Welcome back to Week Twenty-Three of the Detroit Power Play Update series. Each week, I’ll take a look back at the previous week’s power play performance. At the bottom, you will find links to the previous editions. The game played on Tuesday, March 12 will be included in next week’s edition.
5v4 Stats Update
The following three charts show the change for each team in the past week, for three metrics: TOI/GP, GF60, FF60. The teams are arranged from the current highest to lowest numbers. The arrow indicates if they have improved or declined over the last week. To be clear “last week” includes all games from the beginning of the season through March 4, and “this week” includes all games from the beginning of the season through March 11. Detroit is highlighted in red.
In four games, Detroit had thirteen 5v4 opportunities. They did a better job this past week of drawing penalties than in most weeks. Looking at the length of their arrow, you can see that they have the greatest positive change (Minnesota is pretty close). Of course, given the dearth of power plays drawn all season, it remains unlikely that the team will finish the season too much higher than they currently are.
The rate of unblocked shots declined slightly from last week. The team continues to be at or near the bottom of the league in generating unblocked shots.
With the three goals this week, Detroit moves up. It’ll still take a lot for them to move too much higher with 13 games remaining (and Larkin out for another week or so).
Five of the power play opportunities came against Tampa Bay, who is the best in the league at preventing goals against when down a player. This isn’t surprising when you watch them in action. Detroit has historically had trouble against them, although at least the Wings scored one 5v4 goal in that game. Colorado and the Rangers are worse than the Florida teams at preventing 4v5 goals, and Detroit went a combined 1 for 6 in those two games. Definitely a missed opportunity, especially in the Rangers game.
Shot Locations
This edition is going to be a little shorter than usual in terms of video clips, but I hope to make up for with with a new visual I’ve been working on.
Before I get to it, this week definitely exemplifies the reason that a good analysis looks at multiple areas. For example, looking at statistics can definitely give you an insight into performance, but it doesn’t tell the whole story. By the same token, relying solely on the eye test is a mistake because it’s impossible to focus on and remember everything you saw.
That’s the reason why these articles have included both statistics and video from the start. If you were to just look at the charts above, you might think that we did better this week than last week, at least for the most part.
If you re-watched the thirteen times we were on the man advantage, you would likely have a different takeaway.
For example, here are the shot locations from last week. Many of the shots were taken from dangerous scoring locations, and a good amount are above the roughly 8% average expected goal rate of a 5v4 shot, according to Evolving Wild’s xG model.
Like past editions laid out in more detail, the EW model, while very good, is limited by the data it is able to access. For example, in a perfect world, it would be able to take into account the position of every player on the ice, including the goalie, as well as the location of the pass. While the shot that Thomas Vanek scored on in the Florida game (top blue triangle) has an xG value of 11.3%, watching the clip shows you that in this instance, that value should likely be higher.
While everyone was excited to see Zadina score his first goal (lowest blue triangle), the model says that it was a below average scoring chance. Looking at the video, which we will do a little later, that looks about right. The pass comes from the other side of the ice, but the goalie looks like he has gotten across in time to make that save. Of course, the model doesn’t take that into account, but I’m pointing out that you can have outliers in both directions.
Access to player and puck tracking would also make what I’m about to show much easier. Our good friend Prashanth Iyer used R to create the ability to make animations of plays. He made a version for Ryan Stimson to build off of for his talk at SEAHAC (Seattle Hockey Analytics Conference) this past weekend. I have had an idea to do something like that in the back of my mind, so I decided to make my own version.
The actual R code is pretty simple and not very long at all. The hard part is that to make each play, I have to step through the play in 0.5 second increments and plot each player and the puck, which takes about 25-30 minutes for a 14 second play. I do think it could be useful to illustrate some topics. For this first one, I’m going to go back to something I’ve talked about before: having two players below the goal line on a 5v3.
This first animation is taken from the last game we played against Calgary. It starts from an offensive zone face-off. Off the draw, Detroit starts to spread their players out, and Nyquist goes towards the bottom of your screen. Calgary is still able to keep their three players in a relatively stable formation. Near the end of the play, when Vanek slides below the goal line, Calgary’s formation breaks down. It’s one thing to have to turn your head as a defender to keep track of one player behind your goal line, it’s another to have to keep track of two when you still have three in front of you.
The reason I think these animations can be valuable is that they take away all the distractions, and allow you to just focus on the player positioning. At the 16 second mark (the animation is slowed down to -1.5x speed), 55 goes to Nyquist, leaving Larkin alone, which forces 24 to move to cover him in the slot. When the puck goes back to Vanek, 24 is in a real bind, and Vanek is able to get the pass to Larkin. Even though Larkin couldn’t put it in, the reason he got such a great chance is largely due to numbers 24 and 55 have to react to both Vanek and Nyquist being below the goal line.
The second animation is a 5v3 against Ottawa. Notice that even though at one point Vanek and Nyquist are technically below the goal line at the same time, the puck movement is slower, and the player movement doesn’t really force Ottawa to have to break their formation. When Mantha moves up with the puck, Nyquist’s positioning doesn’t help at all.
Since Mantha is moving the Senators defense towards the top of our screen, Nyquist sliding out to the left would put him in position to do something. While I still think that having 2 players below the goal line is a good idea in general at 5v3, the overall idea needs to be to force the defense out of their position through quick puck movement and smart player movement.
For example, what if the other players had done this when Mantha moved towards the top of the screen with the puck?
Look at each defenseman and look at the decisions each would have to make. All you need is for one of them to make a bad read.
I’m still tweaking these animations, and I’m hoping to keep improving on the design. I hope to have some more for you in the near future.
Let’s Go to the Videotape!
There won’t be too many clips this week because the bad things would be repetitive. For the most part, it was poor execution that was letting the team down. Bad passes, the inability to control a pass, missed reads: it wasn’t that they were typically trying to do the wrong thing, it was that they weren’t executing properly.
Here’s an example. This breakout has worked very well recently with these three players doing something similar. Here, Andreas Athanasiou leaves the puck a little further away from Dylan Larkin than typical, causing Larkin to have to change his trajectory. The ripple effect is Larkin making a bad pass, probably because the timing was off between him and Athanasiou.
Here’s another example. This clip starts with Filip Zadina making a nice pass, looking for a Vanek deflection. But when the puck goes back to Niklas Kronwall….I don’t know what he’s thinking. That pass is never getting to Zadina.
To be fair, everyone seemed a little off; it wasn’t just a couple players. In the next clip, this pass from Nielsen was not a bad one. We’ve seen plays just like this lead to sustained zone pressure. Vanek’s in the right position, and he normally has no trouble receiving this pass.
I could show 20 more clips, but I don’t see the point.
Additionally, not having Dylan Larkin after the Avalanche game showed just how important he’s become to that top unit. When doing my in-depth analysis, I think I was able to isolate the second largest source of the top unit’s issues (after Larkin’s absence).
Abdelkader should be nowhere near the power play, especially not on the top unit. I’d take any other forward except maybe Witkowski there. Give Christopher Ehn a shot on the second unit and bump Zadina up to the top unit.
Before I get to the good clips, please keep in mind that just because there are three bad clips and three good clips, that doesn’t mean there were as many good plays as there were bad ones. There were far more bad ones.
Everyone, however, was excited to see Zadina score his first NHL goal, and even though it probably should have been saved, it was good to see him get in a good position and take a quick shot.
Also, notice he was out there with most of the players from the top unit. Just saying.
Closing out this edition is two goals by Thomas Vanek. The first one was a very impressive tip. We’ve seen him make incredible deflections before, but it was still a little surprising he was able to use his backhand to deflect this puck in. As has been said numerous times in this series, Kronwall does not have the mobility he once did, but he can still make nice passes.
The second Vanek goal of the week came off a defensive breakdown, which left Vanek wide open, but it was still a nice pass from Nielsen. Also, let’s be honest. How many times have we seen a player miss that shot?
Also, I just want to point out Abdelkader nearly negating this goal.
So Far
Even though we scored three times, the power play overall was worse than last week. A lot of this has to do with Dylan Larkin being out. It also has to do with the coaching staff not making the proper lineup adjustments on the power play, in my opinion.
For the upcoming week, I’d like to see a top unit of Hronek – Athanasiou – Mantha – Rasmussen – Zadina. I’d also like to see Abdelkader on the bench during the power play, but I doubt I’ll get my wish there.
Previous Editions
Week One
Week Two
Week Three
Week Four
Week Five
Week Six
Week Seven
Week Eight
Week Nine
Week Ten
Week Eleven
Week Twelve – No Edition This Week
Week Thirteen
Week Fourteen
Week Fifteen
Week Sixteen
Week Seventeen – No Edition This Week
Week Eighteen
Week Nineteen
Week Twenty
Week Twenty-One
Week Twenty-Two
