Film breakdown: Michigan's 2019-20 penalty kill
Breaking down the tape and the numbers behind the system
Welcome to Fresh Ice, a data-driven college hockey newsletter. Each week, I’ll be publishing fresh (no pun intended) content that focuses on educating and explaining all aspects of the game. Haven’t subscribed yet?
Monday, I announced my Michigan hockey penalty kill tracking project and laid out the plan for sharing my data. Wednesday, I broke down the penalty kill system from the 2018-19 season to get a better sense of what was going on. Today, we’re diving into the same breakdown of the 2019-20 season.
This is what I wanted to do from the beginning of this project. All of this started when I decided I wanted to break down the system Kris Mayotte brought from Providence to Michigan when he got hired last offseason, so I couldn’t be more excited to finally dive into it.
As I talked about on Wednesday, penalty kill data can be fairly noisy because of the small sample size — Michigan only spent a total of 228 minutes shorthanded as compared to over 2,000 total minutes played — but there are still ways to evaluate performance. For this project, I’m combining manual tracking with film breakdowns to present a comprehensive look.
The basic numbers
I’m working with a 27-game sample here and I only tracked data at 5v4 to keep the comparisons as simple as possible. In the chart below, you’ll see a simple breakdown of the basic numbers both from my tracked 27-game sample and from all 36 games. All of the data in this chart is based on 5v4 play. It does not include any shots, goals or power play opportunities at 5v3.
These basic, underlying numbers present an at-a-glance explanation of Michigan’s penalty kill. It gave up less than one shot per penalty minute, and the goaltending was good. This tracks with Michigan’s overall PK performance in the 2019-20 season — its 84.3% kill rate tied with Denver for 15th out of 60 teams. It also tracks with Michigan’s overall goaltending in 2019-20, which finished third in the nation with a .934 combined save percentage.
The tape (and more numbers!)
To go beyond the basic numbers, let’s take a look at the tape and break down the systems Michigan used on the forecheck and in the defensive zone.
On the forecheck
If you read Wednesday’s breakdown of the 2018-19 system, you’ll notice the format of this one is already different: offensive zone forecheck and neutral zone forecheck aren’t in separate sections. That’s because, according to my tracking, Michigan applied pressure in the offensive zone just 12.3% of the time. You can refer back to Wednesday’s post for a more detailed breakdown of what I considered an offensive zone forecheck while tracking.
The majority of the time when the Wolverines did pressure in the offensive zone, it came late in the kill — maybe 10 to 15 seconds left — and they’d look to keep the puck in the offensive zone for the rest of the time.
In the neutral zone, Michigan mostly used three different formations — a 1-2-1, a 3-1 and a 1-3. The 1-2-1 look was most common at 39.4% of the time, followed by the 3-1 (35.2%) and the 1-3 (20.3%).
What’s obvious at first glance is that none of these looks put a ton of pressure on the opposing power play. In the 1-2-1 and the 1-3 there’s one forechecker available to exert pressure and influence decision-making by the other team, while the other three skaters read and react to where the puck is going to try to deny the entry. The first two clips show Johnny Beecher and Jake Slaker pressuring hard from the 1-2-1 look, but with just one skater pressuring, Michigan’s ability to pressure hard is naturally limited as compared to something like the same-side press we broke down on Wednesday.
In the 3-1, there’s no skater in front of the formation to pressure. With three skaters across the blueline, the goal is to deny the entry, and the skater behind the blueline is ready to chase down the puck, whether it’s carried in or dumped in. Off the eye, the 3-1 forecheck led to a large proportion of dump-ins and failed entries, and that’s mostly backed up by the tracked data. The data isn’t broken down by forecheck type, but Michigan overall forced 4.60 more failed entries and 7.28 more dump-ins per 60 minutes in 2019-20 than it did in 2018-19.
Even in the 1-2-1 and 1-3 looks, the emphasis is on disrupting the entry as opposed to intense pressure. Pressure is an added bonus because it can frustrate the opponent into turning the puck over or making a bad decision, but both looks are well-positioned to stack the blueline and disrupt entries regardless of how much pressure the forechecker is able to exert.
The defensive zone
When set up in the defensive zone, Michigan used a box formation 31.4% of the time and a diamond formation 62% of the time. Within those setups, there was a pretty good amount of flexibility to allow for aggressiveness and pressure.
In September, Mayotte said the system was almost exactly what he ran at Providence, with a few additional wrinkles and changes. In 2018-19, his last season at Providence, the Friars finished 6th in the nation with an 86.6% penalty kill percentage.
“For the most part, any system is the same. But if you talk to 10 different coaches, what they emphasize when they teach the specific system is the subtle difference,” Mayotte said last September. “Some talk about pressure first and foremost. Some talk about stick positioning first and foremost. Some talk about reading. There’s a couple different systems. There’s high-out pressure, there’s a little more passive where you’re gonna block a few more shots and just sit in shot lanes and try to be inside. And then what you see in the NHL is what they call the trap down or triangle and 1.
“I’m more of a straight line, go press, force them to really beat you and make a play. … Making reads, when we have a chance to go press, be aggressive. One go, all go type thing so that pressure is not for naught. I think this team’s pretty quick and pretty smart, so it’s taking advantage of that and hopefully forcing the power play to play a little faster than they’re comfortable playing and force mistakes that way.”
You can see that philosophy at work in the video.
While there are certainly clips that show a more passive style, when the moment is right for aggression and pressure, Michigan’s penalty killers are free to take advantage of it. In the first clip, Luke Morgan in the top corner goes high in the zone to pressure the Ohio State skater at the blueline multiple times, and the same can be seen in the second clip.
In the third clip, the Wolverines are pretty aggressive and in constant motion as the puck moves around the zone. In a diamond formation that’s vulnerable to cross-ice passes and one-timers from the far post, being in motion, pressuring and moving into shooting lanes is essential to killing the penalty. Even in more passive situations, you can see that the penalty killers are always reading and looking for a moment to be more aggressive and pressure.
Overall, the impression is that it’s a flexible system adaptable to any given situation and with a lot of opportunities for the players to make their own reads and react to what they see in front of them. When compared to the defensive zone video from Wednesday, you can tell that there’s just more freedom and flexibility, which is important as more teams switch to using multiple looks on the power play and adding wrinkles to their game.
Other thoughts…
Of course, I’d be remiss if I didn’t address the goaltending. The saying is that your goaltender has to be your best penalty killer, and Richter finalist Strauss Mann was absolutely that for Michigan. Even as the penalty kill improved by 31 spots, Michigan’s goaltending went from second-to-last to third in the country. Just in those few clips above, you can see that Mann is locked in and ready to make save after save — and his team is fully willing to put their bodies on the line for him and block shots.
From the press box, having covered the team in both seasons, it was fascinating to watch as Michigan settled into the system and really learned and embraced what they were supposed to be doing. In the first game of the year, they’d been working in the new system for less than a month, and you could tell. There was just a feeling that they were having to think hard about what to do in a situation rather than just reading and reacting to what they saw. By the end of the year, it had settled in, and Mayotte told me in early February that he thought this kill could’ve been up there with some of the best he’s coached.
I hope you enjoyed the second installment of this series! Next up, I’ll be breaking down the biggest statistical differences between the two seasons based on the data I tracked. If there are any questions you’d like to see explored in this, please reach out on Twitter and let me know — I am open to any and all ideas!
Still haven’t subscribed?