by PuckStopsHere on 06/27/11 at 06:53 PM ET
In my first sabermetrics and hockey post of the summer, I listed the top 20 adjusted +/- ratings for 2010/11. When looking into some of the top players we found that they had benefited from a high saves percentage on their team and a low saves percentage from their opponent’s team. These saves percentages, though partially driven by the player in question, are very unrepeatable from year to year. A much more repeatable number is the number of shots directed at goal or Corsi rating.
In order to see how important saves percentages are to the players who got top adjusted +/- ratings, I thought I would list the top 20 adjusted +/- ratings along with the 5 on 5 saves percentages that existed for and against when a given player is on the ice. His +/- rating benefits from a high saves percentage for his team and a low saves percentage for his opposition.
Sv Pct For
Sv Pct Against
Every player on this list had a better saves percentage for his team when he was on the ice than their opponents did. In most cases it was by quite a significant margin. None had as big a margin as Adam McQuaid who’s Bruins had a .945 saves percentage when he was on the ice, but their opponents went at a .883 clip. I think the Bruins goaltending and some good luck was largely responsible for the high saves percentage for his team and I find it hard to imagine that a three goal 15 point defenceman can impact an opponent’s saves percentage to make it as low as .883. McQuaid’s position here is likely only due to the respective saves percentages.
For every player who has a better saves percentage for than against there is a player who has a better against saves percentage than a for one. None of these players appear on the top 20 list. There are also a considerable number of players with roughly the same saves percentage in both directions. None of them appear here either. In order to appear on a top adjusted +/- list in 2010/11 it was necessary for your team to have a better saves percentage than your opponents when you are on the ice at even strength. There are no exceptions. The closest is the New York Islanders pairing of Frans Nielsen and Michael Grabner who both had a 5 point better for saves percentage than against. McQuaid was 62 points better. The adjusted +/- leader Toni Lydman was 44 points better. The average player on this list was 29 points better. Given that the average player in the league has the same saves percentage for and against, this list is strongly selected for players who had a saves percentage advantage and this saves percentage advantage is not very repeatable. The fact that we see defensive defencemen in Adam McQuaid and Toni Lydman putting up the worst opponent’s saves percentages, it is hard to see that as a skill that they were responsible for and not a fluke occurence.
+/- is useful to find the players who succeeded in their roles. It is even more useful after adjustment so that players on different teams can be more easily compared. It suffers from the problem that it is driven by an unpredictable factor in saves percentages. Largely they are not repeatable. Repeatable elements may exist but they are strongly drowned out by random noise. This is why Corsi is preferable to +/-. It is not subject to the same random variations and is thus more repeatable. It is possible to correct after the fact for repeatable skill a player has on saves percentages, but these repeatable portions are difficult to isolate. Nevertheless this is something hockey sabermetricians have tried to isolate. This is a topic I hope to address in the future.
Be the first to comment.
Add a Comment
Please limit embedded image or media size to 575 pixels wide.
Most Recent Blog Posts
About The Puck Stops Here
Who am I? A diehard hockey fan.
Why am I blogging? I want to.
Why are you reading it? ???