My personal feeling (always open to being out-voted) is that the variation in goalie scoring is not so wild as to create a structural problem. Experientially, I feel like the number of games can *generally* be anticipated at the start of the week, but I have no real evidence for that and I'm sure we've all been surprised on occasion. To explore Matt's larger point a bit about the variation in weekly scoring, we can actually go to the ESPN system itself (for an 'all you can anecdote' festival of small samples). Pick your favorite goalie and your favorite skater, and use their 'Charts' function to compare weekly scoring. You can see Rinne vs Ovechkin below.
Goalies do have higher peaks and lower valleys (crossing the 0 plane in most cases), but its partly their higher averages (4 vs 2 games at 2.4pts creates more visual up-and-down than 4 vs 2 games at 1.9pts). If Ben has the data from his last analysis, we could run some simple descriptives to get a better account. Still, I'm not convinced the flux is totally out of proportion with skaters. I'd be interested to see more examples as well as more rigorous evaluation.
(PS: Nice meme, Matt)
No comments:
Post a Comment