How often do you outshoot your opponent by 23 shots...and not win?
by Hawerchuk on Feb 18, 2010 10:13 PM EST
in Zzz...2010 Olympics
Email
PrintCanada narrowly escaped with a win against Switzerland, taking everything down to the shootout tossup. But they outshot Switzerland 46-23, which, in a close game, normally indicates puck possession superiority. But does it lead to a high winning percentage? Going back to the 1995-96 season in the NHL, the answer is...no:
| Outshot | Win% |
| 25 | 0.628 |
| 24 | 0.575 |
| 23 | 0.579 |
| 22 | 0.587 |
| 21 | 0.642 |
Even with a massive lead in shots, a team should barely expect to win 60% of the time, which is probably just slightly higher than Canada's odds of winning the shootout.
12 comments
|
0 recs |
Do you like this story?
Comments
You will find that as shots go up, shooting percentage goes down (save percentage goes up). So, those 23 shots difference are more like 15 or 13 shots difference or something.
Likely playing-to-score effects? I have read somewhere that the SH% goes down for trailing teams even though their shot rates normally go up.
Playing to score effects are definitely significant. If a team is trailing, its shot rate goes way up and its shooting % goes way down.
What’s different here (and which isn’t captured by the analysis) is that more than 50% of this game was played tied. So we shouldn’t see the effects of playing to the score.
Oh definitely, if you’re talking specifically about Canada vs. Swiss it was domination by chances (despite the Canadian players not paying much attention to detail).
I think the scoring chances project should be good to explore some of this shot quality stuff, generally speaking I think you could define shot quality as a chance-to-shot ratio. I know Vic has said that Corsi and chances at EV track closely while the score is close, probably a different story though when it’s not close.
This seems odd as it wouldn’t jive with the lower SH% when trailing. Unless all the scoring chances end up being of the “low-quality chance” variety in which case I think we’re back to square one with the chances project. Damn.
Well...
it was pretty obvious to me in the third that the Swiss were playing for the tie. Basically no forecheck, sitting back in the trap and not wanting to trade chances. This wasn’t a normal style of play for a tied score.
Many years ago, I tracked shots faced and save percentage, by goalie. There was a strong correlation. Now, I did not include score differential to make sure it was a tie-game. But, I would be surprised if that was the major reason.
It’s possible that a goalie is better per shot, the more shots he faces.
Correcting for PP bias
I think part of the reason for the lower winning percentage is that the difference is often due to a difference in penalty minutes. Perhaps taking out that effect would give you more of the difference in winning percentage one would expect.
Correction...
That was definitely not clear. What I meant to say was that a difference in shots might be more dependent on the difference in penalty minutes served (negatively influenced) than who was the better team. What you could do is something simple like run a logistic regression with difference in shots and penalty minutes on the game outcome (win/lose). Another thought would be to weight shots depending if they came during a power play/shorthanded vs even strength. Not sure what the weights would be or how exactly to estimate them, but with proper weights, that would definitely work.
by James Piette on Feb 19, 2010 2:40 PM EST up reply actions
Comments For This Post Are Closed
Secondary Sidebar
User Tools
FanPosts
Community blog posts and discussion.
