Wright, Slafkovsky or Cooley? | What science says

Shane Wright and Juraj Slafkovsky?

Well seen. But these definitions also apply to two gifted individuals drafted one after the other in 2016: Pierre-Luc Dubois and Jesse Puljujärvi.

At the time, with the third overall pick, the Columbus Blue Jackets preferred the Quebecer to the Finn. This decision stunned observers, who saw in Puljujärvi the reincarnation of Teemu Selänne. Their darling had dominated the World Junior Championship, with 17 points, the title of top scorer, that of the most valuable player and the gold medal. Besides, he was already playing in an adult league. While Dubois, cut from the Canadian team, had fun in the junior against weaker opposition.

It looks a bit, a lot like the current debate on the Canadian’s first choice, doesn’t it?

Today, there’s no doubt the Jackets got it right. Dubois is down to 239 points. Puljujärvi, 98 points. How could so many observers have been so wrong?

The Liiga is a stronger league than the QMJHL. Nobody will dispute it. Except that analysts have overestimated the performance of Puljujärvi in ​​Finland, and underestimated that of Dubois in Cape Breton.

Researcher Katerina Wu, from the University of North Carolina, studied this specific case. His conclusions: “Puljujärvi had obtained 28 points in 50 games. It was the fifth total in the Liiga under 20. Dubois was playing in a junior league, but he still made 99 points in 62 games. According to the equivalence coefficients, Dubois was statistically ahead of Puljujärvi.

“Of course, this is not the only criterion for evaluating recruiters. Dubois’ height and his physical game certainly weighed in the decision as well. But presumably the Blue Jackets had a better idea of ​​how each player compared to another when they preferred Dubois over Puljujärvi. »

Pause.

Let’s step back three lines.

“According to the equivalence coefficients. »

What is that ?

It is the Rosetta stone of hockey. The algorithm that translates the production of a prospect, in a secondary league, into points in the NHL. The catch is that no one has yet found THE right recipe that guarantees the best selections for sure.

All teams work with the same basic ingredients. They first assign a strength to each league, based on the performance history of its players in the NHL. For the Liiga, for example, the index may be 44%. For the QMJHL, 11%. Then we multiply this index with the number of points per game of hope. That’s it, an equivalence coefficient.

Now, as with cabbage soup, the secret is in the seasoning. Analysts will raise the formula with the size of a player, his playing time, his production at equal forces, etc. They can also take into account characteristics specific to each league. Examples ? The age of the players. The number of parts. The format of the championship. The size of the rink. The quality of goalkeepers. The reliability of statistics. In short, it’s a difficult exercise, complicated by the fact that hockey players are always dependent on the gestures of their teammates – except for penalty shots.

You will have understood that these algorithms are imperfect. Still, they provide a good basis for discussion.

And what do they tell us about this season’s draft?

Many, many interesting things.

I have a big weakness for Byron Bader’s model. It calculates the odds of a prospect becoming an NHL star. A star, according to its criteria, is a forward who will average 0.7 points per game in the NHL, or a defenseman who will finish his career with 0.45 points per game.

According to his model, in 2016, four players had more than a 50% chance of becoming a star: Auston Matthews (99%), Clayton Keller (74%), Matthew Tkachuk (70%) and Alex DeBrincat (70%). Jesse Puljujärvi was at 38% and Dubois at 27%. On the other hand, Dubois had a slightly better chance of establishing himself in the NHL.

What does the model suggest for the 2022 cohort?

Six prospects have more than half the chance of becoming a star.

Nemec, Wright and Cooley are the three players most likely to start. And Slafkovsky? Bader’s model only gives him a 12% chance of becoming a star. It’s not much. Less than for Nick Suzuki (27%), Jesperi Kotkaniemi (37%) or Cole Caufield (66%). It must be said that the Slovak had little playing time at the start of the season.

Chace McCallum, who has done some work on the equivalence coefficients, also raised a red flag last week about Slafkovsky and fellow Liiga prospect Brad Lambert.

“The argument against statistical models is that we find it perfectly normal that Slafkovsky and Lambert didn’t produce much because they were playing in a professional league. That these guys faced such strong competition at a young age is a good sign. Except it’s wrong. Due to their low production [dans la Liiga], players like Slafkovsky and Lambert are unlikely to meet the expectations of the rank at which they are drafted. The quality of the league does not compensate [leur production]. »

In another model of equivalence, the DRAFTe of Thibaud Chatel, Simon Nemec and Joakim Kemell occupy the first two ranks. Logan Cooley is 8^eShane Wright is 10^e and Juraj Slafkovsky, 18^e.

My prediction?

Since their appointment, the new bosses of the Canadian hammer the importance of advanced statistics. In this context, a selection of Juraj Slafkovsky in the first row would be surprising. It would be the result of an irresistible crush on Jeff Gorton or Kent Hughes.

It’s possible.

But the most likely scenario for the Habs remains a selection of Shane Wright or Logan Cooley.