The player with 4 points was near the bottom of that chart! Tells you everything you need to know about it's value
It's really annoying when people post charts like this without context, because it's guaranteed that people will misinterpret them and give advanced stats a bad name.
So some context:
a) While it appears the bar sizes adjust for ice time, they don't take into account the increased variance you'll find with low ice time and fewer samples. It's no coincidence that the guys at both extremes had little ice time.
b) Adding to a), one player in one game is a very small sample size. If you were to put error bars around each player's number they would be huge and immediately flag the whole chart as not statistically significant.
c) Adding to b), it's score-adjusted C+/-, which means data points when the game isn't close are weighted less. Effectively making the sample size even smaller than it already was, so while possibly more accurate, the error bars become even farther apart.
d) Zone starts. Bozak started 71% of his shifts in the offensive zone, Nylander 55%, and Kadri 43%. That is obviously going to have a significant effect on Corsi +/-. It would be alarming if Bozak didn't have a good Corsi +/- that game given the coach hardly ever had him on the ice in the defensive zone.
e) Quality of teammates/competition is not adjusted for.
Disappointed in this Sean Tierney guy for pushing these so hard.