clock menu more-arrow no yes

Filed under:

Mark Barberio and super simple player evaluation

New, comments

The Montreal native makes an interesting case of using stats to define a player's role

Jasen Vinlove-USA TODAY Sports

In the midst of an underwhelming season, Mark Barberio has been quite the success story for the Montreal Canadiens. Left unqualified as a restricted free agent by the Tampa Bay Lightning (an organization which, until very recently, has been great at evaluating defencemen), Barberio signed a one-year, two-way contract with the Habs and worked his way up to the NHL club after a terrific stint in St. John's (20 points in 26 games).

Scott Cullen wrote about Barberio in his most recent article on TSN.ca:

He’s been very effective in 13 games, with positive goal differentials in the midst of the Habs’ recent collapse. If Montreal considers shaking up their blue line, they could do so with an eye towards a more regular role for Barberio.

I tend to agree with Cullen's assessment, as Barberio has done very well in Montreal both from an eye-test and a statistical perspective. His strong skating ability allows him to jump into the rush and serve as an offensive option, and he's posted far superior possession numbers to former first-round pick Nathan Beaulieu, by all accounts the best under-25 defenceman in the organization.

All this to say that Barberio has been a wonderful find, but how can we reconcile the fact that a player picked up from the "trash heap" is out-performing highly touted blue chippers, and how can a team make the most of pleasant surprises such as Barberio to improve across its lineup?

Coaches and stats: The Minimum Effect Dose

A consistent theme when I talk about analytics with coaches is that they are mostly uninterested in the raw numbers, but intensely curious about the "so whats" that the stats suggest.

At the simplest level, hockey is a game of goal differentials. Things such as shot differential (Corsi/Fenwick), zone deployments, and transitional metrics (zone entries and exits) are only useful insofar as giving us an educated guess of which players drive goal differential — a markedly superior guess than traditional plus/minus, by the way.

Essentially, when you look at a player's track record through the lens of advanced statistics, you can come to the conclusion that there are only three types of players:

  1. Those who drive play all the time (and who makes most teammates better);
  2. Those who drive play with the right linemates, and;
  3. Those who do not drive play, and who tend to drag most teammates down.

Every NHL team's forward and defence corps is made up of a combination of the three types, and the Habs are no exception. P.K. Subban and Jeff Petry are good examples of Type 1 players; put almost anyone on their left side and you have yourself an above-average pairing. Beaulieu and Barberio are at least Type 2s, and more NHL experience may help them improve enough to anchor a top pair at some point. The recently acquired Victor Bartley and the much-maligned Alexei Emelin will probably remain Type 3s for the rest of their pro careers.

Generally speaking, it's not too important for a head coach to know his players' relative Corsi or zone start percentages. As long as he gives his Type 1s plenty of ice time (preferably away from each other), support them with compatible Type 2s, and shelter his Type 3s, he'll be able to put together the best, most balanced lineup given the personnel at his disposal.

A note on Andrei Markov and Tom Gilbert

You may have noticed that I did not mention Andrei Markov or Tom Gilbert in the previous section. Aging affects different players in different ways, and circumstances suggest that both Markov and Gilbert should be handled very carefully.

Since a picture tells a thousand words, I'll start their stories by comparing them to themselves at a different point in time.

Did Markov really go from being an average second-pair defenceman to being one of the best in the league in his mid-thirties, after major knee surgery? Or can we attribute the change to playing alongside Subban rather than Emelin?

Right now, Markov is playing in a role that he can thrive in: providing offence alongside Subban, who has the speed and hockey sense to help him with shot suppression. But this is only contingent on the fact that Petry and Barberio can keep the other two pairs above water. Playing in a slightly different lineup, and in a different role, the 37-year-old will be exposed.

lt would be mean-spirited to say that Tom Gilbert effectively retired from the NHL the moment he left the Florida Panthers — and Brian Campbell — but it would not be far from the truth.

Gilbert's case shows how a player's reputation usually lags behind his production. He was a strong yet under-appreciated Type-1 defenceman with the Edmonton Oilers, regressed toward Type 2 as he hit 30, and was propped up by an elite possession player for a year in Florida before being signed by Montreal based on his past accomplishments.

At this point, he could still be a useful third-pair guy on a few NHL teams, Montreal included, but it might be a better idea to give Greg Pateryn more ice time, so that we have a sample size to work with when trying to determine exactly what type of player Pateryn is.

***

Jack Han is the Video & Analytics Coordinator for the McGill Martlet Hockey team. He also writes occasionally about the NHL for Habs Eyes on the Prize. You can find him on Twitter or on the ice at McConnell Arena.