• A Shareholder’s Lament: Part I

    by  • September 5, 2013 • Hockey • 5 Comments

    For three seasons, the Toronto Maple Leafs were in pretty good shape when Dion Phaneuf and Mikhail Grabovski were on the ice. Between 2009-12, the Leafs played 910:51 with those two players on the ice, piling up a Corsi% of 52.6% and scoring 63.1% of the goals during that time. 52.6% is a better than average Corsi% for a first line and Grabovski actually did better away from Phaneuf in terms of Corsi% than he did with him, although it’s worth mentioning that the level of competition was probably higher with Phaneuf on the ice.

    In 2013, the Grabovski/Phaneuf pairing posted absolutely putrid numbers – in 269:39 of 5v5 TOI, they posted a 41.4% Corsi% and got just 32% of the goals scored. Goals are, in a lot of ways, luck driven – sometimes you get a bounce, sometimes you don’t. Corsi% doesn’t lie though and the Leafs were horrible. Consequences ensued – Grabovski was bought out at a price of about $1.8MM per year over the next years, albeit with no salary cap penalty. MLSE’s PR guy over at TSN has been musing darkly about Phaneuf’s future in Toronto if he wants $6.5MM.

    Screen Shot 2013-09-05 at 10.57.24 PM

    You can see from the table at left just how good it was with Grabovski/Phaneuf on the ice together in the first three years and how bad it was in 2013. As a Rogers/Bell shareholder, I’ve been kind of horrified at all of this. For the next eight years, $1.8MM that could have been paid out in dividends to people like me will instead be paid to Mikhail Grabovski. I’ve got nothing against Grabovski but I didn’t invest in Rogers/Bell for them to waste money on guys who aren’t generating value for the shareholders like me.

    If nothing else, this seems like a sort of catastrophic business failure and one for which there seems to have been little in the way of public explanation. I’ve been thinking a lot lately about the intersection of data and what hockey teams do and thought that this might be an interesting test case: can we take data about Phaneuf/Grabovski, translate it from numbers into an understanding of how things went when they were on the ice and isolate what was different in 2013 from previous years? I think we can. We’ll see.

    One of the cool things that we can do with time stamped data is generate a count of how many times there were “shifts” with 1 shot attempt for (SAF), 2 SAF, etc. We can also do this for shot attempts against. I say “shifts” because the data doesn’t let us precisely isolate shifts – I group shot attempts within 60 seconds of each other as part of a shift. I think it works pretty well although, unfortunately, it doesn’t let us count shifts with zero shot attempts.

    Nevertheless, it does give us some insight. One of the things I like about this is that it lets us see if there’s something that’s changed in terms of the shot volume. If, for example, all of the shifts on which the Leafs got multiple shot attempts dried up, it might tell us that they weren’t as good at recovering pucks as they once were. That looks to have been a bit of an issue with some of the Oilers in 2013. It doesn’t seem to have been with Grabovski and Phaneuf.

    Screen Shot 2013-09-05 at 10.58.56 PM

    The bottom two rows of this table make the most sense to focus on. What we see is that, on those shifts on which the Leafs got at least one SAF with Phaneuf/Grabovski on the ice in 2013, they got exactly 1 SAF 67% of the time. This was actually an improvement over the pre-2013 figure of 70% – lower is better here, because it means that fewer of your shifts were 1 SAF shifts and more were multi-SAF shifts.

    In fact, with Grabovski/Phaneuf on the ice in 2013, the Leafs were slightly more effective than they’d be historically, once they’d generated at least one SAF. You can evaluate this by asking, for example, “How many SAF would the Leafs get over 100 shifts with at least one SAF in 2013 versus pre-2013?” It’s an easy enough question to answer arithmetically and it turns out that the answer is 142 in 2013 and 141.6 pre-2013. Essentially the same.

    What if we look at things in terms of SAA? Again, 2013 looks pretty similar to pre-2013.

    Screen Shot 2013-09-05 at 10.59.02 PM

    The Leafs were slightly worse in 2013 in terms of preventing shifts from becoming multi-SAA shifts with Phaneuf/Grabovski on the ice than they were in pre-2013. The percentage of single-SAA shifts was down slightly and there were a few more multi-SAA shifts. If you ask yourself the question I posed above but in terms of SAA, How many SAA would the Leafs get over 100 shifts with at least one SAA in 2013 versus pre-2013?” you see a slight increase in the rate at which the Leafs allowed SAA, given that they’d allowed at least one: from 152.2 to 157.9.

    This SAA/100 shifts with at least one SAA is kind of useful as a metric in terms of how good you are at cleaning up rebounds and clearing your zone. What sort of catches my eye here is that the Leafs strength with Grabovski/Phaneuf on the ice up until 2013 is that it must have been built on there being many, many more shifts where they got shot attempts than the opposition because they weren’t outshooting by being better at penning the opposition in their zone.

    Let me explain that. Prior to 2013, the Maple Leafs got 141.6 SAF/100 shifts with at least one SAF with Phaneuf/Grabovski on the ice. They allowed 152.2 SAA/100 shifts with at least one SAA with Grabovski/Phaneuf on the ice. If they had an equivalent number of shifts on which they got shot attempts as shifts on which they allowed them, then Grabovski/Phaneuf would have had a sub-50% Corsi%. They didn’t have a sub-50% Corsi%, so they must have made up for what they lacked in terms of turning single-SAF shifts into multi-SAF shifts in volume.

    Screen Shot 2013-09-05 at 10.59.15 PM

    Turns out, that logic holds true. The Leafs had a very consistent edge in terms of the percentage of shifts with Grabovski/Phaneuf on the ice and at least one SAF (as opposed to at least one SAA) that completely collapsed last year. If the Leafs had generated and allowed multi SAF/SAA shifts at the same rates as they did prior to 2013, their Corsi% with Phaneuf/Grabovski on the ice would have been 41.8%. Accordingly, I think that we can reasonably conclude that the small change in the rates at which those events occurred had a slight impact on their Corsi%

    This is helpful to know because it narrows the issue: the Leafs’ Corsi% last year with Grabovski/Phaneuf on the ice didn’t collapse because of a change in the rate at which multi-SAF and multi-SAA shifts occurred; it collapsed because the Leafs suddenly became extraordinarily poor at generating the first SAF and preventing the first SAA. If you’re blaming Korbinian Holzer or Mike Kostka or Jay McClement for this, you need to come up with a convincing explanation as to why their impact was felt in terms of the likelihood of the first shot attempt occurring, but not really on subsequent ones.

    This is a first step in figuring this out, but I think it’s an important one. The question that we need to answer gets simplified a little bit by this I think – it morphs from “Why did the Leafs Corsi% with Grabovski/Phaneuf fall from north of 52% to below 41%?” to “Why were the Leafs so much more likely to allow at least one shot attempt against and so much less likely to generate at least one shot attempt for on a per shift basis with Phaneuf/Grabovski on the ice in 2013?”

    We’ll dig into that next week.

    Email Tyler Dellow at tyler@mc79hockey.com

    About

    5 Responses to A Shareholder’s Lament: Part I

    1. clrkaitken
      September 5, 2013 at

      “If you’re blaming Korbinian Holzer or Mike Kostka or Jay McClement for this, you need to come up with a convincing explanation as to why their impact was felt in terms of the likelihood of the first shot attempt occurring, but not really on subsequent ones.”

      Grabovski’s Zone Start % in 2012/13 was 36.7%. Compare that to 53.0, 50.2, and 55.1 in the previous years you looked at.

      Phaneuf’s Zone Start % in 2012/13 was 41.3%. Compare that to 50.0, 49.8, and 54.8 in the previous years you looked at.

      (Behind The Net)

      Now I’m not sure how you would be able to tease out a Grabovski/Phaneuf Zone Start %, but given what we know about Carlyle’s usage tendencies last season, and their overall Zone Starts, and their overall high level of competition, I think you can reasonably conclude the 2012/13 MG/DP Zone Start % is also considerably lower than the three previous years.

      So, if as you’ve concluded – the Leafs’ Corsi% last year with Grabovski/Phaneuf on the ice didn’t collapse because of a change in the rate at which multi-SAF and multi-SAA shifts occurred; it collapsed because the Leafs suddenly became extraordinarily poor at generating the first SAF and preventing the first SAA – then the answer may lie in the zone start data.

      Said another way; 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?

    2. Flanders Breivik
      September 5, 2013 at

      I’ve harped on Carlyle’s defensive zone strategy before, so I won’t get that far into it, but Grabovski’s ozone start percentage the last four years went 55.1, 50.2, 53.0, 36.7, and then Phaneuf went 54.8, 49.4, 50.0, 41.3. The shift starting in their zone is going to be something of a big deal.

      An interesting point is that despite everything, Grabovski still had a positive Corsi Rel in both the regular season and the playoffs (Phaneuf was negative, but I also think he was hampered by way, way shittier linemates, getting more time with Bozak, Kostka, Holzer, etc). Something else about Phaneuf’s defense partners, quality of competition, and zone starts, is that if you go down the WOWYs, there are very few players who aren’t way better without Phaneuf. The goons, Gardiner, MacArthur (both are way worse apart), and Holzer. The fact that there’s such a massive swing for guys like Orr/McLaren is kind of indicative, since they so rarely get defensive zone shifts.

      On Grabovski, I know it’s hardly indicative of anything, since we know McLaren and Orr are ECHLers, but I think it’s hilarious that Jay McClement gets defensive zone starts with Grabovski, and otherwise played with Orr and McLaren, who got easier minutes, and yet he’s at 42.2% with Grabovski and 29.1% otherwise.

      Once I’ve become better acquainted with math (tenth grade dropout here), I’ll try to confirm my theories on Carlyle and defensive zone play by seeing if Corsi percentage’s correlation with zone starts gets stronger coach by coach.

    3. Pingback: Is success partly to blame for Dion Phaneuf's and Mikhail Grabovski's poor corsi last year? - HockeyAnalysis.com

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

    5. Pingback: Phaneuf and Grabovski’s shifts | Extra Skater Blog

    Leave a Reply

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