Under Randy Carlyle, the Maple Leafs were a test case as to whether a team can win despite being dramatically out shot largely through controlling shot quality, both for and against.
There is ample evidence to suggest that under Carlyle the Leafs were in fact an above average shooting percentage team and may even in fact have been able to suppress shot quality against to some degree. Unfortunately, their ability to be significantly out shot outweighed any ability to have a positive impact on shot quality, and they were at best a mediocre team.
Randy Carlyle is gone now and it seems that new head coach Peter Horachek is, at least partly, influenced by the Leafs new analytics team, which has a goal of significantly improving their puck possession game. For the most part, the Leafs have accomplished this and have dramatically decreased their shots against totals and mostly maintained their shot for totals. The change has been quite dramatic even if the actual results in the win-loss column haven’t been there.
I am skeptical that a team can dramatically change its style of play to boost Corsi without taking a hit in the shot quality department, but that is what makes the Horachek version of the Maple Leafs a new test case for Analytics. The Leafs performance under Horachek will be dissected in every way possible in the upcoming weeks and months. Today, I am going to start that by looking at player zone start usage and how things have changed from Carlyle to Horachek. In order to minimize score effects (Leafs have not held the lead in games very often under Horachek), I am going to look at 5v5 close situations only (close is defined as within one goal in first and second periods and tied in the third).
The following is a usage chart under Carlyle showing what zones each player started in when he was on the ice for a face off (on the fly shift starts are not considered).
Players are sorted by OZFO%, so Jake Gardiner had the highest percentage of his zone starts in the offensive zone (30.5%) while Percy had the lowest (19.5%). All the other defensemen ranged between from 23.3% to 27.6%; overall, not a huge zone start discrepancy between them. The same is mostly true for defensive zone starts. Generally speaking, Carlyle didn’t employ a heavy zone start bias among his defensemen. How about Horachek?
While sample size might be a factor, there appears to be a larger spread in usage under Horachek. Robidas leads the team under Horachek with 41.3% offensive zone starts, while Polak is down at 26.1%. The others range from 36.5% to 27.3%. It appears Horachek is using his defensemen in more targeted situations.
To investigate which defensemen have seen their usage change the most, I am going to look at what I call “Percent of Team” statistics, which is the percentage of all events by the team that the player was on the ice for (only in games that the player participated in). With better possession play there has been a higher rate of offensive zone starts, so using percent of team statistics will account for this. So, here is how the Leafs defensemen zone start usage has changed from Carlyle to Horachek.
I have included time on ice as well so we can get an idea of who has had their ice time changed (which would in turn impact their zone starts) and who has seen their ice time increased/decreased. The big winners in terms of ice time are Rielly and Franson while Holzer and Robidas are the big losers in ice time.
Franson and Phaneuf are being depended on much more significantly by Horachek in defensive zone starts at the expense of Rielly (who was higher but not as much as his increase in ice time) and Polak. Robidas saw his offensive zone rate jump significantly. The biggest changes though are with who is on the ice for neutral zone face offs. Rielly, Gardiner and Polak have seen big jumps in neutral zone face offs, while Franson and Holzer have dropped significantly.
While some of this makes sense to me, some of it doesn’t. We have seen some evidence that Phaneuf is good at preventing opposition movement up the ice, in which case it would make sense to start him more in the offensive zone and less in the defensive zone (plus, he has a good shot which he can use in the offensive zone). I like that Rielly and Gardiner are getting more neutral zone starts to take advantage of their puck moving skills, but Gardiner also saw an increase in his offensive zone starts (more than his increase in TOI), which limits how much he can utilize his good skating and puck moving ability. We saw in the game against the Oilers what Rielly can do with his good mobility; I think Gardiner can do that, too.
Here are the charts for zone start percentages under Carlyle and Horachek for Maple Leaf forwards.
Very similar to the defensemen in that, aside from Trevor Smith, there hasn’t been a huge difference in zone start usage under Carlyle. Certainly not anywhere close to what some other teams are doing.
There are some pretty significant changes in zone start usage under Horachek from Carlyle. I’ll list some of the most significant observations:
- There is significantly more variability in usage just like we saw among defensemen. This may be partly due to sample size, but clearly Horachek has a more defined sense of how he wants to use his players.
- Booth went from having the lowest offensive zone start percentage to the highest offensive zone start percentage.
- Komarov went from the second highest offensive zone start percentage to the lowest zone start percentage. Essentially, Booth and Komarov switched usage roles and then had them amplified.
- Holland is getting more defensive zone starts.
Now, how about their ice time and zone start percent of team stats:
Winnik and Panik have seen the most significant boost in ice time, and Winnik is getting put on the ice for far more face offs in all zones. The biggest losers in ice time and offensive zone starts are Komarov and Clarkson, Nonis’ $8.2M tandem. Holland, Santorelli and Smith have also been relied on more heavily under Horachek, while Bozak and JVR have been less relied upon.
Impact on CF%
It is a little too early to draw signifcant conclusions about whether Horachek’s changes will impact goals for and against (early indications are not great, although the sample size is still small), but we can look at how CF% has changed with reasonable reliability. Here are the 5v5 close CF% changes from Carlyle to Horachek.
Players are sorted by greatest improvement in CF%. Thus far, the biggest beneficiaries of the new system appear to be Gardiner, Polak and Rielly, while Phaneuf is the only defenseman to see a (slight) drop off in CF%.
Look at that, a stat that shines a positive light on Bozak. In general, despite the near zero goal production, the Leafs whole first line has seen a dramatic improvement in their CF%. That said, there was a lot of room for improvement with those three as they were near league-worst bad. Those who put a lot of stock in Corsi better start jumping on the Bozak bandwagon as only Winnik has had a (marginally) better 5v5close CF% under Horachek.
Conversely, the only two forwards to see a hit to their CF% under Horachek are Smith and Komarov.
Clearly things are different under Horachek. Players roles have been adjusted and more well defined and the impacts on Corsi have been positive, and in some cases dramatic. Are the Leafs a better team under Horachek? Some signs say yes, but early results scream a resounding no. Time will tell, but there is a clear difference between Horachek and Carlyle and going forward these changes can be used as yet another test case for analytics. It will be interesting to see how it plays out.