This is partly why I tend to disagree with the whole "versatility is your worst stat" bandwagon on all the guides. Sure, it does not increase your damage (or healing) as much as the other stats. On the other hand, it reduces the damage you get from all those AOE damage you will take during every single fight and thus helps your raid a lot. But I guess it's hard to value.
Blizzard apparently does their sims with bots. Dunk a whole raid on a boss with scripts to pilot them using "normal" clients, and see how they perform. Someone should try to get sims more like that.
I never have been a stellar dpser, but I generally find that from my experience, what is simmed is often what happens in reality.
If you have a gap of 5k dps where 30k is top, it's 1/6th, or over 16% difference. Unless you are constantly moving during the fight and/or everyone plays perfectly on every global available, both unlikely things to happen, let alone at the same time, relative values of the classes are respected. Gimmicks are gimmicks and don't count, but if the time on target is similar, real-raid results mirror the sim results. Having played a shadow priest in MoP, there was no way for me to top any meter whatsoever, except when playing with people 30 ilvls below me.
What I find baffling is the huge difference from top to bottom and the lack of actions taken to correct them for a long time, being an ex priest I can definitely notice that Blizzard never gave much of a fuck and if you happen to be playing one of the sucky classes, tough luck waiting to get "fixed".
The same is true for tanking or healing, some classes end up being much better than the others for this or that reason and (equal skills on the table assumed) if you raid the hardest content, where you already wipe hundreds of times, you want to squeeze every little advantage you can get. This wouldn't be an issue except that people that run LFR refuse to heal this tank or accept that healer because it sucks (when they are abominable players and should just stfu), this becomes the common knowledge (ignoring gear scaling) and widespread idiocy takes precedence over sensible reasoning.
Reading monk forums now, you'd think Brewmasters are made of toilet paper, but they can tank very well any available content (much better after the last buff), the reality is that most players have their heads so much up their asses, it's amazing they can still breathe, they can't tank normal, god knows who did that silver proving grounds for them and they can't face heroics because they can't play juggling the various short cooldowns.
Another wonderful common "mistake" is taking the guide on a website and think that's the only way to play. "Brewmasters have to pick Serenity" (why?) is already ingrained in the head of wannabe monk tanks that go spouting their bullshit on the official forums, the reality is that if you can play with both arms and are not brain-impaired, Chi Explosion is better in every conceivable way: you purify every time you use it (which is usually every 3rd or 4th global cooldown), you have permanent shuffle and you do a fuckton of damage *and* you can decide to do it single target or in ae. Imo a tank that does 10 to 20k dps is better than one that does 8 to 10k.
Serenity is a cheap way to avoid playing the BM mechanics by stacking 7-8 globals worth of shuffle so for 40ish seconds you can ignore and just spend chi on purifying. Woo hoo, so skilled man...
Stat priority is also out of whack for a lot of people: "avoid haste and crit, get mastery and versatility", good idea, so you are chi starved and basically can't use elusive brew? Amazing plan there, those 3% more staggered damage is going to save the day...
Ok I whined enough, I just wonder when people will realize that playing and testing on their own is better than reading and regurgitating the (unverified) bullshits you read on some website like it's the fucking bible of wow.