NHL Draft 2018: Late-round swings for San Jose

One need only look at the Sharks’ stock of draft picks this summer — a first and two each in rounds four through six — to get excited about the idea of late-round gems. Doug Wilson and his scouting staff have added Joe Pavelski, Justin Braun, Jason Demers, Joakim Ryan, and Kevin Labanc (among a handful of others) to the team’s system with picks in rounds four through seven. Rudolfs Balcers, Alexander Chmelevski, and Ivan Chekhovich might be adding their names to this list soon.

Yet, while we remember good players who were drafted late, we must remind ourselves that early-round players are most likely to find success in the NHL. That’s in part because the NHL’s Central Scouting does a pretty efficient job of figuring out who will be drafted each year. Still, guys like Chekhovich go sliding every year. With everyone’s final rankings in, we can look at some different numbers to see which players might be available to San Jose during day two of the draft, when they get six picks to spend on lottery tickets.

Evan Oppenheimer just introduced a metric he calls “betweenness.” What betweenness tries to determine is who is responsible for his team’s points. Oppenheimer takes the example of Jonathan Cheechoo and Joe Thornton. Cheechoo loaded up with 56 goals the year he won the Rocket Richard trophy, but most of those goals were assisted by Jumbo. Cheechoo’s production was heavily dependent on Thornton’s passes that season. Thornton, on the other hand, helped a wider network of teammates produce points and was not heavily dependent on any one player to rack up his own numbers. In this case (Oppenheimer assumes — he hasn’t tested this on NHLers yet), Thornton would have a high betweenness score, while Cheechoo would have a lower mark, despite both players adding quite a few points to their ledgers.

This idea is helpful in evaluating prospects. Did Player X load up on goals in his draft year because he played with a great teammate, or was he responsible for his own production (and his teammate’s)? Conversely, this metric might help us identify players whose scoring rates weren’t very high, but who were otherwise a catalyst for their teammates’ production. Betweenness scores range from 0 (Cheechoo) to 1 (Thornton). Oppenheimer adjusts each player’s rating for the number of games he played.

He also created adjusted scoring rates for each player based on age, league, and era. Finally, he scaled each player’s betweenness score and adjusted scoring rate and averaged them. I found that the 5v5 averaged rating gave us, at a glance, a solid representation of player skill for his draft year. For instance, the 95th percentile of players in that averaged column (from as long ago as 20 years) includes: Connor McDavid and Sidney Crosby at the top (their numbers blow everyone else’s out of the water), Dylan Strome, Tyler Seguin, Sean Monahan, Jordan Eberle, Mathew Barzal, the list goes on.

(A note for sharks fans, 95th percentile draft year scores also include Josh Norris, Ivan Chekhovich, and Timo Meier.)

There are a fair share of no names in that top five percent, too. This is far from a sure thing. But it looks like we may have found something that can help identify glossed over talent. Of the first-year eligible skaters in the CHL leagues and the USHL, the following skaters’ scores also fall at the 95th percentile and above:

• Filip Zadina
• Andrei Svechnikov
• Tyler Madden
• Matej Pekar
• Joel Farabee
• Oliver Wahlstrom
• Allan McShane
• Pavel Gogolev
• Jonathan Gruden

Svechnikov, Zadina, Farabee, and Wahlstrom are all first-round locks. We’re interested in the other five skaters here.

Below is a table that shows the other five skaters, their draft-year age (how old they are relative to September 15, 2018 — the draft cutoff date), their league, their estimated 5v5 primary points per 60 minutes (thanks to prospect-stats.com), their shots per game (all situations), and their NHLe, or NHL equivalency score estimate.

NHLe looks at how much scoring players have retained when moving from one league to the NHL. For instance, the average player going from the OHL one year to the NHL the following year has retained about 32% of their scoring. A player scoring 1 point per game in the OHL would be expected to score .32 points per game in the NHL the following year, which is about a 26-point pace over 82 games. No one has gone directly from the USHL to the NHL, so no equivalency score exists. Rob Vollman, author of Hockey Abstract and the person whose tweet contains the NHLe information, did mention that USHL players eventually reach about 45.5% of their USHL scoring once they get to the NHL.

prospect-stats.com, Rob Vollman, Evan Oppenheimer

You can see that the USHL guys’ numbers are probably blown up a bit because there is no translation factor for year over year. If we used the OHL translation factor for the USHL guys, Gruden’s NHLe score would simmer down a bit to 36 points (still impressive), Pekar would come in at 25 points, and Tyler Madden at 18.

For shots per game and estimated primary points per game, both OHL players’ numbers were among the top 10 of about 118 draft-year skaters who played at least 25 games this year. The USHL skaters’ numbers were a bit more dispersed, but the lowest any of them ranked was 18th of about 95 skaters with 25 games.

You can see that, with the exception of Gruden’s primary scoring, these statistical profiles are very similar to one another.

NHLe translation factors aren’t meant to serve as projections as standalone metrics. They offer a baseline comparison number and help us determine which NHL feeder league has tougher competition. NHLe scores might, incidentally, still offer us a glimpse into these players’ potential future.

Based on Byron’s look at NHLe translation factors, age and historical success, we can look at what might happen if the past is any predictor of the NHL careers.

Just about all of these guys are pretty close to that 20 - 29 NHLe before age 18 (relative to the draft date) category. Of that historical group, 56 percent have gone on to make the NHL. Fourteen percent have become stars, and 44 percent have failed to make the league. That’s about as boom-or-bust as you can get. Gruden might even be more representative of the group higher up, where 68 percent of guys have historically made the NHL and 24 percent have gone to become stars.

Neutral zone play might help us differentiate between two players

We can also take a peek behind the hood at how some of these guys generate their points. Mitch Brown tracked a bunch of OHL and USHL players this season. Not all five of these players were in his sample, but Gruden and McShane were.

Mitch Brown

These are smallish sample sizes — just nine games for each player — and, as Brown’s article states, these figures represent what the player accomplished this past season. They do not necessarily predict any sort of future performance. McShane is a bit more balanced, but Gruden seemed to be better at driving shot and scoring chance differential this past season.

Let’s the mix eye test with numbers

It would be unwise to select prospects on the basis of numbers alone or eye test alone. Thankfully, The Blue Bullet Report melds the two. By applying his formula based on draft value pick, position, scoring rate, and player involvement in team goals to an amalgamation of 17 scouts’ eye tests, Blue Bullet Brad developed his own ranking.

McShane and Gruden are given second-round grades in this format. Pekar and Madden look like third-round picks using this metric, and Gogolev doesn’t make the top 100 prospects as evaluated by this model. Finally, we can look at how each player is ranked by the various scouting services to understand where he might be picked.

Where might these guys be picked?

McShane is ranked 50th among North American skaters by Central Scouting. The other services and scouts that ranked him give him an average rank of 65th, an early third round mark. McShane dropped from 33 during midterms to 50, so either other player movement shifted him around or scouts soured on his second-half play.

Pekar is ranked just below McShane among North American skaters, at 55. Other services/scouts give him an average rank of 93, which would be the last pick of the third round. Pekar flew up the final draft board from his midterm spot at 87, surely one of the highest risers since the halfway point. That might boost him into contention sooner rather than later.

Gogolev falls to number 79 among North American skaters, and other services are believe he’s similarly ranked. If Gogolev’s draft position is in line with his rank averages, he’ll be taken around pick 109, which is the middle of the 4th round. Gogolev fell four spots from 75 to 79, an essentially unnoticeable slip.

Tyler madden comes in at 71 for his average scouting service score, despite being ranked at 39 among North American skaters by Central Scouting. The 71st pick falls early in the third round. He rose from 49 among North American skaters at midterms, signaling that perhaps scouts understand what the numbers (beyond just his points per game rate) are telling us.

Finally, Gruden is ranked 46 by Central Scouting among North American skaters. He finds his rank all but stagnant, moving up two spots from his 48 position during midterms. The other scouting service and scouts grades average out to about 75 overall, which is also early in the third round.

The folks at Canucks Army ranked three of these skaters among their top 100 prospects. McShane is 33rd, Gruden 38, and Madden 92. It sounds as though Madden is ranked so far down because his skating abilities are top-notch, but his production, particularly where it concerns his play away from his most common linemates, is not terribly exciting.

Despite his impressive scoring, Canucks Army’s player cohort success model only gives Gruden a 28 percent chance of making the NHL, though he does seem to offer some upside worthy of more than a mid-round pick. McShane’s prospects of success are similar to Gruden’s, which helps explain why he’s ranked so similarly.

What should San Jose do?

Unfortunately for San Jose, it appears there’s a decent chance these players won’t last past the third round and into the fourth, where the Sharks have picks. Of these five, Gruden’s profile seems the most promising, while Pekar and Gogolev seem to be the players most likely to slip into later rounds. If we remember where we started, though, there is a chance that one of these players possesses some star qualities, even if his scoring rates aren’t special. It will take a few years before we know the fate of any of these five skaters, but keep an eye on them come Saturday. You might be witnessing the next diamond in the rough in the making.