It's Good To Be Home

by Dirk Hoag on Feb 28, 2007 6:06 PM EST in Statistical Analysis

Thanks to emailer Chris for another article inspiration:

I was telling my friend that hockey is one of the few [the only?] sports where the rules state that the home team gets an advantage -- line changes and face-offs. Since the home team gets to put their stick down last for a faceoff, does the home ice advantage help, hurt, or make no difference to the "good face off" men?

Besides hockey, certainly baseball rules give the home team an advantage by batting last, but outside of that, you're right, most sports do try and stay neutral in terms of home & visitor play. To answer Chris's question, I used the Game Summary files from the 2005-06 season, as well as the first 730 games of this season (give or take a couple games) to determine overall faceoff win percentages for the home and visiting team, broken down by location on the ice:

**FO Win% 2005-06**
	Vis	Home
Offensive	49.29%	52.41%
Defensive	47.59%	50.71%
Neutral	49.95%	50.05%
Total	49.00%	51.00%

**FO Win%, Current Season**
	Vis	Home
Offensive	47.73%	52.08%
Defensive	47.92%	52.27%
Neutral	48.77%	51.23%
Total	48.19%	51.81%

Interestingly, we do see a slight dominance by the home team here, that persists from one season to the next. Given an average of roughly 60 faceoffs per game, this translates into a couple extra draws per game being won by the home team. So, Chris, indeed, it seems like the home team does enjoy a benefit in the faceoff circle.

Before we get too excited, however, it should be noted that faceoffs don't seem to be hugely important through most of the game. Javageek over at Hockey Numbers demonstrated that "quick goals" (occuring within 8 seconds of a faceoff) are pretty rare, and I'd tend to agree with that analysis, as at least for a short while after a faceoff, players on the defensive side should be in good position (they haven't had time to make mistakes yet). What I really need to do is revive some of my prior FO work to look at actual shooting percentage versus what Shot Quality would predict. For example, Javageek notes that shooting percentage in the 8-second window is about half the norm, but if most of those shots are 50- to 60-foot slappers, that's not really out of line. I'll see if I can dig into that issue and provide an update in the next couple days...

Tweet Comment 2 comments | 0 recs |

Do you like this story?

Comments

It could be that we undervalue face offs by just looking at goals score within a short time frame after a face off win.

Perhaps we could take our cue from some advanced baseball statistics that quantify events in terms of their contribution to the currency of the game which, in hockey, would be goals scored.

In the same way that a two out walk has a certain value in its probability of resulting in a run, a face off win would have a certain value in terms of its contribution to scoring a goal. A face of win in the offensive zone results in a puck possession in the offensive zone. If we can look at the number of offensive zone possessions it takes on average to score a goal, then we have a measurable value for the face off win and loss. The same would be true for face offs in the neutral zone.

I am not sure the data are out there to complete this analysis, but it would be interesting to see how many "goals" players contribute (or lose) with their face off skills.

by Tom Farmer on Mar 1, 2007 10:55 PM EST reply actions