How much was that hit worth?

As part of the labor dispute, the NHL removed some of the advanced stats it was recording from its game feed several years ago. That includes hits, giveaways, takeaways and blocked shots – NHL owners didn’t want low-scoring (and otherwise unappreciated) players getting bigger contracts because they led the league in these categories, so they stopped releasing the information to the public. Enough history – the 2005 CBA solved these problems and the NHL brought back its statistics for the 2005-06 season, which allows us to figure out how much each of these events are worth.

First, let’s look at hits. This chart shows the cumulative number of goals scored in the time after a hit. Goals For are shown in green, Goals Against in blue, while the average long-term scoring rate is shown in red.

What this shows is that a hit has no effect on the rate at which the opposing team scores (the green line follows the red line) while it reduces the rate at which the hitting team scores. There are two ways to interpret this: the first is that a hit has no inherent value – it doesn’t necessarily result in a turnover, so it has no real effect on the opposing team’s ability to score; at the same time, it takes the hitting player out of the game for some amount of time, which reduces his team’s ability to score. On the other hand, the NHL does not record which zone the hit took place in, so the average long-term scoring rate is a bit misleading – if a team is more likely to hit in its defensive zone (when it doesn’t have the puck), then their opponent’s expected scoring rate will be higher than their own, and the resulting GF and GA rates are in line with that expectation.

What about blocked shots?

Again, we have a similar situation to hits. Blocking a shot appears to be a negative event that increases the likelihood of the opposing team scoring. But this is misleading because shots can only be blocked in the defensive zone, so the opposing team was already more likely to score. The shot blocker is temporarily out of the play, which reduces his team’s ability to score.

Missing the net?

Here, we have a slightly improbable result: taking a shot and missing the net increases the likelihood that a team will score in the future. But it’s not really that strange – you can only take a shot if you have control of the puck in the offensive zone, which has a very high expectation of scoring relative to the neutral offensive event, taking a face-off at center ice. Thus shooting and missing the net is very likely to be a precursor to shooting and scoring.

Giveaways and Takeaways are nearly symmetric events, though giveaways involve passing the puck directly to the other team, while takeaways usually happen through bodychecks or work in the corners. Here are the cumulative goal-scoring rates:

Giving the puck away (purple) is slightly more likely to result in a goal against than taking the puck away (blue) is to result in a goal for. This implies that giveaways result in much more wide-open scoring opportunities. In each case, the team losing the puck scores at roughly the same future rate.

Cumulative Score Change

The following plot shows the change in score over time as a result of each of the recorded events.

The effects of a giveaway, takeaway or missed shot last approximately 10-15 seconds before the score change flattens out (which implies no further average score change, or the game’s equilibrium offensive state.) Hits and blocked shots appear to have a longer-term effect. In this context, we can see that the value of a blocked shot is actually quite positive:

This chart shows the cumulative score change as a result of a missed shot and a blocked shot. The long-term effect of missed shot is approximately -0.017 goals/missed shot, while a blocked shot costs -0.012 goals/blocked shot. Since missed shots and blocked shots both occur in approximately the same offensive area, blocking the shot is worth at least +0.005 goals. If we were to account for all other shots (ones on net, and ones that result in goals) we would see that blocking a shot is a very positive event relative to allowing a shot on goal.