Edit: This is updated and is current for 3.3.3.
Hopefully you all will find this at least interesting and possibly helpful:
This is a chart of all current balance dps talents, glyphs and set bonuses that I could work out a way to test (either with Simcraft or reasonably exact data). 4t9 was not included because of it's additive nature, so it was nearly impossible for me to determine how much it added. The ordering is done by average dps per talent point (or in the case of glyphs or set bonuses, just the dps increase).
: I only calculated the personal dps bonus. Even if the talent provides a raid dps buff, you are assumed to have it from some other source. For example: Improved Moonkin Form on this chart is only showing the spell power benefit, the haste is a raid buff and is ignored. Moonkin form isn't shown at all because the personal benefit is 0 (assuming you have some other source of 5% crit).
I used Simcraft to calculate all of these. The base wowhead profile I used was this:
The spec was chosen so that I could more easily move talents around to get values and to provide a more consistent baseline.
My original result for Celestial Focus was weird (the final point showed an actual dps loss). This is a simulation artifact - what happens is at certain very exact values of haste, adding more can be a dps "loss" because of interactions with Eclipse, latency and spell queueing. In real life these will not happen reliably (although you can use Simcraft to find and plan around them if you wish). In order to get a better, more useful number for Celestial Focus I changed a gear piece to use a difference haste value (697) and that is represented as CF_697haste in the above chart - if you are looking for Celestial Focus numbers this is what I would use.
Keep in mind that the above chart is based on my current gear and isn't going to be accurate for you - however, it should still give a good idea of the relative value of everything displayed.
Edit: If you are interested the raw data I have uploaded it here. The formatting is bad (non-existent, really) so view at your own risk.
Lets Make it Offical
3 months ago