HatesSetoguchi
Registered User
I'm new here, so I apologize if this has already been analyzed, but I looked at some numbers a while back on whether a team should or shouldn't pull the goalie when they are down 2 goals. The question is this: Does that small chance to get at least 1 point outweigh the goals you expect to give up?
Goal differential is the third tiebreak at the end of the season. If points are tied, ROW is considered (and ROW should be close if points are tied). After ROW, points earned in games between the clubs is considered. If this again is tied, then goal differential is considered. This is rather far down the chain of tiebreaks, but is preserving your goal differential worth the minimal chance at gaining a point?
Of course, there are times when you absolutely need to try and get points (late in the regular season or playoffs where goal differential doesn't matter), but I'm going to try and look at generic games early in the season and between two average teams ('average' to mean overall statistics from 2011-2012 regular season).
Using 2011-2012 stats, teams scored every 32 minutes with 5v5 play, every 10.5 minutes with 6v5, and every 3.5 minutes scoring on empty nets. That means that you are around 3 times more likely to score when you pull the goalie instead of leaving him, but are still scored on 3 times more often than you score when you pull the goalie. These numbers are similar to the ones found in the paper given by Mathletic in 2009, page 4 (7.83 times as likely to score on empty net than 5v5, and 2.67 times as likely to score on 6v5 rather than 5v5).
During the 2 minutes the goalie is pulled when down 2 goals, it is estimated that the team down scores 2/10.5 = 19% of the time. I went on to estimate (first part being chance of scoring 2, second part being that other team not scoring) [(2/10.5)^2]*[1-(2/3.5)] = 1.6% of the time tying the game, and thus earning 1.5 points. I did this style of computing for the different cases (no goal, 1 for losing/winning team, 1 each, 2 for losing/winning team), and compared overall goal differential to points earned, and came up with this statistic:
Pulling the goalie, over a long period of time is expected to earn 1 point for every -15.4 goal differential.
Or, if you add 1 point to any team, could they still win the tiebreak after suffering a 15 point goal differential swing (for instance, Minnesota potentially could have tied Detroit in points at 56 this last year, but would've had -20 DIFF compared to Detroit's +9). A better case would be 2000-2001, where Boston (9th) and Carolina (8th) had 88 points. Although there wasn't ROW back then, if Boston had given up 10 fewer goals, they would have surpassed the Canes for a playoff spot, which these goals may have come from empty nets. Now I doubt that is the case, but something to consider.
To be honest, I don't think there are many cases where pulling/not pulling the goalie when down 2 goals would get a team into/out of the playoffs, so this may be a moot point, but it's something I thought was interesting.
I know my analysis is not perfect, but any thoughts?
Goal differential is the third tiebreak at the end of the season. If points are tied, ROW is considered (and ROW should be close if points are tied). After ROW, points earned in games between the clubs is considered. If this again is tied, then goal differential is considered. This is rather far down the chain of tiebreaks, but is preserving your goal differential worth the minimal chance at gaining a point?
Of course, there are times when you absolutely need to try and get points (late in the regular season or playoffs where goal differential doesn't matter), but I'm going to try and look at generic games early in the season and between two average teams ('average' to mean overall statistics from 2011-2012 regular season).
Using 2011-2012 stats, teams scored every 32 minutes with 5v5 play, every 10.5 minutes with 6v5, and every 3.5 minutes scoring on empty nets. That means that you are around 3 times more likely to score when you pull the goalie instead of leaving him, but are still scored on 3 times more often than you score when you pull the goalie. These numbers are similar to the ones found in the paper given by Mathletic in 2009, page 4 (7.83 times as likely to score on empty net than 5v5, and 2.67 times as likely to score on 6v5 rather than 5v5).
During the 2 minutes the goalie is pulled when down 2 goals, it is estimated that the team down scores 2/10.5 = 19% of the time. I went on to estimate (first part being chance of scoring 2, second part being that other team not scoring) [(2/10.5)^2]*[1-(2/3.5)] = 1.6% of the time tying the game, and thus earning 1.5 points. I did this style of computing for the different cases (no goal, 1 for losing/winning team, 1 each, 2 for losing/winning team), and compared overall goal differential to points earned, and came up with this statistic:
Pulling the goalie, over a long period of time is expected to earn 1 point for every -15.4 goal differential.
Or, if you add 1 point to any team, could they still win the tiebreak after suffering a 15 point goal differential swing (for instance, Minnesota potentially could have tied Detroit in points at 56 this last year, but would've had -20 DIFF compared to Detroit's +9). A better case would be 2000-2001, where Boston (9th) and Carolina (8th) had 88 points. Although there wasn't ROW back then, if Boston had given up 10 fewer goals, they would have surpassed the Canes for a playoff spot, which these goals may have come from empty nets. Now I doubt that is the case, but something to consider.
To be honest, I don't think there are many cases where pulling/not pulling the goalie when down 2 goals would get a team into/out of the playoffs, so this may be a moot point, but it's something I thought was interesting.
I know my analysis is not perfect, but any thoughts?