Good Players, Bad Corsi

[seabass45]

Are there any instances of this happening, and if so, why?

The reason I mention is Rusty Klesla, considered a solid defenseman in spite of his recent concussion issues but generally looks bad according to Corsi for the last few years. As an Islander fan I haven't seen him play recently but if he's good, why would he have a low Corsi?

Click to read more...

 

[Larry Hoover]

Yes. There are tons of examples. Phil Kessel for instance.

The statistic is just a single metric. It's essentially a proxy to measure puck possession. Historically, it has shown that most teams with a high Corsi are good teams and most good teams have a high Corsi. It's far from a rule however. Some players are strong off the rush and just need a single shot to score. Others play a Corsi+ style of game and throw pucks on net.

Also, context must be considered. Some teams in general are better for Corsi, the situation in which a player is used, who they play against and with, etc...

 

[seabass45]

Yes. There are tons of examples. Phil Kessel for instance.

The statistic is just a single metric. It's essentially a proxy to measure puck possession. Historically, it has shown that most teams with a high Corsi are good teams and most good teams have a high Corsi. It's far from a rule however. Some players are strong off the rush and just need a single shot to score. Others play a Corsi+ style of game and throw pucks on net.

Also, context must be considered. Some teams in general are better for Corsi, the situation in which a player is used, who they play against and with, etc...

That's what I thought too. The guy who occasionally does stats posts for Islanders Point Blank is very hardened on Corsi, refusing to believe that Klesla might be a good player because "corsi bad player bad", in his mind, and that's been his guiding principle in evaluating all players. He hasn't seen Klesla play, obviously, and doesn't seem to care what anyone who has seen him play thinks, so all we have is Corsi without context.

(I'm not anti-Corsi, BTW, I'm a data analyst so all this stuff fascinates me. I just wish more hockey stats geeks were more skeptical and open to discussion.)

 

[wgknestrick]

There are 2 things that go into making a good NHL player

Shot quantity AND shot quality

Corsi is just one factor (and may be slightly more important and repeatable), but it is just as foolish to reject the 2nd factor.

Here is an excellent article explaining why:
http://hockeyanalysis.com/2013/09/16/why-corsipossession-should-not-be-the-goal/

Basically any forward that is one dimensional with his shot IE Kovalchuk (and probably on a bad team) would be a good example IMO. Most "good players" on bad teams will have this issue. 1 good player is not enough to keep the puck out of his own zone (and keep his shots against low).

As for the "best players" who fit this criteria; I compiled the last 3 years of data and compared GF% to SF% at 5v5. Any player that has "abnormally" high GF% vs their SF% will show up here. The mean is 1.0 (actually .995) for GF/SF

Kadri is #1 with 47% of SF, but 60% goals for (1.27 GF/SF)
Crosby #2
Halischuk
Hansen
Semin
Vrbata
Burrows
Mackenzie
Kunitz

(generally these are good players since they have high Goals For values, and just about every good player is on it)

Parros is last followed by
Jackman
Steckel
Winchester
Hall
Larose

IE bad players who are just way worse than their corsi would suggest.

 

[hatterson]

Klesla has also historically had average to hard competition (from a corsi standpoint) with average linemates (again, from a corsi standpoint).

Although his corsi rel isn't great, he does usually post a solid on ice shooting percentage. Granted some work shows that defensemen have a limited ability to influence this number, but he's been consistently above average on his team in this regard.

Phaneuf is another example of good player, bad corsi. His corsi rel has typically been below average on the Leafs, but you also need to consider that he plays incredibly hard minutes (consistency top 5 in the league from a corsi standpoint) and usually has linemates who aren't amazing from a corsi standpoint either.

Gomez is an example of the other side of this coin. Usually good corsi rel number, but an extremely low number of shots are converted while he's on the ice and most, if not all, consider him a bad player.