Ries: The Value of a Hit

In my previous two articles in this series, I looked at the approximated value of a takeaway and a giveaway. Both events when totaled up over the course of a season carried relatively low weight on the outcome of a team’s success. Using a similar approach to what Gabe Desjardins did when looking at the value of offensive zone faceoffs, we found that in order to gain an additional 2 points in the standings, a team must add 486 Takeaways or commit 351 less Giveaways. Today we will use the same approach to estimate the value of a Hit.

There are times when Hits are necessary for dispossessing the puck from the other team, and sometimes they can be meaningless. We do know that outhitting your opponent may be a bad indicator, since it suggests your team had the puck less. This is also shown when looking at correlations between Hits and Points. Over the past two full seasons, there is a -0.17 correlation between the two stats. I suspect we will find that Hits similarly are low value plays, and may even be less valuable than Takeaways and Giveaways

Here’s a quick refresher on the method I used:

I looked at play by play logs from the past two seasons of regular season games and tracked how soon after a Hit a shot for and shot against occurred. If there was a stoppage of play, I would “reset the clock” to zero and wait for another Hit. Below are the per game SF and SA values for two seasons of data:

We see some interesting results right away. Since a team sees more shots against than shots for after a hit, at a basic level we would infer that to mean you are costing your team each time you hit another player. Let’s calculate that value:

There were 119,321 Hits over the last two full seasons (48.5 per game). Summing the shot differential value (SF-SA) for each second on the Hit graph above and dividing by 48.5 we get -0.38 shots per additional Hit. Multiply by 0.089 (League-wide shooting %) = -0.003 goals per Hit (or -0.3 Goals per 100 Hits). Three goals per point and two points per win means we take 6 / (-0.003) = -1792.2 Hits per additional win.

The New York Islanders racked up the most hits last year, with 2685 (thanks largely to hitting machines Matt Martin and Cal Clutterbuck who had 725 hits between the two of them). So even if we suggest that our conclusion is true, we can play out the silly scenario that the Islanders would have added 1.5 wins by going the entire season without hitting anyone, which we know is not feasible in the slightest.

I think our results are slightly surprising, but it likely is the fault of the way we set it up. Consider the scenario of a player hitting an opponent in the defensive zone. Either the puck is turned over, which has the now offensive team breaking out from their own zone, or the puck stays with the same team who are still in a ripe position to get a shot off since they are in the offensive zone. It’s not necessarily the fault of the hitter that the puck was in his defensive zone to begin with.

This would be a logical explanation for our results above only if more hits occurred in the defensive zone than offensive zone. If hits happen more in the defensive zone than the offensive zone, then of course a team will see more shots against than shots for after a hit. Fortunately, play-by-play logs keep track of this, and so we can explore that aspect more in future articles.