An analysis of Phil Kessel's advanced stats

(Ed note: got this great piece from Anomandaris84, I helped him format the charts to this site, the work and all credit for the rest goes to him!)

**

Following the Kessel trade I decided to take a look at some numbers to try and find out, whether looking at shot data I can find any support for his reputation of being lazy and/or simply terrible defensively. Before I get into it, let me add a little disclaimer:

Disclaimer: This post has gotten pretty long and there's more data I could've looked at. Also, there are a lot of numbers in it, so if that's not your thing, you can jump right to the conclusion. With that being said, I always tried to describe as good as I could the numbers that I used. So I hope that for readers that aren't very familiar with some of these numbers it's possible to follow along as well. I'm not a pro at using these stats, so let me know if I messed up somewhere:)

Corsi, WOWY and common linemates

First, I decided to take a look at the impact Phil Kessel had on his most frequent linemates over the last 3 season (12/13-14/15, only even strength). I chose that timeframe, because over that time, the 4 forwards he played the most minutes with, all played for Toronto for the entire 3 seasons (all WOWY data from stats.hockeyanalysis.com). These are:

Kessel spent around 75% of his line playing with James van Riemsdyk and Tyler Bozak. Because of that, he's only spent a small portion of his TOI with Nazem Kadri and Joffrey Lupul. As most of you probably know, Toronto has been a pretty bad team over this time-frame, so the Corsi numbers aren't good for each of these guys. However, I'm more interested in the effect Kessel had on his linemates and not in how good the Corsi numbers are. As a quick reminder, Corsi means shot attempts (shots on goal, missed shots, blocked shots) for or against a team.

Shots for

As you can see in this first graph, the only guy that has a better Corsi for % (= shot attempts for a team relative to all shot attempts) away from Kessel than with him is Kadri, and the difference is very small. So, Kessel seems to at least not drag his teammates down. However, the knock on Kessel, as far as I know, has never been that he doesn't generate shots, but that he gives up as many or more shots than he generates. So instead of looking at CF% I'm going to look at CF/60 and CA/60 which means shot attempts for and against a team per 60 minutes of 5v5 play.

Shots against

Now, as you'd expect, Kessel has a very positive effect on his linemates when we look at CF/60. That's not surprising, given that he's been the most dynamic offensive player on his team. More interesting to me are the CA/60 numbers, because these numbers might give us a hint on whether Kessel gives up as many or more shots as he generates. The numbers are a bit mixed, Kadri and Lupul give up slightly less shots away from Kessel, whereas JVR and Bozak give up more shots when not playing with Kessel.

What should we make of this? The numbers presented so far seem to suggest, that Kessel at least doesn't stand out as a horrible defensive player when compared to his most frequent linemates. However, maybe the reason for that isn't that he's about as good defensively as Bozak, JVR et al., but that whenever these guys weren't playing with Kessel, they had to play harder minutes. After all it makes some sense that Kessel, being the most gifted offensive player on his team, would get more offensive zone starts vs the weakest possible competition to maximize his offensive output.

Zone starts

Whereas these other guys might have had to play tougher minutes, whenever they weren't playing with Kessel. To see whether that's true, we have to look at how these guys were being used in the 3 seasons  from 12/13 to 14/15. So let's do that (graph courtesy of www.war-on-ice.com):

I only included players with more than 1,000 minutes TOI of 5v5 play. On the X-axis you can see how often a player started his shift in the offensive zone compared to starts in the defensive zone. On the Y-axis you see the average time on ice/60 minutes of the opposition's players. This is a proxy for the skill level of the opponents, because, as one would expect, good players usually play more minutes than bad players. The color and size of the dots are Corsi % and time on ice, but they aren't very important with regard to usage.

As you can see, Phil Kessel did indeed start a lot of shifts in the offensive zone, however the difference to his 4 most frequent linemates Bozak, JVR, Kadri and Lupul isn't very big. Bozak and Lupul did have less favorable starts though when away from Kessel. When we look at quality of competition, Kessel, Bozak and JVR all have pretty similarly tough competition whereas Kadri and Lupul get weaker competition when away from Kessel.

What does it all mean? The fact that Bozak, JVR and Kessel are so close to another is mostly given to the fact, that they spent most of their time on a line together. It does seem that Bozak gets tougher assignments when he moves away form Kessel. Kessel spent way less time together with Lupul and Kadri, who both had slightly tougher starts but also weaker competition, so I wouldn't say that their assignments when away from Kessel were much tougher. To conclude, I don't think you can convincingly make the case that Kessel's most frequent linemates played much tougher minutes when away form him.

Examining the Kessel defense narrative

So far it doesn't seem to me like you can make the case that Kessel's much worse defensively than his most frequent linemates. However there's still one more thing we can look at. So far we've only looked at shot attempts, but that doesn't tell us anything about shot quality. Maybe Kessel is so terrible defensively that his team gives up way more high quality shots to the opposition with him on the ice than when he's off the ice. To look at this we can use the awesome hextally graphs from www.war-on-ice.com:

What you see on this graph are shot rates against Toronto at even strength with Kessel on the ice (left) and off the ice (right). The shots are sorted into 3 areas. A red dot means that Toronto gives up more shots from that area than league average. Numbers over 1 mean the same thing. Toronto gives up a ton of shots, whether Kessel is on or off the ice. However, the rate of high danger shots against is quite a bit higher when Kessel is on the ice compared to when he's not (1.52 vs 1.23). So this seems to support the notion, that Kessel tends to play a wide open game, creating chances but also giving up a lot of high danger shots against. However, keep in mind that when Kessel is on the ice, that usually means Bozak and JVR are right there as well, so it's tough to say, what Kessel's effect actually is. Lucky for us Kessel played his first 3 seasons for the Bruins and spent a lot of time over the three seasons playing with quality forwards like Marc Savard, Milan Lucic and David Krejci

And lo and behold, there's only a minimal difference in high danger shots against with Kessel on and off the ice (1.08 vs 1.06). So when put together with good linemates in Savard and Lucic, the number of high shots against goes down a lot compared to his time with Toronto.

Conclusion

The numbers do lend some support to the notion that Kessel is bad defensively and gives up a lot of high quality shots. However, to me the whole issue seems to be blown out of proportion and is probably more of a product of Toronto's awful shot rates over the last few years. Kessel's most frequent linemates of Bozak and JVR don't do better without him in any case. If we look at Kessel's time in Boston, the shots against hextally with him on and off the ice is a lot better. Kessel did just fine playing with Savard as his center, I'm 100% certain he'll be just fine playing with Sidney Crosby or Evgeni Malkin.