I made an estimation of the effects of the differences in shot counting across the league. First I took every game of the BTN era and looked at the (corsi) shooting percentages. I then calculated the save percentage of the goalies in all other rinks but the one where the game took place. Then I calculated the shooting% of the shooting team in every other rink.
We now have a couple of things:
– An estimation of the goalie’s shot-to-goal ratio across 29 arenas.
– An estimation of the shooting team’s shot-to-goal ratio across 29 arenas.
And of course we have the actual shot-to-goal ratio of the game. We then subtract the actual ratio from our ‘projected ratio’, the average of the above mentioned ratios. The difference between the ‘expected’ shot-to-goal ratio and the real one we attribute to the scorekeeper(s) at that particular rink, as he or she is the variable we’ve isolated in this experiment. By looking at thousands of games like this, we can make a rough estimate of the kind of effect the scorekeeper is having on shooting/save% on games at that arena.
Why shot to goal ratio? Because goals are something that the scorekeeper has no influence over, and therefore give us a pretty simple and easy barometer of the ratio of shots to goals. Simply put, if there’s a strange difference in the shot to goal ratios at a particular arena from average, we can bet it’s something about the arena that is making it do that.
Here are the numbers
All numbers include even strength shots only, so to eliminate the effect different game states may have on the totals. The home effect is the % in shooting percentage that this experiment attributes to the scorekeeper, and vice versa for away. A postive number would mean that shooting percentages are higher then they should be in the rink, meaning that there is likely shot count inflation going on. The rink bias shows the difference in this effect between the home and away teams.
The Pittsburgh Penguins are the worst practitioners of rink bias when it comes to shot counting according to this research. There are 0.58 too many shots being counted for the Penguins per 100 while at home, while their opponents are losing 0.43 shots out of a hundred relative to the other 29 arenas. Rogers Arena in Vancouver also seems to be counting shots rather partisan-y. The Maple Leafs, conversely, are getting too many shots counted against them while at home, one of the few teams to give a shot counting bonus to the opposition.
Using the above index for reference, I recalculated every team’s shot totals in every rink this season adjusting for shot totals.
The effect? Basically meaningless. The biggest movers are the Ottawa Senators, who are adjusted at the top, who gain an eighth of a percentage points, and the Pittsburgh Penguins, who lose a ninth of a percentage point.
Rink effects exist. Some arenas have shown a consistent ability to suppress or inflate shots, but as we can see the effect is 10s of shots a season. Low 10’s. It isn’t enough to make a real effect on corsi totals, but it’s still something to think about. Beyond just checking the scorekeepers work with every shot, this is probably the best analysis we can do on rink effects on corsi. The numbers above are estimates of how we think corsi totals might be effected.