The Smythe is probably the trophy where raw counting is the least accurate way to view things. Honestly,
@Hockey Outsider 's matrix does a good job so that x=/=x.
Roy had 4 runs where he could have reasonably won the Smythe (1986, 1993, 1996, 2001), but only one of those was so elite that no one else could be argued for it (1993). 1986 is his next best run.
Joe Sakic could have very easily won both 1996 and 2001. Patrick Roy could have easily won both 1996 and 2001. Neither would have been historically odd votes, even if
nothing on the ice changed.
Roy gets a lot of reputation from 3 Conn Smythes, but he could have easily been a 2 time winner without actually him doing anything different.
Conversely, Gretzky could have won 4 Conn Smythes without actually doing anything different. Hell, Gretzky could have won in 1983 (10 points ahead any Islander, tied for Potvin for +- lead) and we wouldn't view it as the worst Smythe ever.
Is Crosby a worse playoff performer if Letang wins in 2016? Is Kucherov better if he wins in 2021? Is Richard viewed even better if the Smythe is created in 1940 and he is a 3 time winner?
The on-ice results matter more than the binary trophy result. Otherwise we get into a weird situation where Crosby 2016>Crosby 2008 or Crosby 2009, when anyone watching the games could clearly tell you Crosby 2008 and 2009 were far better runs.