They don't pick 25 random players for raids. What I meant is that each raid differs from every other even if you have the same composition and people playing 100% of the time. Dps will change based on a lot of things. That's why we have sims. They basically neglect every aspect of an encounter (or they could be made to account for most of them) and provide you with hundreds of hours worth of data withouth you having to do anything more than tinker with options for half an hour or so irl. They can account for almost everything. Unlike rarw, you can edit your rotation and possible downtimes, ua usage, etc. during a fight in kahories which makes it vastly superior and very accurate.
It's okay if you don't want to believe in it. I didn't some time ago but there's realistically no way for anyone to provide hundreds of hours worth of data to compare item A vs item B other than a sim. There's no point of comparing dps figures between two dks as results can be extremely deceiving in some cases
Late edit: http://imgur.com/X1EJE7G
Some dummy wacking. There was disease damage going off on another target and it equates to ~an extra 207 dps points.
http://imgur.com/Y07RLMg
The dummy one is from today, lk is from last month.

