Hello everyone and welcome to Arctic Ice Hockey's very own Friday #FancyStats Mailbag! Once a week I'll be answering a few of your questions on underlying metrics. Also, I will try to explore any concerns about the Jets that can be investigated through data.
Is there a stat comparing goals on original shot compared to goals scored on rebounds?
In our first mailbag we discussed how most of our data comes from NHL game reports. The NHL does track some data like player shot location -which tends to have some imprecision issues (note: imprecision and inaccuracy are different)-, shot type (wrist, slap, or snap), player who took the shot, who was on the ice when it happened, and what time it occurred, but currently there is no indication of whether a shot is a rebound or not.
There are two ways you can try to gain this information.
You can estimate it using ice time between shot occurrences. Rebounds by definition are when a player shoots immediately after a goaltender's save. The lack of time for a goaltender to get reposition and prepare is why rebounds are so dangerous. I used Greg Sinclair's estimates to account for rebounds when I made this article on Ondrej Pavelec's 2012-13 season.
The only other way is to track the information yourself.
Chris Boyle -author of SportsNet Shot Quality Project- has been tracking information like this for past seasons. He's been trying to establish what variables affect a goaltenders save percentage outside of talent and by how much. Here is an example of the data he tracked on Pavelec:
Our own Travis Hrubenik is tracking this stuff on Pavelec this season with his Jets Goaltender Performance series. Here is one on Michael Hutchinson he published today.
There of course is the issue of significance. It's been established previously that the sample required for significance in save percentage is fairly large (about 2000+ shots against) due to the natural variance in goaltending. Whether slicing the data into smaller pieces leads to more or less significance is still undetermined.
I recently wrote this article at Hockey-Graphs on the current state and future of goaltending statistics.
I was curious as to what the percentage is for how often a team scores when they pull their goalie?
The real answer is "it depends". Pulling a goalie has been shown to increase both scoring rates for and against. However when you are trailing with very little time remaining, the chance of losing a game by two goals instead of one is not much of a disincentive. The numbers are highly dependent on how long the goaltender has been pulled. As time increases, the chance of tying the game increases quickly, and then after a while it begins to decrease.
Arctic Ice Hockey founder, Gabriel Desjardins, wrote about this strategy on a website that was then called Puck Prospectus (now Hockey Prospectus). When the website changed, most of the formatting has been altered and the images are now gone, but you can still read the original article.
War-on-Ice has even created a tool to estimate when is the best time for a team to pull their goaltender. You can play with different rates and see how it affects the probability that the trailing team ties the game.
In a general sense, Desjardins found that most teams try to pull the goalie somewhere between 40-90 seconds, and that the maneuver on average increases the trailing team's probability in tying the game by about 300%.
Thank you everyone who wrote!
That is all for this week. If I didn't get to your question this week, I promise I will eventually reply. The plan is to answer every single question given to me, although sometimes they may slide a week or two depending on the load of questions I receive.
If you have any questions you want looked at, email me at garrethohlaih (at) gmail (dot) com.