I didn't see this post until now and was in the other thread. Anyway, here's some information for people who are looking to setup a fair comparison between the stats.
The key is to keep the secondary stat budget the same across different setups.
This simulation has 35,000 secondary stats.
This simulation has 22,000 secondary stats.
This simulation has 16,000 secondary stats.
Because of the difference in secondary stats, of course the first one will do more dmg. To run single-sim tests like this, you'll want to make sure the secondary stat total is the same in each one.
I also setup a 'stats' batch run that leaves crit where it is and varies haste, mastery and versatility. Single target, default talents. Here's the report: http://www.askmrrobot.com/wow/simula...d615c0378e21a9
Here's another followup post I made in the other thread (yeah, I'm quoting myself, is that weird?!)
Before I post this, I want to say I am not trying to argue for or against a stat. I am not our theorycraft expert, that would be Swol over at AMR.
However, I *AM* trying to set up fair comparisons for testing, which is something I do all day. That being said, here's a comparison for your first example, the hypothetical one:
- Professor's sim with 35,000 total stats, 0 Mastery and 30,809 Vers: 317k DPS http://www.askmrrobot.com/wow/simula...af151a5aacb6e8
- A sim that also has 35,000 total stats, flipped to have 30,809 Mastery and 0 vers: 391k DPS: http://www.askmrrobot.com/wow/simula...af95dbf7d75603
- Equal amounts of both, 404k DPS http://www.askmrrobot.com/wow/simula...fa32b6bf82657e
But a better way to look at it is that batch report I linked, which professor, you pointed out that you could give up 2-4% DPS and get the other benefits from a little bit of vers. For some players that might be worth it. http://www.askmrrobot.com/wow/simula...d615c0378e21a9
edit: just so you all know what I'm looking at. On any of those regular reports, scroll down and expand the 'stats' section. You can see the stat overrides there.