We’re just one day away from the event of the summer, the great meeting of hockey minds to decide the future of NHL teams across the continent — the total eclipse of our hearts, as hockey fans, as we watch and hope that the right 17 (or 18 or 19)-year-old kid takes the stage next to GM whomever, smiling in the spotlight, basking in the glow of thousands of adoring season ticket holders.
And no draft season would be complete without hundreds of rankings and thousands of Twitter pundits’ 280-character talent appraisals. The rise of video platforms, junior-league stats websites and growth of social media have made it easier than ever for amateur scouts to pore over hours of tape and numbers and disseminate their opinions of who should be drafted when and why.
That the NHL draft community has grown so much in recent years is a boon to anyone attempting to analyze the incoming group of teenagers and decide who they’d like their team to pick. Because we now have dozens of player rankings and models comparing current would-be draftees to historical players, we are privy to a sort of wisdom of the masses we haven’t had before.
Adjusted Scoring Rates Help Narrow Down Potential Picks
One statistically proficient writer at Canucks Army, Jeremy Davis, has created a draft model that compares current players’ production to historical cohorts. His projection system uses situation, era, age, and league- (SEAL) adjusted scoring as a way to standardize point-per-game rates across various leagues and player ages for all draft eligible players. This year, Davis has compiled a consolidated ranking of the top 31 players and has run the top-100 prospects through his algorithms. All of this, along with Will Scouching’s “Involvement rate (INV%)” — the percentage of team points for which a player tallies a primary point — are available here.
Here, according to Jeremy’s consolidated rankings, are the players who should be available to the Sharks at pick 21 this June. Some of the individual experts’ rankings have changed since he published this, but as a proxy for who will be around toward the end of the first round, this ranking should suffice.
Would we prefer it if someone ranked higher, like Jesperi Kotkaniemi, were to fall? Would it be nice if Doug Wilson traded down to amass more picks? Sure. Rather than try explore all hypotheticals we’ll limit ourselves by looking at who might be available if teams pick according to collective wisdom.
What I’ve done is chart each of these players’ SEAL scoring rates against those of the top-100 prospects’ scores. The scores are graphed along a normal distribution curve, to see how each of these guys’ adjusted scoring rates compare to this year’s percentiles.
If the Sharks pick a forward with in the first round, here are the guys who are considered first-round skaters who should be available to them. What we’ve effectively done here is combine the eye test (consolidated industry ranks) with statistics (SEAL-adjusted scoring) to provide a more holistic view of who these players are.
Serron Noel and Martin Kaut do not appear on the chart, but you can see their scores alongside their names in the key and imagine their lines for yourself. Immediately, two players stick out. Akil Thomas and Martin Kaut, both with SEAL-adjusted scoring rates of 1.55 points per game. One standard deviation above average for all 64 forwards in the top-100 is a SEAL-adjusted scoring rate of 1.46 points per game. Thomas and Kaut are the only two who fall above one deviation by this metric.
If we want to differentiate between the two, we can look at other metrics. Kaut recorded a primary point on 15.7 percent of his team’s 5v5 goals this season. Thomas, on 20.7 percent of his team’s 5v5 goals. However — however! Kaut only played about an estimated 11 minutes of 5v5 hockey per game this season, where as Thomas loaded up on nearly 16 minutes of skating time per contest. It’s more likely that Thomas would be involved in a higher percentage of his team’s goals with that disparity.
Luckily, folks, there’s more. Davis’ model uses player statistics to look at two other, among a host, of things:
- Players’ expected likelihood of success “based on how many similar players reached a 100 NHL game threshold while under team control (typically seven seasons after being eligible for the draft).”
- Expected value, which is “designed to approximate Points Shares per 82 games, weighted by likelihood of success.”
Basically, the player’s likelihood of success represents how “safe” a pick he is, or how likely it is he will become an NHL player. A guy’s expected value is a little more representative (at least how I understand it) of someone’s eventual NHL ceiling.
Akil Thomas’ expected likelihood of success is 42 percent, by Davis’ model. Kaut comes in at 31 percent. The model also gives Thomas a slight edge in expected value. Despite their similar adjusted scoring rates, it appears Thomas is more likely to not only reach the NHL, but also to realize a higher ceiling. It’s close, however, and either player would likely make a solid pick on draft day.
Forward Pick: Akil Thomas, center, Niagara IceDogs
When drafting NHL defensemen, there is evidence to suggest draftees who score at higher rates have a higher likelihood of achieving NHL success. And, while scouts do a good job of figuring out what non-scoring traits make defensemen good prospects, they often overvalue those traits relative to prospects’ scoring rates.
While scoring rates for defensemen are only part of the picture, it seems we can use scoring to help separate players ranked similarly by scouts.
Here we can see Ryan Merkley well ahead of the rest of this group. He comes with off-ice concerns, and some scouts believe he is a bit deficient in his own end, but his scoring ability seems pretty undeniable here. Rasmus Sandin is also a fairly exciting prospect, if Merkley goes earlier in the draft.
Davis’ model believes Merkely offers both a higher ceiling and likelihood of making the NHL (playing 100 games). However, both players clocked in at around 20 estimated minutes played per game. Both players’ teams recorded a worse goal differential with them on the ice than without. This is a number that some in the scouting community cite as evidence of defensive deficiencies.
Finally, Merkley recorded a primary point on 14.7 percent of his team’s 5v5 goals, vs. just 9 percent from Sandin. The message there is that while both players have work to do in their own end, Merkley is more gifted offensively. Scoring rates, if we remember, are helpful in predicting NHL success. Assuming both players can work on their defense, it would be prudent for the Sharks to pick the player who seems to contribute more to his team’s goal scoring (Bonus: Merkley boasts the coveted right shot that that is so rare in NHL defensemen these days).
Defense Pick: Ryan Merkley, right-handed defenseman, Guelph Storm, OHL
Scoring rates only tell part of the story, but shouldn’t be ignored
Scoring rates aren’t everything when it comes to prospect evaluation, but they are a benchmark that teams shouldn’t ignore completely. Thanks to player cohort success models, like that of Jeremy Davis, we can combine adjusted scoring rates with other statistical measures of player effectiveness to gain a more complete understanding of a player. Combine that quantitative analysis with qualitative analysis (player rankings by scouts/experts who watch game tape), and we can go into June’s draft with a good idea of what a prospect might bring to an NHL organization’s system.