RPPM Changes Will Change Our Stat Priority

I haven't gone through all the changes and how items are being "re-balanced".  (I'm still catching up on things since my vacation.) This affects the performance of my RPPM gear, the value of prepotting, and all the gems, reforging, and some enchants.  In a nutshell, this is a rather huge change.  Haste will not be as good.

Originally Posted by Blizzard (Blue Tracker / Official Forums)
There are some further changes to the RPPM system in development that we'd like to share with you. As you know, RPPM is something we've been doing a lot of iteration on, and we have another iteration that should be hitting the PTR soon:

As you may know, RPPM proc rates typically scale with haste. This was done because historically, attacking faster meant you had more chances to proc something, so got more procs, and we wanted to preserve that effect. However, most procs before RPPM were such that either their effect didn't also scale with haste, or their proc rate was predominantly limited by an ICD. Many of our RPPM effects thus far have had neither of these limitations, such that they effectively 'double-dipped' on haste, benefiting twice from it. In some extreme cases, the proc was designed such that they actually triple-dipped.

As RPPM effects have become more wide-spread and more impactful, this has caused a variety of problems. Primarily, it has skewed stat balances toward haste rather significantly. It's also a compounding problem where many of these procs stack multiplicatively with each other, causing insane burst when all of these procs go off together. That can be fun, but also raises the skill cap on your performance, and makes gearing choices more restricted to ones which stack together optimally.

For 5.4 we're going to change both new and existing RPPM procs to not double-dip on haste. Benefiting once from haste is fine and expected, but not twice. For example, suppose you have two hypothetical procs, Flamekissed and Villainy:
Flamekissed's proc rate scales with haste, and its effect says "Chance on hit: Deal 500 additional fire damage". This is fine, because only the proc rate scales with haste; the effect doesn't.
Villainy's proc rate also scales with haste, and its effect says "Chance on hit: +5000 Agility for 20sec." This is not fine, because both the proc rate and the proc effect scale with haste. The more haste you have, the more attacks you do in that 20sec period which benefit from the increased Agility.

If both the rate and effect of a proc scale with haste, we're going to remove the haste scaling from its proc rate. In these cases, we'll compensate for an expected amount of haste by increasing the base proc rate. For any procs whose effect does not scale with haste, their proc rate will continue to scale with haste as before. However, we're also revisiting the proc rate tuning on all existing procs that were made overbudget due to the addition of Unlucky Streak Prevention (which ends up increasing effective proc rate by 9%). These changes should bring RPPM procs back to being on-budget and tied with traditional ICD procs in value.

This will obviously have a noticeable effect on most players performance; don't panic. We're going to be adjusting damage/healing/tanking performance with these changes in mind.

These RPPM changes should make it to PTR soon, and you'll be able to find the exact changes to each such effect there. We now show RPPM proc rates in the tooltip of the effect, which should make it easy to find.

....

However, we recognize that not everyone likes the RPPM gameplay, and are going to greater lengths to provide alternative options in Siege of Orgrimmar. In addition to RPPM-based trinkets, there will be trinkets that use the older internal cooldown (ICD) method, as well as a few on-use trinkets if you'd prefer to have an even greater degree of control over when your trinkets are active. There's even a couple trinkets with completely passive effects.