Sunday, October 12, 2008

This Week In Corsi

Perhaps this will become a regular feature. Perhaps not. And while it's only two games worth of data, here's a look at the Corsi Rating for each skater on the team on a per game, per sixty minutes rate:

Not good, Flash, not good (and not surprisingly, the five skaters on the wrong side of zero here are also the only Caps on the wrong side of zero in plus/minus).

9 comments:

Anonymous said...

Aesthetically, with players who are in the negative you should put their names on the right side of the zero line.

That would make that chart look SO cash.

JP said...

I couldn't agree more... but I'm limited by my Excel skillz. I'm working on it, though.

JP said...

There. That's better.

exwhaler said...

So, according to this, Flash had an ineffective game last night? I would question the value of the stat based on a game-to-game basis.

Also, "small sample size" are words to live by when using statistics.

JP said...

I agree on sample size, ex, but at some point it becomes a trend, so it's worth watching. Through two games, the top two lines have identical +28 combined scores and the fourth line is a combined +11. The third line, however, is -18, and I don't think you need a metric like this to tell you it's been the worst line of the four at even strength.

exwhaler said...

Sorry, but Fleichmann played a solid and aggressive game last night. The fact the Corisi Rating says he was the worst offensive player on the ice makes me seriously consider discrediting the value of the stat for individual players. Does it take into account that Fleichmann was paired with the defensive line for much of the third period while his regular linemates sat? Does it take into account the weak play of Clark and Nylander in that game and how that underminds Flash's own play?

I've worked in research-based publishing for more than a decade. From my standpoint, the Corsi Rating is useless. It attempts to judge individual players based on play that's reliant on line chemistry, which is a fatally large analytical hole. It also doesn't consider the roles of each line or player. It tells you nothing about a player's quality of play. It only tells you how many shots occurred while the player happened to be on the ice. It's weaker than the plus/minus stat.

JP said...

The rating doesn't say that Flash was "the worst offensive player on the ice" - it says he was the worst forward taking both ends of the ice into account in terms of the shots created and denied when on ice at even strength. That's all. Take it with as much NaCl as you need.

exwhaler said...

And that is exactly what I take issue with. The stat is inaccurate in regards to an individual's play.

pepper said...

That Bradley is so high up there is very interesting. Says a lot about grinders who can effectively cycle away in the offensive zone.