NHL EQUIVALENCIES

As we see more and more players selected from the USHL, BCJHL and the New England leagues whose names change with the seasons, it’s becoming more and more important to get an idea about equivalencies for these boutique leagues. Scott Reynolds of Copper and Blue tackles a lot of things guys like me wouldn’t care to, and his NHLE’s (here) in 2011 looked like this:

  • Reynolds: KHL (multiply offense by 0.83), SEL (0.78), CZE (0.74), FNL (0.54), NCAA (0.41), WHL (0.30), OHL (0.30) and QMJHL (0.28); this article for the translations from the USHL (0.27), AJHL (0.16), and BCHL (0.14); and this article for the translations from High School hockey (0.0625). Unfortunately, especially at this early stage in development, there are a lot of leagues that don’t have an NHL equivalency number yet including the U20 leagues from Europe.

That last part from Reynolds is a growing issue. At this year’s draft, Craig Button called the Swedish J20 league the ‘best junior league outside the CHL’ which is both fascinating and (from what I can see) unproven. How do we go about this? Does it make sense to gather the J20 kids who move to SHL season-over-season and use that as an NHLE?

What about the USHL? The .27 number seems aggressive based on conversations I’ve had with scouts (Tom Lynn, others) and observers of the game. The idea here isn’t to call anyone out, or to imply work done in the past has no value. However, what is the best plan of action for this?

Rob Vollman’s Hockey Abstract a year ago was very helpful in updating major junior and the Euro leagues, plus the AHL. I used his AHL numbers and Euro numbers through the draft window, but saw some resistance to the CHL numbers. Where are we, as a community on this?

Thoughts?

written by

The author didn‘t add any Information to his profile yet.
Related Posts

21 Responses to "NHL EQUIVALENCIES"

  1. John Chambers says:

    I think NHLE is absurd as a concept the more I see it referenced. It can be a useful guide for comparing draft age players across different leagues, but that it implies that the player could slide into a major league lineup and post numbers, even with linemates named Tangradi or Gazdic, playing 9 minutes a night, does not equate.

    Basically anyone who can legitimately put up 30 pts over an NHL season should be in the league, and there can’t possibly be 25 – 40 teenagers, 60 or 70 minor leaguers, and a whole infantry of Russian leaguers that are a better option for the Hurricanes’ bottom-6 than Riley Nash.

  2. geowal says:

    I think nhle for second tier leagues will look a little wonky, because I think by in large only the best will be drafted (less than one per team?) versus the chl with supporting players as well as the stars getting drafted.
    Off topic: still no thoughts on the oilogosphere on Jason Williams? Bueller? Bueller?

  3. Romulus Apotheosis says:

    “Rob Vollman’s Hockey Abstract a year ago was very helpful in updating major junior and the Euro leagues, plus the AHL. I used his AHL numbers and Euro numbers through the draft window, but saw some resistance to the CHL numbers. Where are we, as a community on this?”

    For my part I trust Vollman’s corrections and updatings to the NHLE numbers by Desjardins… but, I also assume that these things are more cyclical and reactive than the latest numbers can avail us of.

    For that I think having GB’s base line, semi-equalized CHL (the Q gets dinged) numbers helps. So… I use both and add everything else I can get my hands on.

    make sense?

  4. Lowetide says:

    Romulus Apotheosis:
    “Rob Vollman’s Hockey Abstract a year ago was very helpful in updating major junior and the Euro leagues, plus the AHL. I used his AHL numbers and Euro numbers through the draft window, but saw some resistance to the CHL numbers. Where are we, as a community on this?”

    For my part I trust Vollman’s corrections and updatings to the NHLE numbers by Desjardins… but, I also assume that these things are more cyclical and reactive than the latest numbers can avail us of.

    For that I think having GB’s base line, semi-equalized CHL (the Q gets dinged) numbers helps. So… I use both and add everything else I can get my hands on.

    make sense?

    That’s where I am, but what about the USHL? The Gabe/Scott Reynolds number has it .27, which is basically equal to QMJHL. Does anyone believe that? Maybe I should take Vesel’s USHL’s number run it though at NCAA levels and see if he can match next season?

  5. Romulus Apotheosis says:

    Lowetide: That’s where I am, but what about the USHL? The Gabe/Scott Reynolds number has it .27, which is basically equal to QMJHL. Does anyone believe that? Maybe I should take Vesel’s USHL’s number run it though at NCAA levels and see if he can match next season?

    I don’t know how to figure on the USHL numbers…

    but, what I’d do for sure is ding the overagers, or boost the kids drafted in their draft year.

  6. RexLibris says:

    I’ve long seen NHLE as a ballpark, to be contrasted within a narrow margin of peers.

    It does not, I believe, translate well over time, but like other junior stats like ppg, IPP and, frankly, size, it helps to colour in the lines on the player.

    I’ve been kicking around the idea of doing some projections for rosters using NHLE, but estimating potential impact of rookies and AHL pros into the NHL seems like a fool’s errand. Which makes LT’s RE series all the more impressive.

    I’m going to take a crack at it all the same, as the exercise ought to be enlightening, but I don’t hold out much hope for my results over time.

  7. Frank The Dog says:

    Hey Rom, unless we subscribe to the fact that goals scored net of giveaways are the final determinant of the success of a given pp then we are surely fooling ourselves.

    Having said that I’d agree that had Eakins enjoyed a better shot % and less takeaways but isn’t that the point? Analytics are there to validate reality rather than contradict it.

    I’d suggest that to unlock Kruegers success it may be more profitable to see what he did to cause the oncrease on the sh % than resort to the statistical disclaimer of plain good luck.

    Edit: To enlarge, what did Krueger do, that increased the sh% of the unblocked shots? Did they come from misdirection, or unexpected angled, or were the shots taken just that fraction quicker that gave the g just that little less time to react? Or all of the above?

  8. RexLibris says:

    Romulus Apotheosis: I don’t know how to figure on the USHL numbers…

    but, what I’d do for sure is ding the overagers, or boost the kids drafted in their draft year.

    I’ve been looking at NHLE on a year-by-year basis.

    For example, I was recently involved in a discussion of Hunter Smith as a prospect. His NHLE looks interesting from his last junior year in the OHL (15.37), but the year before, his first draft-eligible year, his NHLE was 1.09. His point total from his draft eligible year to his 19 year-old season (draft year) went from 2pts to 40.

    Now, statistically you could take the average of the two, but no self-respecting statistician would be content taking that approach.

    So the alternative is to rely on the old-school method and look for intangible circumstances, like the fact that his physique closely resembles a brick outhouse and his point production increased as a result of his imposing size against younger, slighter opposition.

    Arguing that it was an increase in ice time or emergent skill seems sketchy at best.

    So using NHLE to evaluate this player is essentially pointless (see what I did there) as it provides a result that does not coincide with other available data. This gets back to our previous discussion a few days ago about integrating analytics with visual observation of the player.

    I’d prefer to use NHLE to track the progress of forward prospects playing several seasons in the same league to determine their rate of improvement in that one category. I wouldn’t apply it to defensemen to be honest, except in so far as to illustrate some measure of offensive awareness or ability to make critical passes that directly impact scoring.

  9. Romulus Apotheosis says:

    Frank The Dog: Hey Rom, unless we subscribe to the fact that goals scored net of giveaways are the final determinant of the success of a given pp then we are surely fooling ourselves.

    I’ve read this many times. I can’t make out what it means. Am I missing a clause or something?

    Frank The Dog: Having said that I’d agree that had Eakins enjoyed a better shot % and less takeaways but isn’t that the point? Analytics are there to validate reality rather than contradict it.

    Not sure about “takeaways” or what that has to do with my article (I didn’t bring them up).

    On “shot %”… do you mean shooting %? if so Eakins didn’t. if you meant FF/60, he did. And we don’t need to agree or disagree about these things, they are simple empirical records, not pieces of analysis.

    Frank The Dog: I’d suggest that to unlock Kruegers success it may be more profitable to see what he did to cause the oncrease on the sh % than resort to the statistical disclaimer of plain good luck.

    This reduces to a shot quality argument: Krueger somehow managed to construct a team epically bad at creating shots, but epically great at scoring goals.

    All the data we have suggest this is not what happened.

    Luck is a perfectly reasonable explanation here.

  10. Romulus Apotheosis says:

    RexLibris: Arguing that it was an increase in ice time or emergent skill seems sketchy at best.

    So using NHLE to evaluate this player is essentially pointless (see what I did there) as it provides a result that does not coincide with other available data. This gets back to our previous discussion a few days ago about integrating analytics with visual observation of the player.
    I’d prefer to use NHLE to track the progress of forward prospects playing several seasons in the same league to determine their rate of improvement in that one category. I wouldn’t apply it to defensemen to be honest, except in so far as to illustrate some measure of offensive awareness or ability to make critical passes that directly impact scoring.

    http://www.eliteprospects.com/player.php?player=117017

    http://www.extraskater.com/chl/players?team=osh

    http://www.inlouwetrust.com/2014/6/9/5784142/hunter-smith-2014-nhl-draft-prospect-profile

    I’m not sure why you’d discount his NHLE for his last season.

    I’d take it in context, as you suggest, with the rest of the stuff you know about him. He doesn’t look like a wise bet to me, esp. in the 2nd.

    As far as D go, as long as we think PPG and discipline splits tell us something interesting about D and their future, I don’t see why NHLE wouldn’t also.

  11. crude says:

    Rom,

    I tried to read your piece, but the Chet was too good.

  12. Romulus Apotheosis says:

    crude:
    Rom,

    I tried to read your piece, but the Chet was too good.

    Awesome!

  13. Frank The Dog says:

    Hi Rom,
    My attempts at answers as follows:
    Rom: “I’ve read this many times. I can’t make out what it means. Am I missing a clause or something?”
    Re-phrased: Ultimately speaking, a perfect power play would be one where every power play results in a goal, and not short handed goal is ever allowed. The closer a power play gets to this goal, the better the power play. As opposed to FF% being the final determinant.

    Rom: “Not sure about “takeaways” or what that has to do with my article (I didn’t bring them up).”
    Clarification: Takeaways resulting in short handed goals against.

    Rom: “On “shot %”… do you mean shooting %?” FTD: Yes.

    Rom: “if so Eakins didn’t. if you meant FF/60, he did. And we don’t need to agree or disagree about these things, they are simple empirical records, not pieces of analysis.”

    FTD: There’s empirical analysis and there’s what you make of it. You are correct that Eakins had the superior FF% and inferior SH%. It is what it is. I disagree that the higher SH% is luck. I’m suggesting that just because we haven’t nailed what the cause of the higher SH% is, does not mean its due to sequential random chance, similar to a coin landing two heads in succession. I believe there is a reason that Krueger’s PP had the higher SH%.

    Rom: “This reduces to a shot quality argument: Krueger somehow managed to construct a team epically bad at creating shots, but epically great at scoring goals.

    All the data we have suggest this is not what happened.

    Luck is a perfectly reasonable explanation here.”

    FTD: We don’t know how Krueger maintained that high SH% over both seasons. But that is because we don’t know, not because Gord stuck His finger in it. The random variation aka “luck” would have averaged out over the course of the year, nor was it bad luck that produced the short handed goals.

    But this is not the end of the world, I just happen to think that it’s not luck, it’s something about the shot itself that resulted in the higher shooting %. Until such time as someone goes and examines the video of Krueger’s and then Eakins’ PP shots and goals, neither of us will have the definitive answer.

    Cheers!

  14. oilswell says:

    1. I think the uncertainties underlying the translations are a concern. Eric T did a good job looking at projection errors and the basics are all there to be applied to the USHL number. Namely: if your sample is small your best estimator is still pretty likely to be widely askew because of the variance in general. This seems to be a likely threat for USHL numbers. The approach to lump all scoring rates into a single composite number may only exacerbate the sensitivity of the measure to sampling problems. Certainly, I think publishing translation factors without also expressing variance is a major omission. Don’t you want to know that AHL-to-NHL at age 22 varies only by 20% compared to 80% from NSBSHL (totally made up stats, you get the point).

    2. Source bias. Players selected in the same draft position from the OHL can be worse than players from the USHL and still find their way on rosters. I know of no data to support this supposition…apart from the USHL equivalencies! ;)

  15. garret9 says:

    I use NHLEs but I always take with a huge grain of salt… one reason is because of this:
    https://pbs.twimg.com/media/BtCXYKuCIAE8Wra.jpg:large

  16. Romulus Apotheosis says:

    Frank The Dog: Re-phrased: Ultimately speaking, a perfect power play would be one where every power play results in a goal, and not short handed goal is ever allowed. The closer a power play gets to this goal, the better the power play. As opposed to FF% being the final determinant.

    This doesn’t solve anything.

    The question isn’t “is it good to score on the PP?” but “how does one increase their chances of crafting a successful and repeatable PP”?

    The answer to that question has gone through several independent studies (cited in my piece). The answer is pretty clear (as of current thinking): FF/60 is the most reliable indicator of future PP success. It simply is.

    thanks for clarification though :)

    Frank The Dog: Clarification: Takeaways resulting in short handed goals against.

    this isn’t relevant to the question I am asking.

    SHG open a whole other kettle of fish.

    Frank The Dog: I disagree that the higher SH% is luck. I’m suggesting that just because we haven’t nailed what the cause of the higher SH% is, does not mean its due to sequential random chance, similar to a coin landing two heads in succession. I believe there is a reason that Krueger’s PP had the higher SH%.

    You need to re-read the part on luck in my piece.

    Luck doesn’t mean “random chance.” I distinguish between this kind of “luck” and the luck I mean.

    Luck is a matter of repeatability. As the studies I cite show, PP SH% and PP efficiency are not nearly as repeatable as shot attempts.

    The take away, is that the most reliable way to build a successful PP model that can repeat its results is to build a model that generates a lot of shots attempts.

    This is basically the opposite of the Krueger years.

  17. Romulus Apotheosis says:

    garret9:
    I use NHLEs but I always take with a huge grain of salt… one reason is because of this:
    https://pbs.twimg.com/media/BtCXYKuCIAE8Wra.jpg:large

    It’s amazing how off the pace the WHL was for a few years there.

  18. Scott Reynolds says:

    Trying to use one number as an NHL equivalency for each league is a very basic starting point that’s going to hide a lot of things. We already know that the number is an average of players who’ve moved from one league to another and that the range is pretty substantial; we know that younger players merit a higher equivalency than older ones; and we know that there’s a bias in the numbers from taking only the players who actually make it to the NHL or other target league (almost no one maintains their equivalency moving from the CHL to the AHL).

    Since the USHL translation is actually based on using the NCAA as the target league, it suffers from that last problem less, but the age and range issues remain, both of which will be important to keep in mind while projecting Vesel’s numbers.

  19. RexLibris says:

    Romulus Apotheosis,

    I guess I should say that I wouldn’t entirely discount NHLE for a defender, but like boxcars on a defenseman, I don’t give them as much weight as I do even for a forward.

    Take Petry for instance. Fans who don’t like him list his lack of offensive production among his many sins, but those who look a little deeper know that his involvement in the offensive side of the game, by way of puck retrieval and movement is a crucial part of the offensive game of the team, even if it doesn’t show up in the boxcars.

    So to that end, NHLE on a defenseman I don’t find particularly compelling. If it is very high it raises eyebrows and prompts more questions. Very low and I wonder about the real value in the stat as I question whether the player is simply placed in important defensive situations.

    Like any stat, it has its uses provided it is assigned and interpreted appropriately.

    As for Smith, yeah, he was a poor bet, in my opinion, based on where he was taken, but it is also a subtle indication that Brian Burke, or at least his team philosophy, holds significant sway within the Flames, either in combination with or in spite of Brad Treliving.

  20. "Steve Smith" says:

    Frank The Dog:
    Re-phrased: Ultimately speaking, a perfect power play would be one where every power play results in a goal, and not short handed goal is ever allowed. The closer a power play gets to this goal, the better the power play. As opposed to FF% being the final determinant.

    I don’t think that goes far enough: the ultimate goal of a power play is to help a team win, so I think we should measure how many goals a power play produces in a winning effort, and use that as our measure of success.

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

© Copyright - Lowetide.ca