THE BOOK--Playing The Percentages In Baseball

Really nice stuff ... but the charts are based on only 2000 shots.  Is that enough?  I get the SE of goals allowed being about 6.5 goals a season, which means that the chart of goalie differences doesn’t show significance (although I still believe that Luongo is good).

I’ve gone through this analysis several times, but ultimately I’m not convinced it tells us anything.  Once you get 3-4 years worth of shots, the difference between goaltenders is basically nothing:

http://www.behindthenethockey.com/2009/10/10/1079378/even-strength-goaltender-shot

So the toughest shooting was 6.5% and the easiest was 5.5% if we include goalies with ~1 year experience.  Over four seasons, it’s quite a bit narrower than that.

The numbers I’ve shown there are rink-adjusted for shooting percentage (most important for MSG) but I’ve done nothing to account for under- or over-counting of shots.  It would be better to use road stats, but then you’ve cut the sample size in half, so you may not come out ahead.

Vic and Tyler saw through all the b.s. I was doing with goaltender shot quality and instead turned to raw EVSV%.  Adding shot quality to the mix does not improve its predictive value.

Also it’s a terrible chart--essentially illegible.