I’ve expanded the on-ice events report with three new stats to build on the existing on-ice shots for distance: on-ice shots against distance, relative on-ice shots for distance, and relative on-ice shots against distance. Note that these distance stats are actually ‘fenwick distance': the average distance of goals, shots saved, and shots that miss the net.
I was curious to see on-ice shots against distance for defencemen in particular — I wondered if any defencemen were able to consistently force opponents to take shots from further out than their team could otherwise. Using the distance relative to their team’s when they’re not on the ice helps mitigate scorers’ bias in shot distance, which can be fairly substantial.
The result? On first look, not much. The list of defencemen with the highest relative on-ice shots against distance doesn’t really match what you’d expect. In 2013-14, Benoit, Hamonic, and N. Schultz had the biggest positive effect on opponents’ shot distances (i.e. forced shots from further out than their teams without them) and Sekera, Guenin, and Jackman had the biggest negative effect. Chara, a guy who you might expect to force shots from far out with his size, is in the middle of the pack at -0.3 ft. relative to the rest of the Bruins.
In addition to eyeballing the list, I wanted to check, statistically, if relative shot distance against was a skill for defencemen. If something’s a skill, it should persist year-to-year. Taking 298 defencemen season pairs from 2011-12 through 2013-14 (e.g. Niklas Kronwall in 2011-12 and 2012-13) I looked at how their relative shot distance on year 1 compared to the same in year 2:
The correlation coefficient (R-squared) is 0.03 which is fairly low: on-ice shot against distance regresses 82% toward the mean year-to-year. To put that into perspective, Eric T found that goal scoring, something that certainly has a talent component, has a correlation of 0.22.
Another way of looking at this is to group players based on their year 1 shot distance and see how the groups fare in year 2:
|Y1 SADist rel Group||N||Avg Y1 SADist rel||Avg Y2 SADist rel|
|1.5 to 2||23||1.70||0.23|
|1 to 1.5||31||1.18||0.56|
|0.5 to 1||35||0.72||0.30|
|0 to 0.5||49||0.16||0.08|
|-0.5 to 0||50||-0.27||-0.03|
|-1 to -0.5||41||-0.79||0.09|
|-1.5 to -1||24||-1.28||-0.29|
|-2 to -1.5||12||-1.78||-0.32|
If on-ice shot distance were completely random we wouldn’t see any trend reading the last column top to bottom. This isn’t the case; there is a trend, but it’s noisy.
When we look at these correlations, we’re capturing more than just a defenceman’s talent, of course: his teammates, zone starts, opponents, etc. are all mixed in. Taking everything into consideration I’d say that defencemen have little identifiable talent for forcing their opponents’ shots to come from further out or closer in. But it would be interesting to look into things deeper, perhaps with more seasons’ worth of data and only focussing on ‘open play’ shots.
Something else I wanted to check was the relationship between shots against distance and save percentage. On-ice save percentage for defencemen is largely luck-driven by about as much as on-ice shot distance is. And distance and save percentage are closely related in general. So, maybe year-to-year changes in on-ice save percentage are driven by changes in shot distance?
This doesn’t appear to be the case. About 2% of a player’s on-ice relative save percentage can be explained by his on-ice shots against distance. Trying to understand this intuitively, we know that the league average shot distance is 36 ft., and the majority of defencemen are in the 35–37 range for shots against distance. The difference in save percentage between a 35 ft. shot and a 37 ft. shot just isn’t that big compared to other factors affecting save percentage.
Summing up, it appears that a) defencemen have little talent for forcing shots from further out, but further analysis could be done and b) there’s little relationship between relative on-ice save percentage and shot distance for defencemen. I haven’t looked at on-ice shots for distance for defencemen or any shot distances for forwards, but the data is now available to do so.