Is Kyle Dubas ‘Good’ at Drafting?

Using NHL Draft Bias Research to Evaluate the Recent Toronto Maple Leafs Draft Record

There’s a tone of smoke about the Toronto Maple Leafs management at the moment. Maybe we will look back and call this period overdramatic, but right now, it’s not good. Many people are questioning whether Kyle Dubas is the right man for Toronto’s General Manager (GM) position. There are a few criticisms, but that’s not the point of this post. Instead, we will look into the one feather in Dubas’s cap his supporters consistently point to.

The point in Dubas’s favour is that he is amazing at the draft table. And, given that drafting well is probably the most important component to sustained team success, that would be a massive point in favour of Dubas. On the other hand, if Dubas hasn’t been good at the draft table, that is a massive knock against his tenure.

To make matters worse, when evaluating a new GM like Dubas, the waters are still muddy on his draft record. I have liked many of his draft selections, but it’s not like their record is so full of All-stars that Dubas’s draft prowess is unquestionable. So, today let’s try to answer the question, is Kyle Dubas actually good at drafting?

Evaluating the Process

To see if Dubas is any good at drafting, we must investigate the process rather than the results. This is an evergreen way of looking at things, but it is especially important when looking at a new GM’s draft picks for two key reasons. The first key reason is that Dubas has only been the GM since 2018, and that is rarely enough time to properly evaluate a draft class. The second is those draft outcomes themselves are noisy and have small samples. So, if somebody was terrible at drafting but got two lucky hits, their record might look fantastic even over a half-decade.

To evaluate “process” I will use my NHL draft bias research from this summer. This research revealed that the NHL draft is exploitable. Specifically, an NHL GM could have gotten more value out of his picks if his draftees were on average:

-          Higher scoring

-          Lower penalty minute

-          Shorter

-          Relatively Young

- Over-age prospects

Since we know players like this tend to outperform their draft position; a good drafting GM will tend to select players like this. That way, they can get more out of whatever draft capital they do have. So, based on the models produced during this research, we will see if Dubas has been adding expected value to his draft picks by often selecting these kinds of players.

Also note, exploiting the draft is especially important for a GM like Dubas because contenders have limited draft capital. So, they can’t simply follow what everyone else is doing to sustain success; they will need to find edges to mitigate the long-term losses that contending teams tend to concede when making win-now moves.

Using my research from the summer, where do Dubas’s Leafs stack up in terms of value relative to expectations? It turns out they top the league on a per pick basis.

On average, Dubas has been adding an estimated 2+ GAR per draft pick. Of course, this will be noisy, but the Leafs draft process under Dubas appears to have been very, very strong. It’s tough to mentally know how much value Dubas should produce with his draft picks because he has only picked 22 times since 2018. But, in the picks he has used to select prospects, Dubas has been fantastic, on average. Note the market has fundamentally changed since I did my analysis using 2007-2015 (the league agrees with my thesis and has started changing its tendencies in that direction). However, Dubas and the Leafs are ahead of the curve, along with teams like Carolina, Seattle and Minnesota. This is especially encouraging to see Toronto’s company include the analytical powerhouse that is Tulsky’s Hurricanes and the Seattle Kraken, who have made investments in multiple well-known public analysts.

Edit: I have been informed by someone who knows better than me Darren Yorke oversees drafting in Carolina. Credit where credit is due to Mr. Yorke.

So, is Kyle Dubas #actuallygood at drafting? I think the answer is yes. This model estimates he has been the best in the league at exploiting market inefficiencies during his tenure. I would be cautious about suggesting he is definitely the best drafter in the league since there is more than one way to go about this analysis. Still, the evidence suggests to me drafting is a legitimate strength of Kyle Dubas.

Hunter Vs Dubas

Where Dubas stacks up around the league at exploiting draft market inefficiencies during his tenure is the most important way to evaluate him. However, Dubas has another important benchmark on Leafs Twitter, Mark Hunter. Hunter is a polarizing figure on Leafs Twitter. Some believe he was fantastic; others believe his draft reputation is incredibly overrated. We can evaluate the Leafs process during the Hunter era with the same data and model. Where did the Hunter-era Leafs stack up against the league? Also, very strong.

Turns out the Hunter Era Leafs were the second-best team in the league by this measure at exploiting the draft market, on average. However, we need to go further because averages can obfuscate trends. We can learn more by looking at the distribution.

So, on average, the Leafs have selected better players than what was expected of them since 2014 or so. This makes sense. Their rapid turnaround in 2016 would not have been possible otherwise. But the distribution of their process has changed most during the Dubas era. The Hunter-era Leafs selected some fantastic players with very high-leverage picks. My model has Matthews as the best prospect of this generation (he’s the massive positive outlier in 2016), with Marner not far behind. These two massively pull up the average expected value of the Hunter-era Leafs. That era also had some selections my model hated, like Fedor Gordeev, Eemeli Räsänen and other hulking defencemen who couldn’t score they selected.

So, were the Hunter-era Leafs good at drafting? They almost certainly weren’t as strong as the Dubas-era Leafs at maximizing what picks they had. How much worse the Hunter-era Leafs were at drafting basically boils down to how much credit you give them for selecting Marner and Matthews. If you assume either or both of those picks were tap-ins, much of that era’s draft prowess disappears. Even if we assume Matthews was a tap-in, they are league-average. Take out Marner too, and they sink to the bottom 10.

On the other hand, Dubas lacks any massive wins but has been fantastic at avoiding selections who are unlikely to live up to their draft position. Using this model, his best picks relative to where they were selected include Durzi, Robertson, Abruzzese and Hirvonen. On the contrary, his only truly negative pick has been Filip Král, which is impressive for a five-year sample.

Dubas is likely going to have to continue to win drafts this way because not only is the league better than it used to be at drafting, he will likely never have a bad enough roster to select an obviously extremely high-value player like Matthews or Marner as long as the two of them are on the NHL roster.

Conclusion

So, is Kyle Dubas good at drafting? The signs point to yes. In fact, using one methodology of estimating which GMs best maximize the expected value of their draft selections on average, Kyle Dubas has been the best drafter in the NHL since he became GM in 2018. There is noise around this number, but we can still be confident in the draft process of the Dubas-era Leafs. They have yet to have an absolute home run draft selection, but they have done incredibly well at avoiding obvious mistakes when drafting. Instead, they seem to consistently make relatively good but not out-of-this-world selections