Concerning passive haste rating items like the Blade of Wizardry and Quagmirran's Eye with a 15% chance to proc for a ~25% and ~30% (respectively) increased casting speed, how does that stack against weapons / trinkets with more dmg and or with %crit/%hit. Anyone have some math on the subject? Ie how much haste is equal to 1 dps or something along those lines. I'm personally using both the sword and the eye and greatly enjoy having haste up on a semi constant basis but I'd like to see the math on the subject. I do realize the effect changes based on spec (10/48/3 current - mage) but there has to be some common ground on which haste ratings can be given a rough dps estimate.


Blade of Wizardry
Binds when equipped
Main Hand Sword
31 - 117 Damage Speed 1.80
(41.4 damage per second)
Durability 105 / 105
Requires Level 70
Equip: Increases damage and healing done by magical spells and effects by up to 150.
Equip: Your harmful spells have a chance to increase your spell haste rating by 280 for 6 secs.


Quagmirran's Eye
Binds when picked up
Unique
Trinket
Requires Level 70
Equip: Increases damage and healing done by magical spells and effects by up to 37.
Equip: Your harmful spells have a chance to increase your spell haste rating by 320 for 6 secs.

Well in order to spur some converstation on the subject - my poor math skills end up with a rough estimate of a 4.5% DPS increase for the swords proc... I ran the topic by another another mage and we hashed it out a bit but he agreed that 4-5% increase should be the end result. From my limited testing it appears the procs do stack, however there does seem to be a PPM limit on the sword and trinket as I didn't seem to get back to back procs but any addtional info on the subject would be greatly appriecated. I'll be downloading proc watcher and see if I can come up with some more stable numbers.

Proc per minute is a term that applies when defining a percent chance to proc on different speed weapons. It does not mean that there is a hard cap on number of times per minute that a proc can occur.

I would be interested to see if they implemented an actual proc per minute on those items however, with longer spell casts having a higher % chance to proc the haste effect.

If there is no actual way to get procs on successive spell casts, what you could be observing is a "cooldown" on procs, similar to what is seen with Windfury.

*edit* Nevermind - http://www.thottbot.com/?sp=33297 if correct, indicates a flat 15% chance to proc. Same with http://www.thottbot.com/?sp=38319

The Blade of Wizardry has been said to have a 30s cooldown. Between that and the 15% proc rate, you can expect a proc every ~50s on a 3.0 cast time spell. You can do better if fishing for procs or using a number of instants, but 6s out of every 50s is not so hot. Far better to get a solid raid drop (Prince's dagger) and use the 2+kg you would spend (or would gain by selling it) on epic craftables and gems.

It appears the sword has a 3 proc /min (maybe a tiny bit lower need to run the proc watcher mod over a longer period of time)

I've heard of the 30 second cooldown on the Blade of Wizardry, but I'm wondering:
(a) does the Eye have a similar cooldown?
(b) can we get confirmation that one proc would start the 30 second cooldown on the other?
(c) if they do in fact share cooldowns, all having the Blade would do would be to push the proc from once every 50sec~ closer to 40sec (in essence having a 30% chance to proc a haste from either weapon or trinket, and both having the cooldown). Meaning, go get the trinket because the sword doesn't add marginal value?


If this is the case, then can we conclude that they don't share 30 second cooldowns? Because one item alone can't proc 3 times per minute if it has a 30 second cooldown on the proc. Is the cooldown less than we suspected, or were you using two items and getting procs from both?

Edit: I'll admit that I may have a chance to buy the sword for 3k and am seriously considering it. I will get the Prince Malch dagger when it drops though, and am wondering whether it's worth the investment (I'm fond of how well spell haste scales with gear) and if it stacks with the quag eye trinket.

The blade doesn't have a 30 sec cooldown.

It's pretty sexy for affliction warlocks (what I'm using it for).

As far as % dps increase, you could quantify it, but that would belittle the benefit you derive from UA spam in PvP and/or SoC spam in either PvE or PvP.

To cross-reference some rogue discussion about the trinket Hourglass of the Unraveller in the Rogue DPS Spreadsheet. Starting at post #546, continuing for a few pages, and ending about post #571.
It was concluded that items with a chance to proc a buff for x seconds have a hidden cooldown of: 3 x length of buff. The hidden cooldown for both of those items should be 18seconds.

So I went ahead and bought the Blade last night for around 2700. I'm pretty happy with it. My only concern would be once I get Quag's Eye, the cooldown will be initiated for both. At the very least, I have an amazingly expensive torch main hand.

Thanks for the link though, and I'm glad that there's an underlying rule to the madness, rather than arbitrary values (if you count Windfury as a "1 second" proc, then the 3 second built in cooldown would match this rule?).

I'd probably rate Quag's eye as the best overall DPS trinket now that the lightning capacitor is being nerfed. I love it as an affliction warlock, even though I feel like I don't get as much out of it as I would were I destruction. It's awesome when it procs off my last DOT in my rotation and I can fire off 3 Shadowbolts at 1.92 seconds each, but it's not so great when it procs at the start of a fight as I'm casting a bunch of insta-cast DOTs and Curses.

I probably wouldn't use the sword unless I had more money than I knew what to do with. The proc is definitely nice, but picking up another ~30 damage, stam/int, and crit probably outweighs it, especially if you already have a Quag's Eye.

I have both and they don't share any cooldowns. I've had them proc off the same spell, and I have had the trinket proc on the cast immediately following a sword proc (and vice versa).

This is so sensitive to context, it's hard to come up with a rule of thumb. It depends on what spells you're casting; thus on your class/spec. It is a function of your DPS, so this scales with your gear (and a = 23 +dam kind of evaluation depends on your gear).

The easiest one to figure out, imo, is a frostbolt spamming mage. to begin...the proc would happen on the 6th spell (6*2.5+.2lag = 16.2 sec); 6 sec of haste; 12 sec of cooldown; repeat;

so that's 6 of 34.2 sec hasted...17.5% uptime * haste%...4.4% for sword, 5.25% for trinket for a frost mage.

using Vontre's spreadsheet (http://www.radiationnow.net/wow/mage_dps.xls ) with his default stats and his deep frost spec, frostbolt spam does 527 dps. 4.4% of that is 23 dps. the extra 28 +dam, 21 crit rating & 15 int provided by the Nathrezim Mindblade adds 17 dps.

So, for a frost spec'd mage, the sword is 6 dps better...which is about the same as +14dam.

it's a lot trickier for other specs/classes. I think new haste procs don't reduce the GCD, so scorches, fully ramped up AB, instants gain no benefit from it. however, those faster spells done to begin might cause the proc sooner?

Followed the same math you provided above but swapped numbers around for pure fireball spam and ended up with the sword doing 5.5 more dps than the talon of the tempest (best dps main hand around - put the best craftable gems into it)

The way I've been thinking of this is that a proc gives me X "extra" seconds of my damage spell. Calculating estimated damage added by the proc and then working back towards a %.

Assumption ** If you have any fraction of time left on your haste buff and start a spell, the entire spell gets hasted (much like how the AB-AB-FireB-FireB-AB rotation takes advantage of this rule). Thus, even with marginal casting lag (.3 seconds), I can still get 3 hasted fireballs off when I'm chain-casting fireballs, resulting is an effective haste duration greater than 6 seconds. This is really the kicker, as pure "uptime haste %" calculations don't show the true benefit of haste buffs.

Because a sword hasted fireball casts approx. 0.7 seconds faster, a "proc" results in 2.1 "extra" seconds towards additional fireball spam. (for now, let's assume one of my fire mage flunkies is in charge of maintaining scorch)

From there it's a matter of taking the average number of procs per minute, and then calculating the "extra" dps:
A- Let's say with the cooldown and casting fireball, I'd see a proc every 40 seconds (18 uptime + cooldown, 6x3+4 seconds buffer). That's 1.5 procs per minute.
B- 2.1 "extra" seconds of fireball
C- For now, let's assume 600 DPS for fireball spam (obviously, the higher the better)
D- 600 damage/second * 2.1 seconds/proc = 1260 damage/proc
E- 1260 damage/proc * 1.5 procs/minute = 1890 damage/minute
F- 1890 damage/minute * 1/60 minutes/second = 31.5 "extra" DPS

31.5 "extra" DPS from the proc would be 5.25% damage added.

28 +dmg, 21 crit rating and 15 int would add around 17-18 DPS as Stein mentioned, or 2.91% damage added above and beyond Wizardry sans proc.

What's interesting is that, if you have 800 DPS from other pieces of gear, the comparative numbers diverge. Assuming you'd stack balanced +crit and +dmg to reach 800DPS, the +dmg would do less for you %-wise, while the haste rating would do more (42 "extra" DPS from wizardry, vs. ~21 from mindblade, if the above calculations assumed 800 DPS instead of 600). Am I correct in assuming that haste scales with +dmg/+crit/+hit, which are all far easier to acquire? Thus, because haste is the rarer of the caster DPS buffs, we should be seeking it out when possible.

Now, I don't consider myself an expert theorycrafter, so please excuse any gross assumption or logic errors. I know that chain casting fireballs is far from reality. Chain-casting frostbolts is a much better example, except that I'm a fire mage, so please forgive =). I guess I wanted to point out that assuming a mage can sneak in a final spell with 0.2 seconds remaining on the spell haste, it effectively extends the duration of the spell haste buff. This is a pretty awesome thing!

I've been working on a simulation to approximate the value of various statistics (compared to a baseline of any arbitrary one) for use with a mod which weighs stats and ascribes a singular value to any item you look at in game... I'd been intending to for a while but this thread prompted me to code in the Quagmirran's Eye proc to see how it would size up, and I got the following results:

(for reference the simulation is coded to do dps following the same logic I do in game, roughly... if no immolate, immolate; if immolated and > 3.0 seconds, incinerate; if immolated and < 3.0 seconds conflag; if oom, lifetap, etc)

Run #1: Base Stats
Run #2: +10 Spell Damage
Run #3: Quagmirran's Eye Proc enabled
+1 spell damage is worth (727.92 - 722.13)/10 = 0.579 dps
Quagmirran's Eye proc is worth (751.01 - 722.13) = 28.88 dps
Thus QE is worth 28.88/0.579 = 49.88 +dmg

Take the results how you will, as the sim is still somewhat under development.

Edit:
Above results assume no cooldown on quagmirran's eye, just a flat 15% chance to proc on harmful spells. Also, if I were feeling really adventurous I'd run the sim for many many different fight durations for all types of stat increases, then integrate the differences in the resulting graphs to get an approximation of net worth, but methinks that's a feature I'll put off writing for a bit.

Edit #2:
For kicks I recoded the proc to have an 18 second cooldown:
Which takes the proc's value down to ~32.21 damage/healing.

Edit #3:
Recoding yet again(!) to have the proc effect reduce the gcd brings the dps average back up to ~751, so again worth ~50 damage/healing.

Haste does affect GCD, i get a 1.1second cast UA etc. which can be chain casted. Heroism(bloodlust) and quagmirrans eye also produces 1.4second shadowbolts, which can also be chain casted.

Been wanting to test what would happen with 4/5 improved corruption and a haste proc.
Anyone tried it ?

I'm not sure if it is different for heals, but counter to Vodrin's experience I found that as a paladin when the spell haste proc on the Scarab of Infinite Healspam kicked in I would run into the global cooldown a LOT. I have a 1.5 sec cooldown, and I usually spam a 1.5 sec heal just by rightclicking on folks, so when my heal was about 1.2 second I hit that .3 second window of cooldown but not casting quite often (distinct error messge generated when in cooldown rather than casting).

I'm trying to get an idea of how much to value mystical skyfire diamond now (relates to how good T4/5 is compared to spellstrike too).

New Skyfire - 5% chance to proc; next spell 50% haste; effect lasts 4 sec? (with assumed 12 sec CD?)

If so, assuming imp fireball spam, 20 casts to get a proc -> 60 sec, 1.5 sec of "uptime"; 12 sec CD; repeat;

so basically 1.5 sec saved every 73.5 sec => 2% more dps; using kelben's 873 dps, that's 17.4 dps.

if that's right, it makes a lot of stuff way better for trash and nearly as good for bosses as the spellstrike hood.

Does anyone know if the arcanums of rapidity function correctly now spells, and do they effect. If so, how significant an effect would having 2 arcanums + glove enchant have on the gcd?

You're wrong about Quagmirran's Eye.

I get "That spell is not yet ready." with chain casting scorches.

There's a lot of bickering back and forth on this topic, but if it doesn't reduce the GCD then, for the purposes of increasing DPS, haste rating is mostly useless for any build or sequence that is primarily based on 1.5s casts. (Flash spamming healers, scorch spamming mages, and and I'm sure there's others too.)

In reality there's probably a small benefit, since lag usually loses you some cast time anyhow. (so you will probably see some increase in DPS.) But it's clear haste rating is far superior in (for example) fireball/frostbolt specs where your cast time is not likely to ever fall below 1.5s.

Honeslty, this feels a little broken to me, but I'm sure Blizz has their reasons.

Here is my theory to explain the two sides of the "haste reduces GCD."

Haste does reduce the GCD, BUT the GCD mechanic differs from the spell duraction mechanic in that using /stopcasting doesn't actually clear us from the effects of the GCD. Did that make sense? Essentially:

People that say the GCD is shorter are waiting for their cast bar to fill up completely. At which time, the client allows them to cast.

People that say the GCD is NOT shorter are using /stopcasting macros. They hit their /stopcasting macro button before their cast bar is full. This stops the spell cast, but the /cast Scorch that immediately follows it is attempted before the client thinks it should be allowed (even though it would work if the stupid client would just send the message to the server).

Does that sound plausible?