• A Shareholder’s Lament: Pt. 2

    by  • September 6, 2013 • Hockey • 10 Comments

    Since I wrote the first part of my look into what caused the Grabovski/Phaneuf pairing to crater last season, two interesting points have been made. The first, in the comments to the previous post, comes from Twitter’s @clrkaitken:

    is it possible that the reason the Leafs (with Grabovksi/Phaneuf on the ice) became extraordinarily poor at generating the first SAF is because they were more likely than not to start the shift in their own zone, which means to generate the first shot they’d first have to take the puck the entire length of the ice?

    Screen Shot 2013-09-06 at 5.29.09 PM

    We can dispose of this summarily. I calculated a Corsi% for the Leafs with Grabovski/Phaneuf on the ice based only on shots taken more than 60 seconds after a faceoff, which is posted at left. Good theory – I believe zone starts and faceoff wins can matter – but there’s nothing to it. The Leafs got trounced even if we cut out shot attempts following faceoffs. That’s not it.

    David Johnson at Hockey Analytics chimed in with something interesting as well. He looked at the Corsi% of Phaneuf and Gabrovski (overall, not just limited to when they were on the ice together) and found:

    …about 30% of the drop off in the Leafs team and individual CF% from 2011-12 season to last season can be directly attributed to changes int he Leafs leading/trailing/tied ice time percentages. This means 30% of the drop off can be attributed to the Leafs being a far better team last year at getting leads and winning games. Or, if you believe that was largely due to lucky shooting you can say 30% of the Leafs drop off in CF% is due to good luck.

    The logic of his post, which checks out to me and is supported by data, is that the Leafs spent a greater percentage of their time leading this year, teams tend to have a worse Corsi% when they lead more and that a chunk of the decline seen in Grabovski and Phaneuf this year can be attributed to that. Intriguingly, his data also shows that Phaneuf and Grabovski experienced a big decline in their Corsi% when the Leafs were ahead or tied and much less of one when they were trailing.

    This is, I think, important. One of the things that was talked about a lot was the possibility that Grabovski, who suffered from some sort of intestinal problem, may have been physically unable to do what he once did. Unless it was an intestinal thing that affected him only when the Leafs were tied or leading, that theory becomes harder to support.

    I’ve been focusing on Phaneuf/Grabovski together, so I pulled the data for them when the Leafs were leading/trailing/tied and it’s pretty intriguing.

    Screen Shot 2013-09-06 at 5.29.32 PM

    When the Leafs were trailing last year and had Phaneuf/Grabovski on the ice, they essentially performed as well as they did in those situations prior to the 2013 season. When they were leading or tied? Where they posted a pretty solid Corsi% in the past, they took a 13-14 percentage point hit. That’s pretty astonishing and given how well they held up when they were trailing, I think it’s a pretty solid data point that might cause us to wonder about the tactics that the Leafs employed in those situations with Phaneuf/Grabovski.

    We know that Grabovski and Nikolai Kulemin were the Leafs who took the biggest Corsi% hits in 2013 compared to 2011-12. What if we break things out and look at the players who have significant time with the Leafs in 2009-12 and 2013?

    Screen Shot 2013-09-06 at 5.29.47 PM

    Screen Shot 2013-09-06 at 5.29.56 PM

    I’m going to summarize the changes in the three situations for these players to make it a little easier to digest. The table at left makes things a little more understandable, I think. We can see that, when the Leafs were trailing, there was a decline in Corsi% across the board for forwards with some history. Grabovski and Kulemin saw huge declines in their Corsi% when the Leafs were tied or leading relative to the past; Kessel and Kadri saw smaller (but noteworthy) declines in their Corsi% when tied and MacArthur saw a noteworthy decline in his Corsi% when the Leafs were leading.

    If you refer back to the big table above, you can see that Grabovski and Kulemin had a history of posting much better Corsi% numbers when tied or leading than did the other Leafs. Many of the other guys had a history of getting outshot fairly badly when the Leafs were tied or leading; that continued last year. While Grabovski and Kulemin dropped below those fellows in terms of Corsi%, they also had much higher to fall from.

    I think we can say that we’ve narrowed the issue again. All returning forwards saw a decline in their Corsi% when trailing – Grabovski and Kulemin don’t stand out there (and, as I noted above, Grabovski/Phaneuf were basically the same when they were on the ice in this situation together in 2013 as they were in prior years). There was a complete collapse when tied or leading with Grabovski/Phaneuf on the ice in 2013 and that’s probably where we should focus our attention. As noted above, I think the evidence here raises some questions about the tactics in those circumstances.

    The elephant I haven’t referred to yet is the Leafs’ third period collapse against Boston. Phaneuf was on the ice for two of the Bruins’ final four goals and Grabovski was out there for three of them. Given what we’ve found to date, it sort of seems as if the Leafs kind of lived the Lincoln line about how you can fool some of the people all of the time, all of the people some of the time but that you can’t fool all of the people all of the time. Their sins, in terms of getting crushed in shots (particularly with Phaneuf/Grabovski on the ice) when tied or leading, finally caught up with them.

    In the next post, I’ll dig into why this might have happened.

    Email Tyler Dellow at tyler@mc79hockey.com


    10 Responses to A Shareholder’s Lament: Pt. 2

    1. Kent Wilson
      September 6, 2013 at

      Thanks for all this. It’s great work.

    2. Jeremy Wright
      September 6, 2013 at

      It is beyond me and my level of comprehension how you are not yet a paid employee of an NHL management team. Fascinating work, all the time.

    3. Saj
      September 6, 2013 at

      Great stuff!

    4. Tach
      September 6, 2013 at

      I would be interested to see if the usage of Phaneuf and Grabovski changed depending on the score situation. Did Carlyle put them out there for more D-Zone draws when tied or leading than in prior seasons?

      I can’t claim to have watched enough Leags games to say if this is plausible.

    5. Back in Black
      September 7, 2013 at

      Coaching may have had an impact here. Regardless of what anyone feels about the relative merits of Wilson and Carlyle, it was frequently noted that Wilson’s Leafs were major outliers when it came to score effects and C%. While almost all teams saw decreased SF once they took the lead, the Wilson Leafs kept on shooting, IIRC.

      If Carlyle’s style has re-introduced the influence of score effects on C%, then we should logically expect most or all Leaf players to have a noticeable drop in their trailing C%.

      • Back in Black
        September 7, 2013 at

        Sorry, that last bit should have been about leading C%, not trailing C%.

      • Tyler Dellow
        September 7, 2013 at

        There was one year where that was true, according to behindthenet.

        	TIED	UP 2	UP 1
        2013	44.85	39.83	41.42
        2011-12	46.48	42.22	41.59
        2010-11	44.48	36.59	44.92
        2009-10	52.83	50.4	48.08
        2008-09	48.86	49.16	48.98

        I think you can see it in Wilson’s first year. Not after that. It doesn’t really make sense to me as an explanation.

        • Back in Black
          September 7, 2013 at

          Huh. My impression was that the phenomenon was more sustained then that, but it looks like I was mistaken. I might make an argument for it in year two, in that the drop may be present but unusually small compared to other teams. I know Hawerchuk described that Leafs season as having an “unwillingness to play to the score” but I can’t find a link other than this. Year three does have no drop-off between Tied and Up 1, but then a huge gulf (small sample?) at Up 2.

          I don’t see a lot of support here for my hypothesis. Carry on.

    6. Pingback: Spectors Hockey | NHL Blog Beat – September 8, 2013.

    7. CHuck
      September 9, 2013 at

      And the reason is… MacArthur?

      I’ve always thought that MacArthur and Grabovski drove the MGK line, with Kulemin tagging along.

    Leave a Reply

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