Maple Leafs Allowed Shot Locations

 

While doing research for my article on the Leafs’ defense in this year’s Maple Leafs Annual (Go buy one!), I compiled a series of graphs that displayed the locations of shots taken on the Leafs net, sorted by which players were on the ice. Naturally, to help alleviate the effects of scoring biases, I’ve only used data from the Leafs’ road games while the team was at even strength. As we’ll see, differences in sample sizes account for many of the differences in patterns, but I believe that there are nevertheless important differences to observe, and that it makes for interesting viewing, one way or another.

This graph reinforces my belief that Aulie’s numbers are going to take a step backwards next season (yes, yes, I know, I may be influenced by a confirmation bias). His PDO tels us that he was the beneficiary of very good goaltending, and to make matters worse, he apears to have allowed a substantial portion of shots against from very dangerous areas. I’d be skating on very thin ice to attempt to assign any kind of particular narrative to this, because there’s no telling if James Reimer’s developing blocking-style of play allowed for more rebounds than normal, or if Aulie should have been better at clearing them away, or even if he was simply letting opponents’ attackers get too close to the net on a regular basis. It could well be a combination of all three.

François Beauchemin’s graph shows only his time with the Leafs, and hence, has a similar sample size to Aulie’s. To my eye, his shots appear slightly more scattered, with  more shots coming from outside the slot and from the points. This suggests that Beauchemin forced his opponents to take more bad-angle shots off the rush, rather than getting hemmed in his own end and allowing shots from the point. Moreover, Beauchemin is left-handed, and would have spent most of his time on the right side of the ice. The concentration of shots on the left hand side of the crease suggest – though are by no means conclusive evidence – that opposing forwards exploited his partners as weaker links.

 

Cody Franson’s information was taken from his time with Nashville, and although he saw substantially less ice time that some of his new teammates, the fact that he dressed for 80 games compensates for that, and makes for an easier comparison with his alphabetical neighbor, Carl Gunnarsson. Similar to Beauchemin’s graph, we can see more shots coming from Franson’s partners’ side than his own.

 

Gunnarsson’s graph appears less scattered than Franson’s, and lines of shots forming 45-degree angles towards the net are beginning to take shape, suggesting that Gunnarsson allowed more shots off the rush.

Unlike other graphs, Tomas Kaberle’s shot locations form very symmetrical patterns, which suggests he was often paired with a defender of similar abilities. Although it appears that a large percentage of his shots against came from a very close range to the net, there are no clusters around the points, which could potentially mean that Kaberle lets players walk in much closer, but that he has less trouble clearing the puck to alleviate pressure.

 

Somewhat surprisingly, Mike Komisarek’s data shows more shots taken from the right side of the crease, which would have been the opposite side he played. Given what we know to be true about Komisarek already, this is perhaps enough to discredit my theory that more shots will appear on the side of a weaker defender. Komisarek also seems to have kept players a bit further back from the crease than Kaberle.

Too small a sample size to say anything, really.

Even with such a tiny sample size, patterns such as a ring around the crease and clusters at the points suggest Lebda was hemmed in his own end often enough. Lines towards the net are already starting to show from shots being taken off the rush. Go ahead guys, take your shots from wherever you want.

John-Michael Liles seems to have largely avoided clusters at the points as well as lines indicating shots off the rush. There are plenty of shots being taken from relatively close range around the net, though he seems to keep players away from the sides of the net.

Phaneuf’s data looks more asymmetrical than Liles’, and many of the shots come from his partners’ side. There is little in the way of point shot or rush-shot build-up.

Given the time that Schenn and Kaberle spent so much time together as a pairing, it’s no wonder that Schenn’s information looks very symmetrical as well. We can observe lines building up heading towards the net, and some small amounts of clustering around the points that may have been more obvious in Kaberle’s graph had he spent the whole season with the Leafs.

  • SkinnyFish

    Instead of shots by grid location (The ESPN way), what about taking shots locations that result in goals and over laying the Leafs shot location onto that heat map?

    You could then take the highest goal concentration areas and assigns them values, and find the average “Shot Value”, the likelihood of shots being taken from high goal areas, for each defenseman.

  • Danny Gray

    Nice work as always JP. Are you planning on tracking shots during the season? Also is there any way to sort the shots by the score. I’m wondering if teams that are trailing often take shots from low-chance areas.

  • How much trouble would it be for you to split the defensive zone into 9 sub-zones (3×3) and then superimpose the percentage of shots from each zone. Because of the different number of shots on each players shot location map it does get kind of difficult to compare players.

    • I suppose it wouldn’t be difficult, though it probably would be time consuming. We’ve actually already done something like that in the MLA, and it seems to me that ESPN also does something similar, but I can’t find it at the moment.

      The only issue with these kinds of graphs is that they never show the exact shot locations. They just group them into a larger area, and there’s no way to see, for example, that many of the shots against Dion Phaneuf come from his partner’s side of the net.

      • Yeah, I just find it difficult to see how different, for example, Aulie’s and Beauchemin’s shot maps are. They look a little different, but it could be deceiving if the shot totals are different.

        • @ David Johnson

          Aulie allowed more shots from close in, and that’s bad. It’s not a startling discovery, but but still interesting to see plotted out, I think.

          @ Danny Gray

          Thanks, though I don’t have any way of sorting shots by score. As for next season, I’ll probably just wait until the end of the year to gather all the data. Just thinking out loud here, but a mid-season review might be OK.