Subban didn’t trust Markov in the playoffs

Subban didn’t trust Markov in the playoffs

With the advent of microstat tracking, we can learn more about the game of hockey than ever before. The level at which we can dissect the game is new, and multi-faceted.

One newly tracked group of statistics that interests me is the tendencies a single player has with the puck on their stick in various zones. Not only do you have a unique way to view the decision making of a player, but you can also compare different stretches of games to see how tendencies changed, how they effected game play, and why.

One such example is the drastic change in P.K. Subban’s tendencies in the defensive zone from the 2014-15 regular season, to the 2014-15 playoffs.

When we look at plays individuals can make in the defensive zone, there are three general categories; passing, dumping the puck out, or an individual play. When we compare Subban’s tendencies from the regular season to the playoffs, there’s a very stark difference.

Simple tendencies

Subban’s most common decision during the regular season was to make a pass, in fact it’s what he did a whopping 68% of the time. However in the playoffs, he passed the puck in the defensive zone just 50.7% of the time, nearly a 20% drop.

As a result, the rate at which he dumped the puck out of the zone nearly doubled, from 16% to 30.2%, and his reliance on his own skill to move the puck increased from 16.8% to 19.3%.

Throughout Subban’s career, even with the changes brought on by Michel Therrien, Subban’s decision making has remained relatively stable in the defensive zone, so this big of a change forces us to ask questions, specifically why his game changed so much.

A common response would be the tighter checking playoffs would necessitate more safe plays like dump outs, but the individual plays are perhaps the most risky, and would happen less if Subban was more tightly checked. Also worth noting is that other defensemen on the Canadiens did not experience this stark change in zone tendencies.

If we break Subban’s zone tendencies down further, we can see what exactly he was doing less, and what he was doing more.


The biggest drop from regular season to postseason was in D to D passes, with Subban passing to Andrei Markov 26.7% of the time in the regular season, and just 15.5% of the time in the playoffs. Why would he stop passing to Markov?

Well if you recall, Markov had a mystery injury to his wrist in the playoffs, which caused him to misplay the puck on several occassions to disastrous results, such as this.

Because coaches are smart and know the game, this lack of trust in his partner would be picked up immediately, and opponents likely cut off Subban from outlet passes to force him to go to Markov, which they knew he didn’t want to do. That explains the drop in outlet passes, as well as the sharp increase in stretch passes.

Subban is an offense-minded player, and because of this, he’s going to want to find a way to turn a defensive play into an offensive one. In the situation he was in during the playoffs, this would lead to him taking more risks to get the puck not only out of trouble, but transitioned into offence, hence the higher volume of controlled exits.

The bigger problem with Subban’s situation though, is that he was still checked extremely tightly, meaning that often he had no option to create offense, and was forced to dump the puck out just to ease the pressure.

You might be wondering why the Canadiens didn’t just shuffled Markov down the lineup if he was so injured, but remember that Nathan Beaulieu was out of commission with a broken sternum, and Alexei Emelin was struggling mightily, those were the only two options the Habs had on left defense, with Tom Gilbert shifted to the left on the third pair to cover for the injured Beaulieu, and help insulate rookie Greg Pateryn.

Perhaps the most amazing thing about the drastic change in Subban’s style of play in the defensive end, is that his effectiveness didn’t really change all that much. He led the Canadiens in points in the playoffs, and the Canadiens controlled 53.01% of shot attempts while he was on the ice, always facing top competition.

All this together is just another example of how adaptable, and dominant P.K. Subban is, as one of the best defensemen in the game, even when forced to play a style that is incredibly inefficient.


  • Steve Jul 15, 2015 Reply

    Have you considered the tight checking may have taken away the pass to Markov? If teams watched how often PK passed to Markov in the regular season, and how often that play translated into an offensive rush, it would make sense they’d try to take that pass away and force PK to carry the puck, dump it out, or force a stretch pass. I’m not saying this is the case, but when you referred to the “common response” this is what I thought was coming. The data certainly correlates/seems to support this hypothesis.

    • Andrew Berkshire Jul 15, 2015 Reply

      That was my first thought, but as it says in the piece, no other defenseman on the Canadiens saw such wild changes to their tendencies.

      • Olivier Jul 16, 2015 Reply

        No other defenseman on the Canadiens has a Norris Trophy as well, it would be normal for PK Subban to be subject to more forechecking from the other team.

  • Eric Jul 15, 2015 Reply

    I would agree with what Andrew said. Even just watching, you could see a change in the way P.K. played during the playoffs. Markov seemed to be much less of an option, and more of a liability every time he had the puck. It was clear he was not playing well, and I think a big part of it had to do with speed. Markov has slowed down quite a bit and it was easily exploited throughout the playoffs. It always felt like he was step behind and was never able to catch up. As much as I like Markov, I think his time as a one of the ‘go to’ defencemen is over and it’s time for people like Petry, Beaulieau, etc. to take over in that role

  • Kerry Jul 16, 2015 Reply

    Now that I think of it, it did seem that PK was trying to carry the puck more rather than pass. The opposition caught on to it pretty quick and stifled him knowing they couldn’t allow him to get his speed up. With his speed and wicked slap shot, he`s a very real threat

  • Jeanlou Jul 16, 2015 Reply

    Markov’s play was so erratic, it even affected your own efficacy, at SPORTLOGiQ ! ; )

    In the first bar graph, the first blue bar shows only 69.2 %, not 79%, which would translate to a 20% drop, not 30%, as the text suggests. The second blue bar shows 16.5%, not 11.5%, and the third one shows 16.2%, not 9.4%.

    Same thing in the second graph, where D to D is indicated as 31.3% in the text but shows as 27% in the graph, etc… Somehow, only the blue regular season bars seem erroneous.

    I suspect that the numbers in the text are accurate and that the graphic artist’s concentration was just thrown off by Markov’s awful performance. :p But the result is that the ill effect of Markov’s erratic play on PK is visually attenuated and will look much worse when accurately represented.

    Great job otherwise!

    • Andrew Berkshire Jul 16, 2015 Reply

      There were some fun excel errors on the first draft, it’s all fixed now! Thank you.

  • David Johnson Jul 16, 2015 Reply

    Interesting stuff. I have one question though. You write:

    “Because coaches are smart and know the game, this lack of trust in his partner would be picked up immediately, and opponents likely cut off Subban from outlet passes to force him to go to Markov, which they knew he didn’t want to do.”

    This is quite possibly correct but is there any way to test this with your tracking technology? What exactly are the opponents doing to “cut off” Subban from outlet passes? Shifting players to his side from Markov’s? Pressuring Subban more?

    Also, it might be interesting to look at Subban’s regular season defensive zone tendencies against only the team(s) they played in the playoffs. Could it be it is just how a particular team defends against Subban and not related to Markov’s injury?

    • Andrew Berkshire Jul 16, 2015 Reply

      We’re working on tracking not only where the event took place, but where all players were on the ice and what they’re doing for each event. Adding that information would take a lot more time, as you can imagine, and the focus right now is on building the database in general, and improving the user interface for when we launch something everyone can use.

      As for what opponents were doing to Subban, from video scouting when I wrote the article, they were doubling up checkers on Subban and leaving Markov open a lot of the time, forechecking him much quicker than usual deep in the zone when most times you’d see a defensemen slide the puck behind the net along the boards to his D-partner, only Subban wouldn’t do that often.

      As for looking at how Subban played against just the playoff teams, I’ll have to check!

  • Rich Compton Jul 17, 2015 Reply

    Either way looking at this article; I cannot wait until we have access to these stats. This is a ridiculous amount of information.

    Well done guys!!!

  • Fred Jul 17, 2015 Reply

    Considering the Habs players 12 playoff games, I think it would be more statistically relevant if we looked at subban’s stats broken down by 12 game blocks (1 to 12, 2 to 13, 71 to 82) That would give us 71 blocks. Then I would be surprised to see where in the curve that those 71 blocks would create would the playoff 12 game block fall.

    I am not saying that Markov played anywhere close to well in the play-offs, I am just saying your method is probably not optimal to do the comparison

Your email is safe with us.

All rights reserved © SportLogIQ   Web Design Strategy by iMediaDesigns