Stuff like Win shares is good for putting players in brackets.
Like, Demar Derozan is not good. He has a well below average ws/48. Good players don't magically have bad win shares.
But trying to compare to elite level players with elite level WS/48 doesn't do much. The defensive win shares just divides team defense by 5, so Dirk gets to benefit from playing with better defenders than him while Kobe played with some bad ones those years (and Kobe has always been a much better defender).
I'm a pretty big Dirk fan as people here can attest (and argued he was the 2011 MVP). Don't think he was ever better than Kobe except for the 2011 postseason run (which Kobe was slowed badly by a bad knee and Dirk went god-tier) and '04-'05 where Kobe was hurt and missed a lot of time.
Dirk couldn't take a team playing massive minutes to Kwame, Smush, Walton, Cook, and George to the postseason in the West those years. Only other players that could have done that in the last 8 years were Lebron and Duncan (if you swap Odom for an equal guard).
edit: I should also say Dirk, as most bigs, benefit from players creating for them a lot more. Granted Dirk for himself creates a lot more than most bigs.
Yeah, win shares are really weird. PER is usually pretty accurate, though it has it's faults too
PER is dangerous. It overvalues shooting and ignores most defense.
There is no end-all statistic. You can look at others like APM and SPM, too.