Predicting Future NHL Scoring Success with NHLe Thresholds & Related Factors

NHL Equivalency (NHLe) is a formula used by some in the hockey analytics community to normalize scoring rates in different prospect feeder leagues. The object of finding a similar “score” for players across different leagues is to help project future NHL scoring/performance. It’s a method developed
by Gabe Desjardins of behindthenet.ca a number of years ago and has been expanded upon by hockey
analytics pioneers like Rob Vollman and Kent Wilson. Here’s Gabe’s original piece to give you a context if you’re not fully versed in NHLe.

Previously, I looked at draft year NHLe and forwards drafted in the first round from 2005 to 2010. In that investigation, I found that of players who had scored a career 0.6 PPG or higher in the NHL (approximately 50 points or more a season) 22 of 32
in total had an NHLe of at least 34 in their draft year.  

In fact, of all the first round
forwards who had an NHLe of 34 or more in their draft year, only five hadn’t
scored at a rate of 0.6 PPG or higher in the NHL to that point, though all had already made the NHL. While
draft year NHLe provides certain insights, I was curious how you could project
future impactful point producers (0.6 PPG or higher) overall, beyond the 1st
round and beyond a player’s draft year equivalency.  

The following analysis provides
insights into the following questions:

  • Are there
    differences between players who score a high equivalency in their draft year
    compared to later on? 
  • Do elite scorers
    tend to hit certain NHLe thresholds (e.g., 30+ or 40+) more often and/or more
    frequently than average, replacement-level scorers and busts?
  • What impact does age have in hitting an equivalency threshold and future NHL success? 

METHODOLOGY

I
created a dataset of all forwards drafted
between 2004 and 2014 (sourced from Hockey DB). 
The data set includes four groups: impact point-producers (IPPs) – 0.6
career PPG or higher; average point-producers (APPs) – 0.4 – 0.59 career PPG;
replacement-level point-producers (RPP) – 0 – 0.39 career PPG; and BUSTs – players
that weren’t successful in making the NHL.
 

IPP, APP and RPP players all
had to have played at least 100 games while BUSTs were drafted between 2003 and
2008, in the first four rounds, and had played less than 50 games. 

I calculated each player’s overall PPG over the course
of his playing career and appended various relevant stats including: draft
year NHLe; last NHLe before turning pro; if they reached an equivalency of 30+,
40+ and 50+; how many times they reached each threshold and the age in which they reached
each for the first time. In total, my sample set included 380 players – 100 RPP players, 87 APP players, 69 IPP players and 124
BUSTs.

**Notes: To increase the sample of IPPs, I included all 0.6+ PPG from the 2003 draft as well.

There has been much equivalency work since Gabe Desjardins original NHLe work in 2004 and further work on NHLes has shown that more recent league equivalencies have changed from Desjardins’ original work. 

However, I used Desjardin’s original ratings for a few reasons – a) my original analysis on 1st rounders used the original league equivalencies and given that analysis is very related to this analysis I wanted to remain consistent between the two analyses; b) the majority of the players in the sample were drafted 5 to 10 years ago, when the equivalencies would have been closer to Desjardins work; c) I don’t use the NHLes as an indicator of how the player is going to score the year after transitioning to the NHL, rather using it in a general sense to predict future career scoring success and d) depending on how the data is manipulated (taking into account age and games played thresholds) the equivalencies for each league can change slightly or even quite dramatically.  

Given Gabe’s original work focused a great deal on young prospects, using his rankings as the basis for this analysis made the most sense.

DRAFT POSITION BREAKDOWN



Draft Position Breakdown

As
one would expect, a significant portion of the elite offensive NHL forwards,
drafted in the past ten years, were drafted in the 1st round, and
the majority of those were drafted Top 10. However, it’s important to note that approximately 30% of this cohort’s elite player were still found
in later rounds, but shared some common traits with players drafted very high. 

Below
are some commonalities that emerge within the four groups mentioned (IPPs, APPs,
RPPs and BUSTs).

NHLe THRESHOLDS & FUTURE SUCCESS


30.40,50 NHLe Thresholds

We
can see that IPPs are far more likely to record a 30+, 40+ and/or 50+ NHLe
season than APPs, RPPs and BUSTs.  Over 90% of players (62 of 69 players) that went on to be impact point
producers had recorded a 30+ NHLe at least once, before turning pro
.  

Additionally, IPPs were far more likely to register multiple 30+ and 40+
equivalencies than APPs, RPPs and BUSTs. IPPs registered multiple 30+
equivalencies 68% of the time compared to 52% for APPs and 35% for RPPs. Conversely,
IPPs register multiple 40+ equivalencies 33% of the time compared to 28% for
APPs and 23% for RPPs. BUSTs recorded a 30+ and 40+ equivalency very infrequently
and, of the ones that did, very few recorded multiple 30+ and 40+ equivalency
seasons. 

So to even make it to the NHL and produce at all, a player will very
likely record a 30+ equivalency (198 of the 256 players – 77% in total) in
their junior career.  IPPs, on the other
hand, almost without fail, will cross the 30+ NHLe equivalency threshold, often multiple times, and will also be far more likely than APPs, RPPS and BUSTs to
cross the much more challenging 40+ and/or 50+ equivalency level as well.

It’s
important to note that half of all future IPPs and nearly all
of future APPs and RPPs will reach the 30+ and 40+ equivalencies after they are
drafted or, at least, eligible to be drafted. This is likely where scouting remains vitally important; to recognize the players that will hit 30+ and 40+ thresholds in their draft +1 and draft +2
years (to continue to develop and improve).

How
do you recognize players before they cross the 30+ threshold quantitavely? For one, a future
scorer may not reach a 30+ equivalency in their draft year but they will likely
be close. Case in point, 78% of all IPPs and 62% of all APPs had an equivalency
of at least 20 in their draft year. Conversely, 40% of all RPPs and only 27% of all BUSTS had an equivalency
of at least 20 in their draft year. As
well, watching for younger players making large jumps in year over year
equivalencies in their 16, 17 and 18 years is a good indication that the player
will cross the 30+ or 40+ milestone soon after being drafted.


Equivalency Jumps

To
illustrate this point, we can look at the average jumps between draft year NHLe
and last pre-pro year NHLe of IPPs, APPs, RPPs and BUSTs that didn’t have a 30+
NHLe in their draft year.  

On average, an
IPP that didn’t register a 30+ equivalency in their draft year jumped up by 21
equivalency points compared to APPs who jumped up by 16 equivalency points,  RPPs jumped up by 15 equivalency points and BUSTs
only jumped up by 6.5 equivalency points. These are important trends to note when you are tracking prospect development; kids who make big jumps year over year are more likely to be NHLers. Kids who only make modest gains are less likely to make the show or be impact players. 

Also, on average, by the time an IPP, who hadn’t hit a 30+ equivalency
in their draft year or before, was set to turn pro (i.e., NHL or the minors),
they had an equivalency of 39 in their last pre-pro year compared to 35 for
APPs,  31 for RPPs and 17 for BUSTs. Therefore, IPPs that do not already have a
30+ equivalency by their draft year will tend to make the largest equivalency
jumps when they go back to junior, college or Euro leagues for their draft+
years.  

However, any player that makes
the NHL and plays over 100 games, elite scorers and depth players, will all
likely make large equivalency jumps in their junior career. Players that don’t make it will tend to make
little equivalency gains and they also are more likely than the other three
groups to fall back in equivalencies between their draft year and future
pre-pro years.

Next,
we’ll look at how age relates to hitting the 30+ and 40+ thresholds and what
age and 30+/40+ thresholds can tell us about future NHL scoring.

NHLe THRESHOLDS & AGE

We
all know that scoring very high point totals as a younger junior player is far
more predictive of future elite talent than scoring high point totals later in
junior. However, looking at the age in which IPPs, APPs and RPPs reach their
first 30+ and 40+ equivalencies thresholds, the age gaps are larger than I
would have expected. 

On average, an IPP reached his first 30+
equivalency a year younger than an APP and just over a year and a half ahead of
a RPP (18.5 vs. 19.7 vs. 20.2 years). Similarly, on average, IPPs reached a 40+ equivalency a year before APPs
and almost two years before a RPPs (19.2 vs. 20.0 vs. 20.9 years).  BUSTs rarely hit a 30+ equivalency or 40+
equivalency and when they do they tend to do so older than the other three groups (30+ –
20.6; 40+ – 21.02).


Thresholds and Age

Of the IPPs
that scored at least one 30+ equivalency, 90% did so before they turned 20 and 70%
did so before they turned 19. Compare this to APPs and RPPs where only 60% and 55%,
respectively, managed the feat before turning 20 (36% and 15%, respectively,
did so before the age of 19). Conversely,
of the BUSTs that did hit a 30+ equivalency, only 37% did it before turning 20
and only 14% did so before turning 19. 

As illustrated by the graphic,
we can see that finding players that hit the 30+ and 40+ equivalency earlier
than their peers is vitally important to finding future impactful NHL scorers. Hitting
a 30+ or 40+ equivalency before the age of 19 is a very good sign: 84% of players to hit a 30+ equivalency
before turning 19 and  89% of all players
to hit a 40+ equivalency before turning 20 went on to be an APP or IPP in the NHL.  

Additionally, 92% of all players to hit a 30+
equivalency before turning 19 and 98% of players to hit a 40+ equivalency
before turning 20 made the NHL and played at least 100 games.

BIRTH MONTH

To drill down a bit further on this topic we can look at birth month in a player’s draft year. 

Malcolm Gladwell, in his
book ‘Outliers’, noted a trend that most elite athletes in team sports that
have made it to or close to the highest level tend to be born in the first half
of the year rather than the second half because at a young age these athletes
are so much more developed than their peers. Gladwell explains that an age gap of several months means that the older
athletes will likely be more physically developed, taller and stronger than
their younger peers.  

Due to their
appeared dominance of the sport, the older kids are also given more attention,
training and encouragement.  As a result,
players born later in the year have a harder time reaching elite levels of team
sports because in their early years they’re further behind the older kids of
the same year. 

Looking
at all the players in the sample included here, this, at least partially,
appears to be the case. Of the 380 players in the sample, 59% were
born in the first six months, 41% in the last six months. However, if we look at the splits of IPP, APP,
RPP and BUST players, things completely switch.  

In terms of players born in the last six months of the year (July
through December), 35% of BUSTs players were born in the last six months compared to 38%
RPP players, 54% APP players and 65% IPP players. To me, this suggests
an IPP player who is born late in the year but is able to keep pace with their
older peers when moving their way through the elite youth hockey system all the
way to the highest level as an elite talent has been exceeding expectations for
10-15 years before being drafted. By the
time the developmental playing field levels off, in their late-teens, they’ve
already pushed themselves further physically, mentally and technically than
their older peers. 

This becomes very apparent looking at the age of
their first 30+ equivalency. IPP players
born between July and December, on average, reach their first 30+ equivalency
by age 18.5. IPP players born between January and June reach their first 30+
equivalency at 19.0. Interestingly, the
age gap between January to June APP players’ first 30+ equivalency and July to
December APP players’ first 30+ equivalency is negligible (19.9 vs. 19.8 years).  

Additionally, across all three groups, a player
born in the second half of the year is more than two times more likely to play
in the NHL directly after being drafted (8 early year players vs. 18 players
late year players), though granted it’s a very small sample.

CONCLUSION

Finding
players that can be consistent impact scorers in the NHL is no small feat. By looking at a player’s pre-pro
equivalencies we can glean numerous insights that can help increase a team’s
odds of finding a point producer or avoiding a bust.  

A
player almost certainly needs to have reached 30+ equivalency at some point if
he is expected to make the NHL and score at all, while the earlier and more frequently they’re able to the more likely the player will go on to produce significantly in the NHL.  he number of players in the past 11 years
that have not reached 30+ at some point and turned into a consistent impact
player is fewer than 10 – which means it’s not impossible, just extremely unlikely. Furthermore, the number of players in total that have
played at least 100 games in the past decade that did not have one single 30+
equivalency in their junior career is less than 60.

It
is possible that a player can do well in the NHL, from a scoring perspective,
without having recorded a 30+ equivalency (e.g., Blake Wheeler, Andrew Ladd,
Anze Kopitar, Patrice Bergeron, Loui Eriksson, Milan Michalek, Milan Lucic) but
the data would suggest that a player should be able to dominate their pre-pro
league, to the level of at least a 30+ equivalency, at some point before
turning pro. If they haven’t and are
expected to be a significant point producer, the odds are already against them,
in a big way. 

Equivalencies don’t tell the whole story of a player and scouting still remains vitally important in determining things like playing circumstances and character. However, NHLe or a similar league standardization systems should make up at least a part of the drafting decision and prospect projections. 

The player that emerges as a big time point-producer in the NHL, seemingly “figuring it out” after a subpar junior career is very rare and the data presented here shows that. Late round picks or undrafted players that do emerge like this will cross over the 30+ equivalency threshold. They may not do it as early as the obviously elite early 1st rounders but they will do it later on (e.g., Ondrej Palat, Pavel Datsyuk, Tyler Johnson, etc), often before they turn 20. 

Conversely, they may do it in their draft year, just like the elite players drafted in the top 10.  However, they might be skipped over because of their size, country of origin, attitude, etc (e.g., Marc Savard, Marty St. Louis, Theo Fleury, etc.). I’d put my late draft pick behind a 19 year old that crosses over the 30+ threshold in their draft-eligible +1 year over an 18 year old with an equivalency south of 10 that’s “on the verge of putting it all together” every time.  This seems self-explanatory but the latter choice is made over and over again, every year.

To conclude, I’ll leave you with a few cohorts pulled from the dataset that summarize the information presented and tell an interesting story …

  1. Of the sample population of players to record a 30+ NHLe in their draft year and were drafted in the 1st round and before the age of 19 … 91% made the NHL (playing 100 games), 85% turned into at least average point-producers. and 59% turned into impact point-producers.. If we up the threshold to 34+ while keeping the other parameters the same … 100% of the players made the NHL, 100% turned into at least average point-producers and 70% turned into impact point-producers.
  2. Of the sample population of players to record a 30+ equivalency before turning 20 (regardless of round drafted) … 91% made the NHL, 66% turned into at least average point-producers. and 37% turned into impact point-producers.
  3. Of the sample population of players to record their first 30+ NHLe after turning 20 … 70% still made the NHL but only 29% turned into at least average point-producers..
  4. Of the sample population of players to never record a 30+ NHLe … only 18% turned into at least average point-producers.

  • jandrewyang

    Side notes.

    Gladwell’s observation was specific to Canadian hockey players and noted that in other sports and countries that it would vary with the birth date eligibility requirements.

    Also, Gladwell has been shown to be more “entertainment” than science.

  • BurningSensation

    Nice piece Byron!

    Any thoughts on those undrafted players who met the NHLE 30 threshold as to who we might target for future camp invites?(Tyler Soy?)

  • OKG

    I think NHLe are cool, but my issue is that these junior leagues don’t measure time on ice.

    How can the production of a player who plays 13 minutes a game on a good team be compared apples to apples to a player playing 3 minute shifts and 20+ minutes a game?

    At the NHL level we have a lot of “per-60” stats that more accurately highlight what players’ production is. That does not exist at lower levels.

    I mean in Junior, sure a 20 year old is an overager and expected to dominate minutes as well as competition… but in SHL or NCAA? These are mens’ leagues where 20 year olds are still redshirting their freshman season or equivalent… there’s politics at play beyond giving young kids an opportunity.

    • clib542

      Did you even read the article? NHLe is calculated different for each league, so yes it takes into affect the difficulty of the league.

      There are also a probability provided above for making the NHL based on the age you get to 30 NHLe, so when the 20 year old is dominating like they should, you can see their history of making it is low if they have not done it before.

    • TX Flame

      Easy, good players with real NHL upside will be relied upon heavily by their coach. If a player plays 13min/g on a good team, why aren’t we drafting the guys playing more than 13minutes because they are clearly driving the bus?

      Also, Providence doesn’t have particularly great offense… so the minutes thing is moot. The best players should be awarded with the most ice time. No team in a lower league is deep enough to give a legitimate blue chip NHL prospect 3rd line minutes, except for maybe the KHL, SHL, and other top pro leagues.

      These factors are fairly robust, with age being the biggest consideration to make when evaluating NHLE… (ie looking at a player with a 30NHLE 3 years post draft versus a 30 at 18, take the 18 year old all day)

      • OKG

        How do you explain away Blake Wheeler?

        Draft+4 NCAA NHLe = 27 (using .41 as the equivalency factor)

        Draft+5 (rookie season) Points = 45

        2012 Points – 64

        2013 Points – 70 (Prorated)

        2014 Points – 69

        2015 Points- 61

        • we can play this game all day… Wheeler is an outlier it’s as simple as that.
          is it possible Jankowski could be an outlier.. yes. It sure is. What is that possibilty.. not very big. I’m not a mathematician but I’d say it’s probably less than 30%
          All statsitics, whether it be NHLE, Janko’s SH% in the NCAA tell us is that it’s very unlikely given his draft position and his stats that he will make the NHL there’s at least 7 out of 10 kids his age who will make it before he will. But that doesn’t mean he won’t for certain make it. It also doesn’t mean he will be the next Blake Wheeler either.

          The odds are stacked against him. That’s what this is saying. That’s what people are trying to say.

          • OKG

            IS Wheeler a true outlier just because he’s a statistical outlier?

            The guy was a 5th overall NHL Draft pick. That doesn’t sound like an “outlier”… his NHL performance is right in line with his draft position, the only thing that doesn’t fit is his NCAA performance “NHLe”.

            This is why you can’t just look at the spreadsheets to make conclusions. They can be used in support of actual scouting reports, not in substitution for them.

          • OKG

            If a player was taken 5th overall out of US high school the way Blake Wheeler was, fans would lose their minds today.

            He also followed up his draft year with a middling season in the USHL.

            He is an outlier in every sense of the word, and no his NHLE didn’t track the way you’d want a top 5 pick to.

          • OKG

            The fact that he was an NHLe outlier, yet a 5th overall pick, should be a good sign that relying on stats to evaluate a player unseen is just dumb. Those fans losing their minds were dumb to do so.

            Amusingly enough, another thing Wheeler has in common with Janko is they both won the NCAA championship in their Draft+3 year, and both scored a goal in the winning game.

            If there’s anything for those fans and us to lose our minds over, it’s letting Wheeler/Janko stay in college for their Draft+4 year and potentially walk as UFAs.

        • OKG

          Very easily. He’s a statistical outlier.

          Drafting and acquiring prospects performing poorly at lower levels, thinking they will be positive outliers solely because they are in our jersey and the eyeballs of our staff are better than the eyeballs of 29 other teams in the NHL is a fast and surefire way to get fired.

          Just because Blake Wheeler happened doesn’t mean we can go take unnecessary risks. We remember Blake Wheeler because he was successful, but we forget the 50 other players with similar performance at lower levels that never made it. Why do we forget them? Because we never learned their names. Survivorship bias.

          • OKG

            Geeze, Not even expected in the first round. I had no idea.

            Listen, none of us our saying not to put any weight in the eye test. But you can find players that pass the eye test that are statistically good bets as well.

            The most damning evidence is that Janko’s little brother destroyed at Stanstead this year, in his draft year. His story is exactly the same (late grower, too small in OHL year, quick skater that dominated). He actually crushed Mark’s scoring record too. He wasn’t drafted. No team risked a 7th. Feaster/Weisbrod thought they were smarter than everyone else, and both ended up out of work because of it.

          • OKG

            That’s not “damning evidence”. It’s proof positive that stats are not ever a substitute for scouting.

            David Jankowski was unrated by CSS. He went undrafted. Not a surprise. Though he does sound like a player who will Tyler Johnson his way to the NHL.

            On the other hand Mark Jankowski was rated as a rising 2nd rounder, ranked 46th by CSS. Red Line Report had him at 28, Craig Button had him at 14. Hockey Prospectus had him 36 and McKeens had him 37. He was not as big a surprise as you make him out to be.

          • mattyc

            I think the evidence strongly supports vaccination, but this one time a kid got autism, who was also vaccinated, so now I don’t put any stock into peer-reviewed scientific research.

            Also, it’s cold today (like 8C below normals for July) so I think this global warming business is wrong.

          • OKG

            I feel that some people might look at a kid who dies due to a vaccination as a minor outlying statistic, whereas I would want to understand what it was about the child or the vaccination that killed him/her, so that in the future, that vaccination can be better understood rather than dismissed, because these are human beings.

            I also feel that vaccinations have nothing to do with Flamesnation.ca

            Oh, and I feel that Vaccinations prickle a little bit.

          • Byron Bader

            Blake Wheeler is a great offensive talent today but he’s definitely an outlier (based on this analysis). He took about twice as long as a normal impact player to emerge as so. Most impact scorers make the NHL right away or after 2 years and are impact scorers right away. Wheeler took three to make the NHL (kind of forcing his hand) and then took a few more years to emerge as an impact producer.

            Also most of the other outliers in the impact group made the NHL very early, Bergeron, Kopitar and Ladd specifically. If the players had gone back to junior for their draft+1 or draft+2 year they likely would have crossed the 30+ threshold, reducing the outliers to a list of 3-4 players.

            Wheeler had four cracks and never came close.

            Why did he become a scorer in the NHL? Who knows. Something about the NHL really clicked for him and made him even better.

            Every once in a while a guy hits a three pointer from half court but it doesn’t happen too often.

          • OKG

            “Why did he become a scorer in the NHL? Who knows. Something about the NHL really clicked for him and made him even better.”

            Just luck, pure luck! lol

            I am now expecting Blake Wheeler to be referred to as Outlier Wheeler or Blake Outlier on NHL broadcasts!

            Byron deserves a Nobel Prize for Sports Mathematics for this article, if there ever is one! Well done!

  • OKG

    Best Janko scenario

    Offer him league min ELC, 1 year term, no signing bonus after next season. Hope he balks at it and signs elsewhere.

    Get a conditional 2nd round pick with a higher probability of becoming an NHLer than Janko ever did.