Not advanced statistics – Grabovski is still better

Advertisement - Continue Reading Below

Over the weekend, Ontario-native and sun-drenched Kitsilano resident Blake Murphy made an excellent case in breaking down the debate between the “advanced” statistics relating to Tyler Bozak and Mikhail Grabovski. It’s true that there are some good metrics out there used for evaluating certain aspects of defensive play that, yes, some NHL teams do use in making player evaluations, but I don’t think that the concepts are particularly advanced. There is definitely a lot of opposition to the objective reaction of Dave Nonis’ recent moves which is based on some mis-understandings.

But I want to circle back to Tyler Bozak, and I want to hammer home this point because Kessel will likely re-sign with the Toronto Maple Leafs and play out a large chunk of his career in Toronto, including the portion of his career where he can no longer be counted on to score 30 goals a season. Without a legitimate No. 1 centre by his side, he has become one of hockey’s most productive wingers, and barring his slump at the start of the season when pucks weren’t going in, an absolute treat to watch down the stretch and into the playoffs.

On the TSN Free Agency broadcast, I was a little peeved that the only regular panelist that seemed to encroach on my belief that Tyler Bozak’s contract was not a positive for the Leafs was Ray Ferraro. In The Reporters segments, Bruce Arthur straight up called Bozak a “lousy player” but that wasn’t about it. I noticed towards the end of the coverage, the tone had changed from “will Nonis get his man” to “Bozak isn’t a No. 1 centreman… but he’s not being paid like one” which sort of misses the point.

I haven’t yet gone to check what other teams are paying their No. 1 and No. 2 centremen, but we’ll have to check that at the start of the season once the rosters are set. The Dallas Stars introduced Tyler Seguin and Shawn Horcoff today, making $5.75-million and $5.5-million against the cap, respectively, and they’ll be the No. 1 and 2 centremen for the Stars. I don’t think that Bozak is even making what would be considered second-line centreman status, but that’s not really the point. The point is that Dave Nonis had the following good players:

Advertisement - Continue Reading Below
  • Phil Kessel
  • Joffrey Lupul
  • Mikhail Grabovski
  • Nikolai Kulemin
  • Clarke MacArthur
  • James van Riemsdyk
  • Nazem Kadri
  • Jay McClement
  • Dion Phaneuf
  • Carl Gunnarsson
  • Cody Franson
  • Jake Gardiner
  • James Reimer

I’m unsure on whether Lupul’s offence really washes out his defensive ineptitude, but let’s keep him there regardless. When injuries forced Bozak and Mark Fraser out of the lineup in the playoffs, we saw a different team. Granted, it was only three and change in games, but the Leafs seemed to work much better as a unit with those working parts making up the core. When Dave Bolland came along, you could have made the argument (and I did) that with Bolland pushing Grabovski and Kadri up the depth chart, the Leafs centremen were improved.

But that was torn down, and it leads to a lot of questions of who was “traded” for whom. Was Grabovski traded for Bozak? Grabovski for David Clarkson? Grabovski for David Bolland? There seems to be three “one or the other” options, but if I were running the Leafs, I would have sprung for Grabovski, Clarkson and Bolland and found a way to work that under the cap, since those are the three best players of the four.

Nonis didn’t. He went with Bozak, Clarkson and Bolland, and regardless of the end of this statement, I think it’s pretty obvious that the Leafs were making a bet between Bozak and Grabovski:

Advertisement - Continue Reading Below

“How Randy uses him you can easily draw that conclusion,” Nonis agreed of deeming Bozak a better fit for the club than Grabovski, denying however that their choice came down to one player over the other. “Not that Grabo isn’t a great player, it’s just for us, Bozie is a better fit; plays more minutes. But it wasn’t about Grabo or Bozak.”

(emphasis mine)

I think it’s an awful bet, and it has nothing to do with Corsi, Fenwick, PDO or whatever a few people in the comments will accuse me of doing. Not everything is about “advanced statistics”, but it’s just basic player forecasting. In the last six years, we have so much more data to work with than we ever did, and a lot of it is breaking down some of the old thoughts about hockey, which has changed so much in the last decade.

Gare Joyce, in his criticism of certain metrics, seemed to side with scouts that have “worked decades in a hockey rink”. It doesn’t seem to matter to guys like Joyce that the game has changed, and perhaps some scouts have kept up with that and some haven’t. The ability to raise the puck with a shot would have been seen as a huge plus in the 1950s, I’m guessing, but with the advent of curved sticks, it’s a real basic skill that anybody can do by the time they’re 8 if they’re playing hockey. There are so many good hockey players coming from so many non-traditional spaces and probably 1000 of them could be successful NHLers, but the league only has room for 690. The 800th best player in the world looks really really good and can do all these things, but he barely looks worse than the 500th best in a neutral setting? How do you make a decision between the two?

Advertisement - Continue Reading Below

You can go to a junior rink and in any given night, watch a dozen guys that you could see being in the NHL. “He’s so fast” “he’s so big” “he’s so talented” but realistically if you’re watching a junior hockey game, just one of them will be an effective player. The gap between success and failure is so slim and in a salary cap league with a limited number of contracts and roster spots, teams need to weigh a lot more information than in the past when making choices.

Do I think that the Leafs weighed those things? No. There is a distinct lack of objective data that would show Tyler Bozak is a better bet than Mikhail Grabovski. I’ve covered that hundreds of times on this blog. Just for the hell of it, here are two more charts. Over the last three years, I looked at the three players that had at least 200 minutes with both Tyler Bozak and Mikhail Grabovski, and compared their goals and points numbers over 60 minutes of even strength time:

Simple, right? It’s not like this is fairly advanced. You are taking the raw points number (you can find them on Hockey Analysis’ player pages, for Grabovski here and Bozak here), dividing it by minutes played with the player, and multiplying by 60. There is nothing advanced about it, and yet Steve Simmons calls Grabovski the “lone wolf” and calls these metrics painfully advanced, because a) it is unlikely he can add 3 and 4 together and b) after being on Brian Burke’s bad side, Simmons wants to suck up to the new regime and break more stories.

As an outsider, I have no reason to feel like I’m on the inside or break stories. I watch the Leafs because I like watching the Leafs and I like maintaining this blog, and when the Leafs are worse, then that just makes my nights worse. I’m quite sure that the Maple Leafs do not use resources like Hockey Analysis to make decisions.

Why? Because Dave Nonis said this, via Jesse Spector:

“If you go back and look at (Bolland’s) junior numbers, they’re exceptional,” Nonis told reporters at the NHL draft, referencing the 299 points Bolland scored during 254 games with the London Knights of the Ontario Hockey League. “He played behind some pretty good people (in Chicago) and I think Randy won’t pigeon-hole him as a third-line center. I think he’ll probably be put in more of a prominent role with us than he was in Chicago.”

Emphasis mine. What’s wrong with that? Well, it’s patently false.

Joel Quenneville did not pigeon-hole him this year. Bolland played all but 33 minutes of his 5-on-5 time with Patrick Kane, who had a 45.0 percent Corsi rate with him, and 56.3 percent without, while Bolland’s second-most frequent linemate, Patrick Sharp, had a 41.9 percent Corsi rate while skating with Bolland, compared to 59.8 percent without.Bolland’s overall Corsi rate was 44.7 percent.

Okay, maybe those numbers are a little more advanced, but it’s worth taking into consideration that Nonis went on record saying that he’ll have a prominent role in Toronto, while completely ignoring the established fact that David Bolland’s two most prominent linemates in Chicago this year were Patrick Kane and Patrick Sharp! That is the key here, not the Corsi, or any other metric. I don’t care about Corsi when looking at Dave Bolland, because for two-and-a-half seasons, he was the team’s primary option in the defensive zone and should be expected to have a low Corsi rate. This year though, he lost his full-time job to Marcus Kruger, and then Chicago went and traded for Michal Handzus to replace him in the lineup. Michal Hanzus is a formerly good two-way centreman that worked his way onto a Chicago roster that was starved for centremen.

Nonis didn’t seem to realize that. I would bet money that Nonis spent more time on Dave Bolland’s HockeyDB page checking to see where he was born than he did tracking video to see how David Bolland was deployed in Chicago’s lineup. It’s not rocket science, and nothing is advanced.

The good news is that the Bolland bet can a) pay off, and the Leafs will have a good defensive centreman this year or b) not pay off, and he’s gone this year. It’s risk-free. Long-term deals to Tyler Bozak and David Clarkson are not, and buying out your best two-way centreman to make room for one of them is catastrophically stupid.

Actually, the move itself isn’t catastrophically stupid. It’s “understandable” for a new general manager who saw his superior get fired by new ownership and figures he needs to lead the team in a new direction. It is catastrophically stupid that the Leafs don’t seem to be consulting any outside data that would have given them the slightest hint that perhaps they were betting on the wrong space on the roulette board.

To quote a line from 21 I guess, the Leafs aren’t card-counting. They’re gambling. They’re hitting on 17 based on a gut feel. If the bet works out, it is not with any information we currently have available, it would just be based on blind luck.

This site uses Akismet to reduce spam. Learn how your comment data is processed.

  • Get over it. This had little to do with Grabovski’s play or his contract. This was clearly a move made because the player and the coach did not get along… all. That’s why he’s not a Maple Leaf anymore. Is Bozak a downgrade? sure. But if you want to blame anybody, blame the coach. What is Nonis supposed to do? Fire the coach because one player is throwing a hissy fit?

      • jasken

        Right because Nonis will fire the coach who brought them to the playoffs for the first time in 8 years AFTER the second best choice(Eakins) has been hired by another team.

        • Nonis doesn’t have any problems replacing the goalies that led the Leafs to the Playoffs for the first time in 8 years.

          Reimer/Scrivens had a team save% of 0.917. That’s 7th best in the League. They are the reasons why the Leafs made it to the playoffs. Yet one got shipped out, and the other is being replaced by a goalie who’s played 62 career games.

          • jasken

            Who said he’s being replaced? Last I checked he’s still on the team and Bernier hasn’t proven a thing. It’s up to Reimer if he wants to be replaced. I’d be more concerned with both of them not living up to expectations than Reimer’s feelings.

        • jasken

          He doesn’t have any problem buying out a player *one year* after he signed a $27 million contract.

          And thinking that the team isn’t trying to replace Reimer with Bernier is nuts. He’s being paid twice as much money. He’s not being paid twice as much money to be the backup goalie.

          • jasken

            He didn’t have a problem without buying Grabovski, because he never wanted him. He tried to get Burke to make a deal last year, to trade Grabovski for a pick Burke refused and signed him not Nonis.

            Bernier is being paid more because Reims is now in his contract year, what he does this year will determine his contract. Bernier was RFA this year and need to be signed.

  • jasken

    It’s not even about advanced statistics anymore: it’s about the Maple Leafs demonstrating utter contempt for information and being mindfully ignorant over their player moves. Concepts like Corsi aren’t all that difficult to grasp; what’s baffling to me is how certain media members (see, for instance, that noted intellectual Steve Simmons) handle them like quantum physics. I wish I rooted for a team that was on the cutting edge of player evaluation, not one that’s deliberately retrograde and that thinks faceoff wins are more important than positive possession stats.

  • That was my one problem with Joyce’s Future Greats and Heartbreaks (which I generally liked) – the fidelity to and even romanticizing of convention and authority. Which isn’t to say I would have liked to read a book that simply crapped on the establishment for the sake of it, but there was a general lack of critical evaluation of the things he was seeing. And that’s despite the fact he spent a year writing about one of the worst run organizations in the league at the time.

  • I think your premise incorrect here.

    1. I don’t know anyone who claims Bozak is a better hockey player then Grabovski. Even Cherry doesn’t think Bozak is worth his contract. You are arguing with yourself here and preaching to the choir.

    2. Fillupa replaced Lecavalier who was bought out. Buy outs happen and no one is ridiculous to suggest Fillupa is better then Lecavalier nor the same with Grabbo.

    3. As a fan of advanced stats, I’m disturbed in the use (misuse) of goal data which is prone to luck effects/SH% and drawing questionable conclusion. Using Corsi we see that Bozak/kessel produced at a 47.3% rate whereas Grabbo/kessel produced at a 34.5% when on ice together in the 2012 season. Both are sub par and to be fair and objective here (which I find lacking by many of those who apply advanced stats), I will note that Grabbo/kessel only played 23 minutes and so small samples will be a problem. I don’t believe in the least that Bozak is “better” then Grabbo as Corsi suggest, but I question the methodology of using goal. Maybe Nonis did look at Corsi (not goal data) and ditched Grabbo for his poor play with Kessel?

    4. Bolland replaces Grabbo as a 3C checker. Kadri/Bozak roles are more consistent year to year. What would work is is Bolland vs Grabbo as 3C rather then trying to compare Bozak/Grabbo who are used differently. And if you must compare them, then use their respective Corsi with Bozak’s as well rather then mixing goals/Corsi where it better fits a specific thesis.

    5. Some of the statement appear presumptuous. For example, how do you know that Nonis does not look at Corsi or other advanced stats? What factors outside of Corsi or goals did Nonis consider? How do you know how much time Nonis spends on hockeydb? You seem to imply Nonis is a buffoon and if so I don’t like this but I really can’t be sure.

  • jasken

    Nice post Cam, you show how offensively talented Grabovski is over Bozak. Which everyone already knew. Considering you are comparing an offensive center to a 2-way center the numbers of an offensive center will always be higher.

    Anyone can score and put up points, its defense that wins hockey games. If you rack up 4-5 goals a night, have no defensive play to maintain the lead, your going to lose.

    Grabovski playing for Wilson was to get goals and pts and that was it. Wilson = good coach with a 3.16 g.a.a

    With Carlyle get goals and pts more importantly prevent them and win games. Carlyle = stupid coach with 2.67 g.a.a.

    Making some myth on mismanaging a player and yes it is a myth. Grabovski was on pace to score 26 goals, his avg along with his shooting %. Where is the mismanage there, the fact his +/- changed, he was on 3rd line so he gets less TOI, so what by moving him to 3rd line leafs dropped .50 g.a.a. that was a drop of 41 g for 82 games over the previous year.

    The leafs if they maintained the pace they were doing would have had approximately 9 more wins then last season if they kept the pace. Its not just about how much points 1 player gets but what is best for team as a whole.

    Carlyle was playing Grabovski the way it would give the Leafs a chance to win every night. You dont put all your apples in 1 basket if you lose that basket you have none. If you balance it over multiple baskets even if you lose 1 still got apples. This is Carlyle’s logic you, and Wilson forget the rest of the baskets want the 1 good 1 your logic is flawed.

    • jasken

      Good point. If Cam wants to look at goal based stats, we need to also consider goals against as a metric for defence. Looking at goal against and +-, we see Grabbo was a -10 while Bozak was a -1. And given Reimer was playing lights out, this suggest that Grabbo was a defensive sink hole there.

      Also I don’t understand why Cam doesn’t show Mclemment. Mclemment who makes 1.5M/yr and who was a 3C/4C and on the same goal and point pace of Grabbo. Mcclements had harder minutes with tougher zone start and similar QOC but Grabbo still could not outproduce Mclemment. That data does not put Grabbo in a a positive light.

      Grabbo’s corsi is better though again Mclemment got tougher zone starts so it is understandable. ANd since Cam loves goal based stats, Mclemment managed to produce at the same rate as Grabbo given his usage.

    • jasken

      Not great for the year 2.99 g.a.a. Yes it was good with Reims 2.60 g.a.a in 2010-2011 playing in front of the other goalies it was horrendous there was no consistency in their play.

      Gustavsson 3.29 g.a.a.
      Giguere 2.87 g.a.a

      This year better but still in my opinion not great. They need to work on either developing defencemen or get temporary solid defence until they are ready. Consistency is a must and the defence core dont have it. Not acquiring a defenceman at all, their hoping Bernier/Reims are gonna stand on their heads.

      Wilson plan was to use corvettes and Burke gets him camaros. Wilson was never in the race he didn’t have what he needed to be.

      Carlyle dont seem to have same problem with Nonis. He asks for camaros that is what he gets. Grabovski was a corvette in Carlyle’s camaro race so he tried to pretend he was a camaro, reality sunk in.

  • Not to mention, over the same span of games I mentioned, Bozak’s goals against per 20 was 1.020. Grabovski’s was .822.

    Leafs forwards over the three year span that are plus players?

    Clarke MacArthur
    Mikhail Grabovski
    Nikolai Kulemin

    You don’t want to go down that road.

  • Not to mention, over the same span of games I mentioned, Bozak’s goals against per 20 was 1.020. Grabovski’s was .822.

    Leafs forwards over the three year span that are plus players?

    Clarke MacArthur
    Mikhail Grabovski
    Nikolai Kulemin

    You don’t want to go down that road.

  • jasken

    Haven’t we beaten this topic to death?

    Let’s move on and talk about other things… like whether or not Morgan Rielly makes the big club or goes back to Moose Jaw – Marlies not an option.

  • Regardless of which player is better or fits better in the coaches system, buying out Grabbo is terrible asset management. He has .59 pts/game in the NHL, good possession numbers and plays Center, a position the Leafs are very weak at organizationally.

    Not only did the Leafs get nothing back in return for him, they used up their last compliance buyout on him, essentially removing their safety net and any chance at exploiting their wealth to profit from getting rid of other teams bad contracts.

    The Leafs then used up all the cap space they saved on Grabbo’s buyout without addressing what everyone, including those praising their signings, had identified as their biggest needs – a #1 center and top 2 D.

    I’m probably focusing on the worst case scenario because I’m pissed off, and as a longtime Leaf fan that’s what I’ve gotten used to. At best though, this is terrible asset management.

  • MaxPower417

    Great article. I’ve had nothing but arguments with people standing behind this decision and I still have no idea why. My favourite quote so far from one such person was that they liked it but not for any tangible or intangible reason. Spiritual reasons then?

  • jasken

    My worry: Kulemin decides he’d rather play for a team with at least one player who speaks Russian next year.

    Also, beyond just who’s better, we’re paying Grabovski, what is it, $0.9M/year to NOT PLAY FOR US for 8 years? If he gets signed for $4.5M, it is pretty obviously a bad move. Interested to see where he ends up and how he performs.

  • Killawatt

    It’s now day 5 since the start of UFA season. According to the TSN UFA tracker, 23 centerman have been signed to contracts.

    Matt Cullen has a job. Saku Koivu has a job. Gordon Boyd is making himself $3M a year to be a center in Edmonton (Gordon who?).

    Why hasn’t Grabovski been signed?

    Are there any theories out there? Are all the GM’s who signed those 23 centerman as dopey as Nonis, or is there perhaps more to this player than his stats, advanced or otherwise?

    What the hell is going on?

    • jasken

      I’ve been wondering the same thing. Some ideas..

      – He’s actually been a problem in Toronto (has anyone heard of this? I hadn’t) so teams don’t want him.
      – His outburst against Carlyle scared teams off.
      – His bad reputation from his Montreal time is still sticking with him (not helped by [expletive] outburst).
      – He’s not happy with the offers he’s getting (since getting less than ~$5M would imply that buying him out was the right thing to do).
      – Teams assume that since he was bought out, he’s not worth $5.5M/yr, but he still wants that kind of money.
      – He’s considering a move to the KHL.

  • jasken

    @mac and @killerwatt

    I said before dont be surprised, was called insane. His outburst towards the coach didn’t help his cause any. If teams that he really dont want to play for are contacting him, he might decide to play KHL. If he gets upset with Leafs for playing on third line how will he respond if other teams coaches put him there. You dont fire a coach if you dont like where he plays his players, you fire a coach for not doing his job “winning”.

    Most Speed, and high skilled teams already have their top 2 centers, him playing on a tough and big gritty teams wont work.

    Kulemin leaving because there are no other Russians? Kulemin is not that shallow. He might leave for being back on 3rd line though j/k.

    • jasken

      I don’t think it would be shallow for a player to decide on where to play based on who he would be playing with. People definitely like to be able to hang out with people who speak their native language. I was at a small party recently with a bunch of people from various Russian speaking countries. It pretty quickly split up into groups based on language. I can’t recall Kulemin ever giving an interview, no idea how comfortable he is in English.

      I forgot to mention that Grabovski also seemed to have some sort of health problem this season, which could turn teams off. Hadn’t heard about him getting married.

      also the possibility that he just wasn’t worth the 5.5 (or 4.5 if you factor in the buyout). I guess we’ll see how he does next year.

  • Back in Black

    Because the Leafs waited to the last minute to buy him out, other teams already had already negotiated deals with the other major UFAs. Since then he’s probably been busy with his wedding – and at this point, what’s the rush?

  • jasken

    1. 11 game goalless streak at the start of the year, after his 5 pts in 5 games when he was on second line.

    2. 10 games scoreless streaks after that.

    That’s 31 games without scoring a goal and 6 pts in those 31 games and wanting to be on top 2 lines.

    1/3 of Grabovski’s goals and pts were in the first 5 games of the season then nothing. He was avg over 17 mins ES a game,played 11 games getting 1 pt.

    No amount of arguing can justify 6 pts in 31 games of play and expect to play top 2 line. Bozak would play better then that on any line, same as McClement who did play mainly on 4th line.

    This is what happened, why he was bought out, why there are questions of risk on picking him up, why GMs dont rely on advance stats to get players.

  • jasken

    Random observation of the day: Lupul, Bozak and Kessel played together more than any other forwards in the playoffs and their fenwick numbers weren’t half bad.

    Bozak w/ Lupul 48.5 FF%
    Bozak w/ Kessel 46.9% FF%
    Lupul w/ Kessel 52.4% FF%

    Similar story with JVR/Grabovski/Kulemin

  • Something that’s been doing my head in: people claiming Grabovski’s incompatible with Carlyle’s system, then trying to argue that that’s somehow a validation of Randy Carlyle as a coach. Isn’t it the exact OPPOSITE? If Randy Carlyle’s a hockey guru, shouldn’t he be able to figure out how to use a valuable player who doesn’t necessarily fit with his vaunted “system”? Based on what we know, a lot of the clashes between player and coach might’ve been avoided if the coach had used the player the way he should’ve been used. Given all that…isn’t this an indictment of Randy Carlyle as a coach?