they prefer low variance returns on haste (e.g. they prefer 1% faster casting all the time to 1% increased chance of 20% faster casting, even though the latter is more haste on average).

(b) They can't guarantee appropriate haste buffs in their 25 man raids to cap out their hots with 5/5 GotEM, so they just stack haste as much as possible.

Personally, I stack haste so I can guarantee the cap with 4/5 GotEM (since I need 1 point from somewhere for Empowered Touch with 3.2 changes). After that, I am stacking crit, on the off chance I have to take over tank healing.

The problem resto druids have is they don't have a real auxiliary stat to stack past the cap. Crit and Haste both give very meager returns because so little of our healing is done with direct healing spells, 3.2 buffs to Nourish nonwithstanding. Mp5 would be ok, except it rarely occurs on spirit gear, and we don't really have mana problems as is. They need to make hots crit or add some sort of fancy gem which lets you reconvert a useless auxiliary stat into a useful main stat (like int/spirit).

Since I am making crazy suggestions anyways, I ll go out on a limb and suggest a new glyph for Regrowth which cuts its healing by X% (say somewhere from 25% to 50%), raises its mana cost, but makes it party-wide like PoH. Would be an interesting choice -- giving up tank HPS in exchange for a weaker PoH which leaves a long hot. I feel like we barely cast Regrowth anymore.

EDIT: Thanks for the correction! I dropped a decimal point, and in fact, you get about 0.2% haste from 1% crit. The reasoning mostly stays the same: crit is a better throughput stat after the cap, but perhaps people prefer low variance returns.

No it doesn't. The increase in NG uptime from 1% crit is absolutely marginal.

Let's say you got 40% crit on nourish. If you're spamming it, you get 3 casts off during one NG proc.
So, haste gained from NG is NG uptime*0.2*100%.
(1-(1-0.4)^3)*0.2*100% = 15.68%
Let's add 1% crit for 41% total:
(1-(1-0.41)^3)*0.2*100% = 15.89%

So 1% crit gave you an average haste increase of 0.21%

Edit: Or for a more general idea: 1%*20% < 1%*100% :P

I do agree that crit is better than haste after hastecap due to living seed mostly, but it is not more haste than haste, even if you ignore the difference in ratings required for 1%.

Yes you are right, of course. Fixed. Crit remains a better throughput stat, though, after haste cap.

You say you're aiming to be haste-capped with 4/5 GotEM, but I'm not seeing how that will be possible without full BiS gear. Looking at your armory right now, you're at 316 haste, while I'm currently sitting at 454. According to to Zoltair's post (linky: Resto (PvE) Healing Discussion) you'd need 541 haste rating to be haste capped with 4/5 GotEM, no Celestial Focus, and full raid buffs. Thats 90ish more required for me, and over 200 more for you!

Going from my current gear to full BiS gear gets me to 550 haste - but that involves two Algalon drops and a Val'anyr. Doing it before then isn't practical in my opinion, as you'd have to gem for haste and the spellpower loss would be too great.

Edit: here's the gear to which I was referring, sans the Algalon-25 belt: chardev.org v6 ~ a World of Warcraft character planner

The 4/5 GotEM haste cap is only 430 if you have 3/3 CF. Spending 4 points in the balance tree is kind of a backwards way to drop 1 point in GotEM.

The 4/5 GotEM cap -- with the haste aura and haste totem, but without without CF -- is 541 haste. We'll probably be able to hit that point in tier 9, but in t8 hitting that point requires some serious focus on haste.

541 Haste is a very easy amount of haste to take on Ulduar. You MUST use 2 pieces with crit instead of Haste Only to take 4t8 bonus, you can replace your chest for the Moonkin t8 chest for example, or if your guild already kill algalon 10 man, he drops the best legs in the game for resto druid, or the chest from vezax hard mode heroic.

For cloak, neck, belt, boot and rings, most of our BiS itens have haste

Neck:

[Sapphire Amulet of Renewal]
[Watchful Eye]
[Charm of Meticulous Timing]
[Evoker's Charm]

Cloak:

[Disguise of the Kumiho]
[Shroud of Alteration]
[Sunglimmer Drape]
[Sunglimmer Cloak]

Bracer:

[Grasps of Reason]
[Unsullied Cuffs]
[Bracers of the Broodmother]
[Esteemed Bindings]

Belt:

[Belt of Arctic Life]
[Belt of the Sleeper]

Boots:

[Boots of Fiery Resolution]
[Spellslinger's Slippers]
[Boots of the Servant]
[Arcanic Tramplers]
[Boots of the Petrified Forest]

Rings:

[Conductive Seal]
[Starshine Circle]
[Fire Orchid Signet]
[Starshine Signet]
[Lady Maye's Sapphire Ring]


With 606 haste you can take 2 points off GotEM, to take that probably you will have to use a Haste weapon, our BiS is [Staff of Endless Winter], our best Haste option is [Icecore Staff]. As i dont raid 25man, ill take the Icecore as soon its drops lol. with 609 Haste, you can take CF and 2 points on Revitalize, you will have GCD capped for HoTs and 1.15 seg cast time Nourish. Thats probably our best output healing build/itemization now.

I know that 606 haste is hard to get meanly becouse we are stucked with 2 crit pieces for t8 bonus, but its possible.

As Allinone has said, NB was created after the document, but it is factored into the spreadsheet Allinone is using. From when I tested rejuv at the beginning of this expansion with all the various talents, it appeared that it was additive with various other talents (GoN, Genesis, etc.). However, I could be wrong, though luckily it's very easy to test because hot ticks don't have a range on them like casted heals. As to the tree of life aura being additive or not, that sounds like a typo in the document. It's computed as multiplicative on the sheet.

Yes, I added the GCD to all the instants because, if you are at the level where you use this kind of optimization, you are likely to be chain casting. They may be instant, but one cannot simply discount haste decreasing the GCD. For instance, the base GCD and a soft-capped GCD is the difference from being able to roll 11 rejuvs on the raid and 17ish, which is significant. The spreadsheet Allinone is using has a "chain casting %" constant that can be used to customize how much you actually chain cast so that the effect can be adjusted. I'm not sure what value Allinone used, however. The sheet also caps it at one second, unlike the document. At the time, I didn't feel like writing out piecewise functions for the cap after how long it took to develop.

The reason why our haste values are so different is likely because of nourish. At high gear levels, nourish actually scales better on haste than on SP, in terms of HPS. However, I do not currently like how NG is calculated on my sheets. As seen earlier (I forget where), it was found that NG almost caps nourish out at 1 second, which would render haste useless. Currently, I'm using a very weak method which calculates the average multiplier on the spell from NG. I'm currently pondering how I would go about calculating %NG uptime on chained nourish casting. Another possibility for why haste might be valued so differently could be that you assume nourish is haste-capped because of NG, which is definately not true, considering that a crit must occur before the buff is applied and thus some % of nourishes you cast aren't capped. Also, the fewer casts of nourish one does in succession would decrease the % of all nourish casts that benefit from NG, as well as the possibility that NG can fall off even while chaining nourish due to an unlucky non-crit chain. So, now you can see why NG is giving me a rough time.

Firstly, haste sources are multiplicative with each other, i.e. 15% haste from haste rating and 5% haste from the wrath totem go like so:

1.15*1.05= 1.21 or 21%.

The cast time, in this case the GCD, is divided by the decimal form of the percent:

1.5/1.21= 1.24s GCD

If you're unsure of whether something has been discussed, it's probably better to search first. Reading all twenty some pages isn't necessary.

Yep, just realized that. Any way you put it though, the numbers on the first page are incorrect. The correct equation for haste is:

New Casting Time = (Base Casting Time)/(1 + (% Spell Haste / 100))

So here we go again:

Wrath of Air Totem - 1.05
Improved Moonkin Aura - 1.03
Gift of The Earthmother - 1.2

1.5 x 1.3 x 1.2 = 1.2978

1.5 / 1.2978 = 1.1558

Therefore you would need 15.58% haste or 511 haste rating to reach a 1.0 second global cooldown.

I'll point you here

GoTEM Mechanics Explained

which is the same link Nofair has in the OP. You are still calculating the 20% reduction as a 20% haste buff. It appears that the 20% of GoTEM gets applied after all other haste buffs, so instead of a .3 reduction, it ends up being less. In the post I linked, with 336 haste rating and 3/3 CF, GoTEM only subtracted 0.2641940, which is less than the simple .3 that would be expected from a 20% reduction.

Also, just a small typo in your calculation, you dropped a few 0's:

You had:
1.5 x 1.3 x 1.2 = 1.2987, which is wrong, but
1.2 x 1.05 x 1.03 = 1.2987. However, since GoTEM isnt the same as a 20% haste buff

1.05 x 1.03 = 1.0815

Try running your formula's again, this time with the correct numbers and calculating the 20% cast time reduction last, instead of off the 1.5. That should bring your numbers a bit closer. This calculation was hammered out with many tests when Wrath was launched, i'd be surprised it you could poke holes in its armor.

Am I crazy, or isn't wearing the Moonkin T8.5 chest a good idea for your offset piece? Basically trading tons of crit for tons of haste.

I've been wearing [Malleable Steelweave Mantle] as my off-piece because I only have 8.10 T8 shoulders, but I'm thinking about using the Moonkin chest instead.

Fir reference: [Conqueror's Nightsong Vestments], [Conqueror's Nightsong Robe], [Valorous Nightsong Spaulders]