Saturday, June 11, 2011

Where is the Cause and Effect?

I was poking around with numbers at to see whether skilled players may affect shooting percentage.

I pulled information from all four seasons' worth of data, and using the filters of minimum 40 games played (for durability), forward (since I'm looking at offense), and at least 12 minutes of 5on5 time on ice per game (top-nine forward, roughly), I got this interesting graph:

The regression line indicates that for every extra minute of TOI/60 these players get, their on-ice shooting percentage increases by half a percent. I'm not sure how I'd go about doing a significance test or confidence intervals here, given different amounts of time on ice for the players, but 814 isn't exactly a tiny sample, I don't think.

To put that in perspective, Alex Ovechkin was on ice for 66 goals for at 5 on 5, the Caps shooting 9.09%. With the Caps shooting an extra half percent, he'd have been on-ice for 70 goals. Account for the extra minute of time on ice and that goes up to 74 goals. Those 8 goals are worth about 1.5 wins, which is not insignificant, though not a giant difference, either.

So what is cause, and what is effect? Are the coaches giving the guys with high on-ice shooting percentage more ice time, or are the coaches simply giving their best players lots of ice time, during which they shoot a high percentage?

Unfortunately, I'm having to manually align the rows for year-over-year correlations, so I'll have those at a later date.

No comments:

Post a Comment