• Dion Phaneuf and the limitations of data

    by  • March 14, 2011 • Uncategorized • 6 Comments

    Old friend James Mirtle had a piece mentioning Dion Phaneuf’s struggles on the PP this week:

    Kaberle and Versteeg played the point on the top power-play unit much of the year, and in their absence, it has mainly been captain Dion Phaneuf taking on the role. A dominant power-play quarterback when he entered the league, Phaneuf has struggled in the role in Toronto, going from more than 30 power-play points in each of his first three seasons to just eight in 51 games this year.

    James’ piece isn’t really aimed at audience of statzis but I suspect that part of the answer to his question lies in Phaneuf having played an ever decreasing amount of 5v3 time, through no fault on his part. 5v3 time has fallen off dramatically since the first season back from the lockout, which coincided with Phaneuf’s rookie season. Calgary scored 13 5v3 goals during Phaneuf’s rookie season; Toronto has 4 this year. 13 goals would probably put you first or second in the league in 5v3 goals in 2010-11; Calgary finished in a three way tie for tenth in that category in 2005-06. Toronto is currently in a seven way tie for 11th in 5v3 goals. Fewer 5v3 goals means fewer points.

    That’s not all of it though – Phaneuf’s 5v4 numbers show a definite falloff.

    phaneuf

    A couple of points there. First of all, his shooting rate is awfully consistent. I’ve mentioned a few times that I’m always amazed at how consistent shooting rates are. His don’t move that much. I was talking to James about this and James mentioned that Phaneuf only has one goal or so in 75 games in Toronto. I’ve got a hard time getting all that worked up about it – on his historic numbers, he should have four or so. Over the small amount of sample that we’re talking about, it’s just not that much.

    Phaneuf’s teams have done much more poorly over the past two years at 5v4 than they did in the preceding two. You can see significantly lower shot rates and shooting percentages. Phaneuf himself has much lower numbers – after averaging 3.5 PPP/60 in 2007-09, he’s averaged 2.6 PPP/60 in 2009-11. A couple of things that aren’t in the table jump out at me as well. Phaneuf’s taking 33.4% of the Leafs’ shots when he’s on the ice at 5v4 this year after fluctuating between 26% and 28% in the three preceding years. Shots by defencemen are a lot less useful, because they tend to be taken from low percentage areas of the ice. Phaneuf has also got points on more goals over the past two years (58%) than he did in the two years before that (52%). That strikes me as sort of curious – if he’s struggling, why is he getting points on more goals and generating a greater proportion of the shots?

    I’m hesitant to say Phaneuf’s lower numbers mean that he’s struggled in the PP QB role. Unlike in baseball, where we measure the things that lead to runs (BA/OBP/SLG), in hockey, a goal is scored and we immediately start assigning credit for it on the basis of the last guys to touch the puck. We aren’t measuring the skills that lead to a goal; we’re inferring that people have skills that lead to goals on the basis of an arbitrary assignment of credit.

    We don’t really measure whatever skills might lead to a good PP in hockey. We know that more shots are, generally speaking, better than fewer, and that’s been a problem in Toronto (although Phaneuf is producing around his historical level). We know that shooting percentage is important as well, although it’s subject to vicious fluctuations. If we really wanted to get an idea of the skills that matter, we’d be charting every touch of the puck on the PP and getting an understanding of where a good PP team looks different from a bad one. Is it an inability to gain the line? A tendency to shoot from bad spots? A tendency to shoot without making the goalie move? With better data, all of this could be answered. As it stands, it can’t.

    About

    6 Responses to Dion Phaneuf and the limitations of data

    1. Sunny Mehta
      March 14, 2011 at

      nice post, Tyler.

    2. March 14, 2011 at

      Great insight.

      We aren’t measuring the skills that lead to a goal; we’re inferring that people have skills that lead to goals on the basis of an arbitrary assignment of credit.

      Great quote.

      If we really wanted to get an idea of the skills that matter, we’d be charting every touch of the puck

      I’d love that. If we could use technology to keep track of who touched the puck, where, and what happened to it, that would go a long way to answering a lot of questions.

    3. March 14, 2011 at

      What happened to your idea of charting passes for the Oilers PP and opponent PP? Was it just to much?

      I LOVE looking at the passing charts that have been developing over at BTN lately. I really think that could be the future of analyzing the power play, and I think that eventually you’ll find out that two things drive good PP’s: 1) the ability of the players to seamlessly move around to support each other on the PP (usually seeing defensemen swapping back and forth with the man on the half-boards), and 2) lateral passing.

    4. Tom Benjamin
      March 15, 2011 at

      We aren’t measuring the skills that lead to a goal; we’re inferring that people have skills that lead to goals on the basis of an arbitrary assignment of credit.

      This is a critical problem with all statistical analysis.

      We don’t really measure whatever skills might lead to a good PP in hockey. We know that more shots are, generally speaking, better than fewer, and that’s been a problem in Toronto (although Phaneuf is producing around his historical level). We know that shooting percentage is important as well, although it’s subject to vicious fluctuations. If we really wanted to get an idea of the skills that matter, we’d be charting every touch of the puck on the PP and getting an understanding of where a good PP team looks different from a bad one. Is it an inability to gain the line? A tendency to shoot from bad spots? A tendency to shoot without making the goalie move? With better data, all of this could be answered. As it stands, it can’t.

      I think there are two factors. First, it is gaining the zone with the puck. Setting up. The second part is after setup, the moving of the puck quickly enough to find an open man and a scoring chance. (The same two elements define good penalty killing, IMO. Does the team make it difficult for the opponent’s to set up? Once in the box, do they pressure the puck and pounce on the smallest bobble?)

      I’m pretty sure that if we could figure out a way to measure those elements, we would find that the good power play will fail to gain entry with the puck somewhat less frequently than the poor PP. Once set up, the good PP will find a scoring chance somewhat more frequently than the poor PP.

      Will that tell us much? It might show that teams are better off taking more risks to prevent the PP from setting up against them. That would be interesting. Still, I think it will all come down to the personnel.

      The Canucks have an excellent power play. To get set up they give the puck to a Sedin who usually carries it in. If the PK denies a Sedin entry, the puck is dumped in and Kesler goes and gets it. Once they get set up, Kesler goes to the net, and the pointmen get ready to shoot. The Sedins piddle around a bit with the puck and then somebody gets a scoring chance.

      Its hard for a team like the Oilers – or any other team – to draw any lessons from the way the Canucks do it.

    5. March 17, 2011 at

      The Canucks are all powerful!!

    6. Tom Benjamin
      March 18, 2011 at

      The Canucks are all powerful!!

      Probably not, but they are damned good. Anyone who likes to think about team building should be interested in Vancouver. If the Oilers end up with a contender some years down the road, everyone will understand how they did it.

      But Vancouver? How did they do it?

    Leave a Reply

    Your email address will not be published. Required fields are marked *