• The Carlyle Effect

    by  • April 14, 2014 • Hockey • 12 Comments

    One of the biggest changes in my views as I’ve gotten deeper into looking at hockey through a number lens is my belief in the impact of coaching. I was initially a skeptic on the value of coaching. I still am, to a degree. In many cases, I think that that which we attribute to coaching is just chance – pucks going or not going in. Are there 30 guys out there who can more or less get it right? I’m inclined to think that the answer is yes.

    That being said, I believe in evidence. The more closely I’ve looked at this stuff, the more I’ve come to think that there are some coaches who, for whatever reason, see their team’s Corsi% fall when they’re there and rise when they leave. I suspect that Michel Therrien is one of these guys, as I wrote over at Sportsnet a little while back.

    I’m also pretty suspicious of Randy Carlyle. I went back and took a look at how the Ducks did from 2011-13 under Carlyle and then under Boudreau. I also took a look at how the Maple Leafs did from 2011-13 under Ron Wilson and then what happened under Randy Carlyle.

    I like to break 5v5 down into seven different game states, based on whether it’s within X seconds of the last faceoff, depending on the location and result of the faceoff and, if it’s not, I use a seventh state that I call open play. That’s all discussed and set out here. So, OZ-, for example, means an offensive zone faceoff loss and we’re looking at what happened in the 21 seconds after that.

    Anaheim first:

    I’ve already discussed the absurd increase following offensive zone faceoff losses (and the fascinating history of Boudreau teams in these situations) here. Boudreau’s team upped their Corsi% in every single facet of the game.

    What’s that worth? Well, the average team this year had a Corsi% of 50% on 3535 shot attempts for and 3535 shot attempts against. So that’s 7070 overall. A team with a a 44.8% Corsi%, which Carlyle’s teams did over his last year and a bit, would have 3167 shot attempts for and 3903 against. Contrast that with a team posting a 48.7% Corsi%. They’d generate 3443 shot attempts for and allow 3627 shot attempts.

    Now convert that to shots. 53.5% of shot attempts this year became shots on goal, so I’ll run things through that way. That leaves a Boudreau team getting outshot 1940-1842 at 5v5 and a Carlyle team getting outshot 2088-1694 at 5v5. Now let’s turn that into goals, assuming an 8% shooting percentage and a .920 save percentage. Our Boudreau team gets outscored 155-147 and our Carlyle team gets outscored 167-136. So a difference of 24 goals, or eight standings points. Oof.

    Remember – I’m just assuming a generic team here. The Ducks have probably been a better than average goaltending and finishing team. In order to justify the Carlyle Tax though, he has to generate better shooting percentage and save percentage to make up for massive extent by which they’re getting outshot. In the years that I’m looking at, the Ducks shot 7.7% at 5v5 under Carlyle, with a .919 save percentage. They shot 8.5% under Boudreau, with a .923 save percentage. There’s a slight edge to Boudreau there – a PDO of 996 versus 1008. I’m not really inclined to believe that Boudreau is responsible for that but you certainly can’t argue that whatever Carlyle was giving up in shots paid off in goals.

    Let’s look at Toronto. Can we identify a similar difference between Wilson and Carlyle?

    It’s the same process in reverse. What was once a below average Corsi% became a trainwreck. And this year, which isn’t taken into account, it got worse. What was a 48.6% Corsi% in Wilson’s last two years became a 43.6% Corsi% team despite, in my view, having significantly better talent.

    I won’t bother with repeating the arithmetic that I did for the Ducks but you can see yourself that, if anything, the hole was deeper this year. It’s only worth it if it pays off in terms of percentages. Did it? Well, Wilson’s Leafs posted an 8.5% S% and a .918 SV% at 5v5 when he coached between 2010-12. Carlyle’s Leafs? They posted a 9.7% S% and a .917 SV%.

    As with the Boudreau/Carlyle example, I’m extremely hesitant to give the coach much in the way of credit for the percentages absent a very compelling argument. The stats guys were all screaming that the Leafs 5v5 S% (the engine behind their playoff run) would regress and, sure enough, it did substantially.

    In the big picture, we’re seeing the same thing. Carlyle’s teams go in the toilet in terms of shot differential. The most important question to ask here is why. For me, I’m a believer in tactics, things that are real and tangible. My best guess is that Carlyle asks his teams to do things that are not beneficial. It’s possible that he’s just a jerk and people despise him and that’s why the numbers are bad but even if that’s the case, it’s going to show up in the play. His teams have to be doing things differently than the teams coached by Boudreau (when he replaced him) and Wilson (prior to being replaced).

    Is Carlyle a bad coach? It’s possible. It’s also possible that he’s a good coach with some bad tactics who would be just fine if the Leafs identified this and worked to figure out a better angle from which to approach things. There are a lot of skills that a coach has to have, of which tactical decisions are just one. It’s possible that Carlyle’s fantastic at other things.

    Either way, if the Maple Leafs are looking for a riddle for Brendan Shanahan to solve that would pay immediate dividends, figuring out why Anaheim got so much better and Toronto got so much in terms of Corsi% be a start. It’d go a very long way to making the Leafs (who I think have a lot of very good players) an actual legitimate playoff contender, as opposed to a team hoping to sneak in.

    Email Tyler Dellow at tyler@mc79hockey.com


    12 Responses to The Carlyle Effect

    1. daryl
      April 14, 2014 at

      Cool. How about Vigneault and Torts?

    2. rickpaolucci
      April 14, 2014 at

      Randy’s numbers were bad all year, yet no one in the organization said anything. A house league B team would have beat them. At least the B team would have showed up if there was BEER after the game. Shame on the WHOLE management team for being totally Pathetic!!!!

    3. MirinovsNose
      April 14, 2014 at

      Do you know what kind of systems Boudreau uses?

      It’s frustrating watching the Leafs constantly chip the puck out of their own end into the neutral zone only to have their opponents bring it back in, or dump the puck into the OZ then try to take it away from the other teams D. It doesn’t seem surprising that a team would have poor possession numbers when a large part of their strategy is giving up possession of the puck.

      It’s also interesting to me that the players Carlyle has had problems with – Grabovski, MacArthur, Kadri, Gardiner, Liles – are guys who try to do something with the puck when they get it rather than chip it ahead.

      I know that people have tracked zone entries, for stretches anyway, has anyone tracked DZ exits? ie chip outs vs controlled exits?

      • May 7, 2014 at

        Hey, thanks for all the cetmonms! I am back working on Go Bonsai after a long hiatus and have already released a new version that looks way better and also runs faster. I also added a debug console by pressing ~ if you feel like playing around with the internals. There will me more updates to come in the following weeks.

    4. voline
      April 14, 2014 at

      Typo in the Leaf chart or the paragraph following:

      “What was a 48.6% Corsi% in Wilson’s last two years became a 43.6% Corsi% team despite, in my view, having significantly better talent.”

      The chart above shows the Leafs under Carlyle as a 45.6% Corsi% team not 43.6%.

      • April 15, 2014 at

        That isn’t a typo. The 43.6% is this season’s number, which is not included in the chart; he says this immediately preceding the part of the paragraph you cite.

        It’s quite possible that Carlyle’s idea of tactics are simply outdated, and are suited to the types of players that were much more common before the 2004 lockout: the pluggers, the dead-puck pylons, the guys who had to absorb tons of punishment to create chances. It would explain a lot about his reputation as a defensive coach when his recent teams are terrible defensively; it would explain why he favors the chip and chase; it would definitely explain why the Leafs management was fixated on “quality chances” – they were quite rare in the dead-puck era and usually involved mucking about for rebounds and deflections, not quick-strikes off the rush and dangles and sharp passes.

        The Leafs look like Lightning McQueen trying to haul around that road resurfacer in Cars. It’s not what they’re really built for.

        • D. Green
          April 17, 2014 at

          Actually it must still have been a typo as the overall Corsi works out at 44.98 not 43.6 and of course it was being compared to the 48.6 figure which was the OP number

          None of it good but maybe a fraction better !

      • May 6, 2014 at

        Thanks for checking out my game eroeyvne! I would love to add more stuff to Go Bonsai but have been really busy with other stuff. I plan to eventually offer multiple tree types, wiring, and even cross pollination. I doubt I will incorporate watering though because real bonsai trees need to be watered almost daily and that would get extremely tedious and repetitive.

    5. Voline
      April 14, 2014 at

      Typo 2, missing word in first sentence of last paragraph:

      “Either way, if the Maple Leafs are looking for a riddle for Brendan Shanahan to solve that would pay immediate dividends, figuring out why Anaheim got so much better and Toronto got so much [worse] in terms of Corsi% be a start.”

      Good work, Tyler. There’s a definitely a pattern at work. Also, I like that your conclusion is more nuanced and judicious than “Randy sux at everything!”

    6. Clarko
      April 15, 2014 at

      So we’re suggesting Carlyle can take a team with supposedly better talent and actually drag its numbers down. Did you consider reviewing the stats for the first couple years that Carlyle assumed the reign as the Ducks coach? Did he have a similar effect, or did the talent supercede his influence?

    7. D. Green
      April 17, 2014 at

      What does surprise me (that I hadn’t realised) is the poor overall Corsi of the Ducks both before and after

      In addition the one area Carlyle gets killed with the Ducks (oz-) is the one area of improvement with the Leafs

      But overall the OP figure seems to dwarf the face off figures in any case so while the face off figures might be important the overall Corsi is the bigger indicator IMO

      But of course all great stuff ……

    8. chelch
      May 2, 2014 at

      It’s May and Randy-bo-bandy is still the coach. So maybe your theory that Shanny will adjust the tactics and keep the coach is sound. Randy is one of – if not the – best line-matching coaches in the league. Randy (and Reimer) is still the reason they took Boston to seven games last year, and why Kessel was so productive.

      But of the available coaches, which one would be at the top of your list, including Carlyle?

    Leave a Reply

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