## How good is this?Edit

Let's base the calculation on the following presumptions:

- . The haste improves all casting, as well as the global cooldown, with 25%.
- . You continually cast one spell every 3 seconds on average.
- . Proc chance is 10% (as stated by Thottbot).
- . All spells are affected by the haste.

Then we have the following results:

- The number of spells that can have given a proc is 6 seconds / 3 = 2.
- The chance of not having an active proc is 90% ^ 2 = 0.9^2 = 81%.
- The chance of having a proc for a spell is 100-81 = 19%
- The average spell haste will be 19% * 25% = 4,75%.

Could someone else comfirm that the proc rate is or isn't close to 10%? With patch 2.1, the spell haste effect is expected to be substantially decreased. But a trinket that increase average spell damage with 2% is maybe the best there is? --LarsPensjo 09:38, 3 May 2007 (EDT)

- Currently about 10%, but itll be 2% in the next patch "iirc" -
**CJ**talk / cont 09:58, 3 May 2007 (EDT)

The 10% and 18s cooldown seem true.
To calculate you calculate the amount of time it takes you to proc on average and then add the cooldown to that time. The % uptime is the uptime divided by the time it takes you to proc it. Note that the uptime is almost always more than 6s when casting spells that can be hasted as with current mechanics if you start casting a hasted spell it'll remain hasted even when the haste effect wears off. I consistently get 3 fireballs off with it.
And it seems to NOT reduce the global cooldown, if you think you can cast scorches etc faster with it make sure you're using /stopcasting proprely to overcome lag first, otherwise your testing is pointless.
Overall this trinket is better DPS than most epics including icon of the silver cresent for any mage (even AB users as long as they don't spam 1.5s cast AB all day long) and probably for warlocks too althogh I didn't calculate for warlocks. —The preceding unsigned comment was added by **Galzohar** (talk • contr). 03:11, 3 July 2007 (UTC)

The internal cooldown on this is not 18 seconds, it's 45 seconds. The 10% proc rate is however correct. I've been using this for six weeks, I've saved every raid and instance combat log I've been to (about 100 in total), and whilst I have numerous time differences between acquiring the spell haste buff of 45.001s all the way up from there, there is not a single one below 45s. Additionally if a 45sec internal cooldown is assumed then I get a mean 10.002% chance of acquiring a spell haste buff on any spell cast which is "harmful" - whether that be CC, applying a DoT (not the tick of course, you don't cast ticks), casting a debuff, casting a direct damage spell, whatever - pretty much anything that the mob in question wouldn't be pleased to acquire. Mossy (talk) 10:32, 17 June 2008 (UTC)

Some corrected information from the first post:

- . This trinket has 320 Haste Rating, or 320 / 15.7 = 20.3821% Haste
- . Assume a cast time(t) of 3 seconds.
- . 10% Proc Chance
- . All spells are affected by the haste.
- . There is an internal hidden cooldown of 45 seconds.

Then we have the following results:

- The time required to proc this trinket is 3 / 0.1 = 30 seconds - this gives us a cycle-length of 30 + 45 = 75 seconds.
- This trinket has a duration of 6 seconds, meaning it is active 6 / 75 seconds.
- This gives us an average spell haste of: 0.203821 * (6 / 75) = 0.203821 * 0.08 = 0.01630568 or 1.630568%.
- The average Haste Rating given is therefor 1.630568 * 15.7 = 25.5999