The analytics are interesting too. From a pure statistics perspective, it actually makes sense to start pulling the goalie at about 8mins left in the third. But fans will have quite an adverse reaction to that (particularly if you do it at home). What’s interesting though is that we’ve gone from pulling with 1min left to 3mins… I suspect that will keep increasing until it converges to the statistical sweet spot, but only a little bit each year so fans don’t react badly.
The modeling is far beyond me, but when down one goal “The crossover point comes at 6:10 remaining. So, at 6:20 you should not pull the goalie, but at 6:10 you should.” And two goals they say pull with 13 minutes left.
Their explanation of their model helps somewhat, and I’m sure the math is all very interesting to someone who gets it.
Thanks for linking this! I thought this was interesting:
There have been a number of papers published on the subject using different models and data, but all
agree that goalies should be pulled earlier than is the usual practice.
The analytics are interesting too. From a pure statistics perspective, it actually makes sense to start pulling the goalie at about 8mins left in the third. But fans will have quite an adverse reaction to that (particularly if you do it at home). What’s interesting though is that we’ve gone from pulling with 1min left to 3mins… I suspect that will keep increasing until it converges to the statistical sweet spot, but only a little bit each year so fans don’t react badly.
Can you plz elaborate? :)
Also not a stats/capitalist/math guy, but a couple years ago I came across this paper “Pulling the Goalie: Hockey and Investment Implications.”
The modeling is far beyond me, but when down one goal “The crossover point comes at 6:10 remaining. So, at 6:20 you should not pull the goalie, but at 6:10 you should.” And two goals they say pull with 13 minutes left.
Their explanation of their model helps somewhat, and I’m sure the math is all very interesting to someone who gets it.
Thanks for linking this! I thought this was interesting:
Not a stat guy, but I do think about the 5 minute mark makes sense.
Agreed, teams will develop defensive strategies over time to mitigate the advantages, or if it becomes too strong the rules will be adjusted.