well thats a point i would disagree with, you aren't entirely wrong with your argument, however you could just aswell say "a ninja is even more dependant on a good team, therefore in an ideal world he should get something out of it", (btw, an "average" ninja would by definition sit around the 50% percentile, the best group in the world won't push them to 99%) also yes these numbers are an average, of the literal top 1%, at this point samurai is superior in 99% off all cases, i really don't see why we should have this as the intentional design goal because "at the top 0,1% ninja may actually pull 20 dps extra" Also one should never take max parses for comparisons, at the very very VERY top their will allways be some way to cheese something, things like intentionally dragging out a fight 5 more seconds so someone gets his big 1500 potency finisher in or catering to a class in a way thats totally unrealistic like throwing shield after shield on a blackmage so he can stand in shit and whatnot
that one actually has some merit, however it misses one crucial point. if we are at the top percentiles there is no dead. you won't pull 99 or even 95 if you died, if you pull 95 with a death you obviously where on pace for world record before, got rezzed in seconds and where not at a critical part of your rotation where suddenly you lost all resources. there will be a bit less of a loss most likely, but seing as even for the really strong raid contributors buffs are like 11-13% of their total damage, that only dampens the actual penalty do to dying by a miniscule amount.
mind you, there are obviously also lower levels of play then 99% or even 95% parses, in fact out of a 100 parses 94 will be worse then 95% but thats exactly why i asked "if even in a best case scenario it is worse" , the aspect of "deaths being less punishing" is true, but only very slightly, and it gets less true the higher up you go.
In an ideal world i would say raidbuffs should overtake personal dps at the highest level of play while falling slightly short before that, you can actually see that happening with bard and mch as at 10% percentiles bard is about 1,5% below mch,than very very slowly is catching up until finally overtaking mch at 99% (by like 7 dps but whatever) in a perfect world i would say bard should be 1% below mch at the low end (so 10% percentile) while being somewhere about 0,5% above it at 95%+. Now obviously there are more than 2 classes, and even i don't think it would be realistic to have the real dps with everything included down to this level of optimization, but we are talking design goals here , not "is it realistic to hit that design goal exactly on point"
and i would agree, it is interesting, and also i don't think the numbers are bad, i don't agree with them exactly, like i think (no matter what number that may be) that sam should actually be equal to blm, maybe even slightly above, as addle is more usefull than feint in most cases and i would also say the "mobility" tax is still like a 100 dps too high but that is arguing semantics, balance obviously can allways be improved and at least some things are subjective and thats fine, the whole reason for such a forum is for feedback and discussion