
17 Jul
2021 NHL Draft Preview
There has been plenty of talk around the 2021 NHL draft. There is a narrative that it is a “weaker” draft. I’m not sure I 100% buy that. There certainly is not a guaranteed star player that will go first overall. It appears people have become accustomed to that being the case with the likes of Connor McDavid (2015), Auston Matthews (2016), Rasmus Dahlin (2018), Jack Hughes (2019), and Alexis Lafreniere (2020) being recent first overall picks. The only year missing from the past six draft is 2017 that saw Nico Hischier go first overall. He didn’t have the same hype around him but has developed into a quality NHL player whose biggest issues in the league have been injuries. To me, the 2021 NHL draft is very similar to the 2014 draft. We’ll get back to that later. The biggest concern I have around the draft is the amount of data that we have on the prospects given the shortened or cancelled seasons around the world due to the COVID-19 pandemic, which will continue until people get the free, safe, and effective vaccine.
Process
I have previously built and back tested a logistic regression model that takes publicly available stats from player performance across leagues throughout the world to determine probabilities of a player playing at least 200 NHL games. I factor in performances during their two years prior to them being draft and adjust their probabilities based on age. Basically, the better a player is at a younger age, the better his chance of success. This is why the model has produced Luke Hughes as the top prospect. There are a couple of cutoffs for data to be included. First, no tournament play is included. It represents far too small of a sample size. Second, a player would have had to play at least 10 games with a team to have those results included.
The model has a major advantage and major disadvantage when it comes to this year’s (and probably next year’s) drafts. The major advantage is I take into account two seasons so players who were good during their mostly complete 2019 seasons get credit for doing so. The major disadvantage is the model is built using publicly available counting statistics and those statistics logically should not be linearly scaled. Therefore, players who were able to play full (or close to full) seasons may have inflated probabilities relative to other prospects in this draft class (Matt Coronato). Conversely, players who played minimal games might have deflated probabilities.
Let’s also get a couple of other things out of the way. First, there are plenty of other quality and some are probably more robust models out there. I am not going to say my is the best and all the others should be ignored. Each one probably has their pluses and minuses. Just because a prospect has a high or low probability it doesn’t mean they are destined for NHL greatness or guaranteed to be a bust. It simply means that past players with similar production profiles played or didn’t play 200 NHL games. Also, I have used tiers to break apart the prospects. Just because one has a 69% chance of success and another a 63% chance, I think either would be a fine choice for a team. They should be picked in a similar spot in the draft and go before players with a far lower probability (say 40%). Finally, I watch a lot of NHL hockey. I do not watch many prospects. My recommendations are simply going to be based off the numbers. In an ideal world, the model results are combined with traditional scouting to better understand and rank players.
Results
As you see, the model’s results are not as strong as many probably would expect for the three University of Michigan players. Although their games played is lower than a traditional college hockey season, many other players saw themselves playing significantly less games than usual. As such, Power, Johnson, and Beniers might have slightly deflated probabilities but based on the probabilities output, it is my opinion that these players might be receiving a little more hype than they should. Again, that by no means translates to them being busts. Conversely, you see a player like Matt Coronato having an extremely high probability. I don’t necessarily believe that means he should be a top five pick, but he is someone who probably should be taken sooner than many people seem to be predicting.

In my opinion, the top three picks should be some order of Luke Hughes, Owen Power, and William Eklund. If I were drafting for each team, I would probably take William Eklund for Buffalo, Luke Hughes for Seattle, and Owen Power for Anaheim. I expect the reality to be Power to Buffalo, Beniers to Seattle, and Anaheim to reach for one of Guenther or McTavish though Eklund or Hughes should be the pick. Leaving Hughes would be ideal for New Jersey as they would be able to get a tremendous fit with an added bonus of pairing him with his brother. Although Columbus seems to love picking off the traditional board, I can’t see how they could justify passing on Eklund if he’s there for them at five.
Comparison to 2014 Draft
I mentioned in my introduction that I see this draft as very comparable to the 2014 draft. In that draft, Ekblad had the highest probability of success (~70%), which is probably not far off of where Power would have been had he played a usual NCAA schedule. There are also a lot of forwards that are fairly muddled at the top with no clear star standing out. If I was going to compare players, I would say Power is close to Ekblad, Beniers is close to Sam Bennett, Eklund should be Leon Draisaitl or Sam Reinhart but probably ends up William Nylander. I also like the overall draft comparison because there are no absolute stars at the top but a fair amount of depth in the first round but drops off fairly quickly after that.
Conclusion
There is even more unknown about this draft than many others because of the impact of the pandemic on prospects’ 2020-2021 seasons with some barely playing a game. However, teams who make smart, data-driven decisions should fair just fine. There have been reports about the Sabres strongly valuing assets in this draft. Even though past history does not afford them the benefit of doubt, I personally, think there is value to be had, especially through the first round. Does that mean they should overvalue picks this year, absolutely not but they also shouldn’t undervalue them either, which seems to be the league wide approach. Seattle may end up in the same boat. If they could accumulate a similar amount of first round picks as Vegas, they too should have a very quality prospect pool to start their franchise with.
All data was collected from Elite Prospects.com
.KYLE STICH is the Director of AFP Analytics. In addition, Mr. Stich is a tax specialist and Director of Operations at AFP Consulting LLC, whose clientele include professional athletes performing services on three separate continents. Mr. Stich earned his Master of Science in Sport Management with a Concentration in Sport Analytics from Columbia University in 2017. He earned his undergraduate degrees in Accounting and Sport Management from St. John Fisher College in 2015, where he has served as an adjunct professor teaching Sport Finance and Baseball Analytics.