• Tired Teams and Corsi%

    by  • December 13, 2013 • Hockey • 2 Comments

    I’ve written a couple of posts this year about the impact of scheduling, both in terms of the impact on the records of the teams that are involved in games despite having played the day before and goalies who play in back to back games. We’ve been able to identify some persistent effects: there’s a real impact on results that comes from having to play when you’re less rested than other teams. Road teams are persistently less rested than home teams. This happy coincidence saves teams money when they travel – fewer days on the road means fewer dollars in travel costs. It also increases the chance of the home side winning and the fans who paid to be at the rink going home happy. At least some of what we think of as home ice advantage is almost certainly a rest issue.

    This issue came up again after two Leaf games this week. On Wednesday, the Leafs played against Los Angeles. The Leafs are one of the worst possession teams in the NHL; the Kings amongst the best. There was a wrinkle though: the Maple Leafs had had two clear days between games, not having played since Sunday. The Kings played the night before in Montreal. Amazingly, the Leafs attempted more 5v5 shots than the Kings: 56-46. Score effects probably didn’t hurt – the Leafs were chasing the game for most of the night but still: it was an eyebrow raising result.

    On Thursday, the Maple Leafs went to St. Louis. This time they were the tired team. The Blues had played Tuesday and then had a night off before the game. The Blues ran away with the game early – they were up two goals within the first thirteen minutes and three by the 16:10 mark of the first period. We’d generally expect strong score effects in that situation. Still, the Blues blew the Leafs away in terms of 5v5 shot attempts: 46-28.

    I had someone ask me if there was a rest effect on Corsi%, so I went and took a look. I used the data for the 2007-12 seasons and coded each game based on how many clear days there were between the game in question and the most recent game played for each team. I’m reporting the results for the situations in which there were at least 80 games. As you can see, there’s a real sample size problem: the NHL has some consistent scheduling patterns.

    It’s pretty much exactly what you’d expect: the home team does best when the road team played the night before. The road team does best when the home team played the night before or when both teams have had a lot of rest.

    I suspect if I re-ran this using close numbers, we’d see that the edge is even bigger – the Corsi% edge when the home team’s rested doesn’t quite seem large enough to me, although the finding that tired goalies playing behind tired teams do worse than rested goalies playing behind tired teams probably factors into that as well.

    All of which is to say, maybe we shouldn’t have been that surprised that the Leafs did so well against the Kings. And may God have mercy on the Oilers, playing their second in two nights and third in four days, against a Canucks team that’s been off since playing Monday night.

    Email Tyler Dellow at tyler@mc79hockey.com


    2 Responses to Tired Teams and Corsi%

    1. Woodguy
      December 13, 2013 at

      Can you re-run it with close corsi please?

      Gross corsi is ok, but close makes the information so much less ambiguous.


      You show the home team has a consistent advantage, even when rest is equal or tipped by a game to favour the visiting team, which is interesting.

      I’d imagine a portion of the consistent advantage of the home team is due to last change.

      Last change more important than rest?

      Seems so.

      • Mr Debakey
        December 16, 2013 at

        The book “Scorecasting” identifies referees/umpires as the consistent across all sports home field advantage givers.
        Its subtle, but consistent – more cards to the visiting teams in soccer, a smaller strike zone for the home team in the Majors.

    Leave a Reply

    Your email address will not be published. Required fields are marked *