The fun part with analytics is when the results are really surprising, and go against the general opinion. Too weird would trigger a signal about your model of course but if those “debatable findings” happen once in a while and have a rational explanation, it is worth discussing them and understanding why.
The DRAFTe model is no different. Because it is based on historical precedents, like all models, and some players come from a unique path (like Cale Makar drafted from AJHL) or have a unique performance that pushes the boundaries of the model’s knowledge about potential.
Also because the model quantified tangible results, even if we tried to go beyond points production with the Win Shares metrics. Some prospects like Brad Lambert just had poor on-ice results this season, but scouts see the potential in the tools and skills of the player.
Ideally, an NHL team would also quantity those observations and mix it with the tangible results to create an overall evaluation score that captures everything we know about a prospect.
But let’s go over the names that fell hard in the DRAFTe model. We took the Elite Prospect ranking as a comparaison. If you did not yet check our their 1,150 pages Draft Guide, we urge you to go read that amazing work.
Pavel Mintyukov (OHL): EP: #7
Mintyukov offensive creation was real, but his -18 5v5 goals differential (though slightly positive relative to his team) limited the impact of his defensive Win Shares. Even with a late year birthdate, he ends at the 62nd percentile historically among OHL defensemen. And, historically, the OHL has had a strong top 30 percentile in terms of probability to reach the NHL. But from a 49% probability for players at the 70th percentile, it drops to 27% for players between the 60th and 69th percentile. With that 27% chance, Mintyukov DRAFTe just could not go high much.
Marco Kasper (SHL): EP: #12
Kasper is a physical forward with good skating and who knows where to go on the ice. However, his performance in 46 games of SHL only ranked him at the 34th percentile historically, giving him a 33% chance of making the NHL. Though he scored a point per game in the Swedish junior league, it only placed him at the 55th percentile historically there, with a 25% chance of making the NHL. So, on one hand there is the hype of a player built for the NHL game, but the on-ice results did not follow just yet.
Brad Lambert (Liiga): EP: #14
There’s no need to go back on all the debate surrounding Lambert. The key here is that his 10 points and -16 5v5 goals ended up in a negative Win Shares, meaning Lambert costed his team points (a fraction to be honest) over the season. He is not the first Liiga draftee in that situation but at the 18th percentile historically, Lambert only benefits from a 33% chance of making the NHL… and still a negative Win Shares to base his DRAFTe value on. Which is why he just ends up at the 1,255th spot… Points alone would rank him 110th.
Calle Odelius (Swedish U20): EP: #15
This is a tough one. Odelius is the second best defensive defenseman according to EP, with great neutral zone play and transition skills. The issue here is that Odelius was on the ice for 52% of goals for his team at 5v5. But his team had 68% of goals without him. That relative -16 just held him back for his Win Shares, only ranking him at the 21st percentile historically among Swedish junior defensemen, a range where just 6% of the players made the NHL. Odelius is a clear case where talent and on-ice results will be seen differently by scouts for sure.
Owen Pickering (WHL): EP: #16
EP describes Pickering as a “formidable (…) 6-foot-5 puck-moving, two-way defenceman with plus-level hands, passing skill, and sense.” But like Odelius, on-ice results did not follow as expected. If the offensive production was modest, his 5v5 goals differential ended at -28. Pickering was on the ice for 38% of goals for at 5v5 this season, while his team recorded 45% of goals for without him. His Win Shares puts him at the 24th percentile historically among WHL defensemen, with a low 13% chance of making the NHL.
Owen Beck (OHL): EP: #21
Beck is another “don’t look at his stats” type of profile according to the EP guide. 50 points in 68 games is not ridiculous but nothing to be crazy about, and his +19 at 5v5 is good but he was part of a good team. At the end, his Win Shares only puts him at the 31st percentile historically, with a 18% chance of making the NHL.
Danila Yurov (KHL): EP: #23
Yurov clearly suffered from how the KHL sees prospect development. If he was dressed for 21 games with Magnitogorsk this season, he only played 3 minutes on average… No wonder he doesn’t have a single point to show for, and a negative Win Shares. His games in the MHL, on the other hand, ranked at the 87th percentile historically, and if the balance of both leagues makes him the 184th best profile in the model, his MHL games alone would have put him at 37th. If we defined the limit at 10 games played, Yurov’s case makes us wonder if Time on Ice should have a minimum as well…
Nathan Gaucher (LHJMQ): EP #25
We explained why the Quebec league has had a rough past decade in terms of producing NHLers in a past article. Gaucher is already playing like he would in the NHL and that is his big strength, according to EP. But, his performance on the ice is only good for the 47th percentile historically among LHJMQ forwards. At that level, only 24% of drafted players made the NHL.
Mattias Hävelid (SHL): EP: #26
Hävelid kind has the same story as Yurov, though he played close to 10 minutes a night in SHL with Linköping. However, with 0 point and a -4 5v5 goal differential, it would simply be the worst performance by a drafted player in the SHL since 2008. His games in junior ranked him at the 65th percentile there and that performance alone would have given him the 85th best DRAFTe instead of the 138th here.
Lian Bichsel (SHL): EP: #28
Finally, a bit of 6’05 gigantic ultra competitive defenseman with the Swiss Lian Bichsel. He mostly played with the men in SHL, gathering just 3 points in 29 games. Moreover, he also only was on the ice for 36% of goals for at 5v5 (-6 in total), when his team was at 51% without him. The hard reality is that such a performance barely gave him a positive Win Shares over the season. At the 10th percentile historically among SHL defensemen, he just benefits from a 33% chance of making the NHL. His junior games actually drag him further down the rankings too, as he could not do better than his team average in terms of 5v5 goals differential there.
Again, NHL teams are fully aware of on-ice performances versus talent in their evaluation. And that is why players with a lower DRAFTe still had success in the past (Brady Tkachuk had the 32nd DRAFTe in 2018, Adam Boqvist was 49th, K’Andre Miller was 52nd)), as their skills went beyond tangible results on their draft year. It is still interesting to see what will happen with those guys a few years down the road.
Liked that post?
And consider supporting us with a coffee !
You can also follow me on Twitter pages here and NL ice Data here