NHL equivalency has become a popular tool used to evaluate prospects by accounting for the quality of the league. For example, a prospect who is playing in the KHL should put up different numbers than a players in the Swedish junior league or an American in the USHL based on the difficulty of scoring in that particular league. For numerous leagues, the NHL equivalency has been determined and for our purposes, we just want the CHL. The equivalencies are:
WHL | OHL | QMJHL |
0.30 | 0.30 | 0.28 |
The WHL and OHL have the same equivalency of 0.30, meaning that a player who puts up a point per game in juniors should be able to score at a 0.30 Pts/G in the NHL. Meanwhile a player from the Q would only scores at a rate of 0.28 Pts/G. Based on NHL equivalencies, the QMJHL is the inferior league of the three and players point totals are inflated compared to the OHL and WHL. To look into this further will examine the save percentage of the three leagues since 1998 (regular season and playoffs combined). Lucky for you I have done the grunt work for this when doing my 2014 rankings:
WHL | OHL | QMJHL | |
SAVES | 759411 | 775258 | 632271 |
SHOTS | 847099 | 862909 | 711262 |
SA% | 0.896 | 0.898 | 0.889 |
We can see that the Q has a lower SA% than both the WHL and the OHL, which helps support the notion that it is an easier league to score in and that point totals are inflated. Therefore, when we adjust the point totals for Q players by a ratio of 0.933 (28/30) we should expect the correlation to get stronger in the regression model from my last article. After adjusting the point totals, the correlation and adjusted r-squared value:
Non Power-Play | |||
Correlation (w/o NHLE) | 0.584 | Adjusted r-squared (w/o NHLE) | .336 |
Correlation (w/ NHLE) | 0.571 | Adjusted r-squared (w/ NHLE) | .321 |
Power-Play | |||
Correlation (w/o NHLE) | 0.571 | Adjusted r-squared (w/o NHLE) | 0.315 |
Correlation (w/ NHLE) | 0.567 | Adjusted r-squared (w/ NHLE) | 0.310 |
Well these numbers were unexpected for those that support NHL equivalency, as the correlation and adjusted r-squared weakened when we adjusted for it. This goes against the notion that the Q is an inferior league and is easier to score in, but instead the model suggests that a point in the Q is worth every bit as much as a point in the WHL or OHL. If that is the case, this brings up the question as to why does the Q have a lower SA% than the other two leagues? One angle to consider is that the Q may be a better league at focusing and developing offensive skills that ultimately produces better scoring ability and more goals. Therefore, this sample of players suggests the QMJHL is not the inferior league some like to believe it is and that adjusting the point totals down for NHL equivalency are doing disfavour to themselves. To do that in the 2015 draft would be one of the worst year’s to undervalue the talent of the Q, as the depth from the Q is arguably the best we’ve seen out of that league in the last decade. This is not the year to forget about the Q.
Would you mind sharing where do you get the PP information? I used to tediously transfer data from the hockey database. There has to be something easier..
LikeLike
I wish I had an easier solution for you as I have done countless hours doing tedious work in getting the PP information. I do not get why the CHL sites will have cumulative stats for PP goals but for assists they do not. All the information for PP stats is in the game summaries and someone with some programming knowledge better than I have may know a way to mine the data.
LikeLike
Analyzing 25 years of OHL stats revealed 3 significant factors in predicting an NHL season with 1 PPG or more: point per game, age and proportion of points vs team total adjusted based on number of games. R=0.71. Would of been interesting to have had the PP info with that dataset. Age, goals and goals per game were not significant. That analysis was done 10 years ago using Minitab software. Always fascinated by the numbers. Thanks for posting your stuff.
LikeLike
That is very interesting. It’s funny what one assumes and what one finds out when they actually run the numbers. My views towards scouting continue to evolve the more I research. I do want to add proportion of team scoring to my sample.
LikeLike
Did not find anything gor Dman with the dataset…
LikeLike
Wouldn’t going back to 1998 going to far? All 3 leagues have changed vastly since then. I’m curious if the analysis going back just from 05-06 forward would look. Also, adding the USHL to the mix would be an interesting addition as well.
LikeLike
No I don’t think it’s going back too far. The junior leagues have not seen as great a swing in league scoring like the NHL. The major difference I notice is in what we value for d-men and what skills are emphasized. That’s been the bigger change for juniors.
USHL has changed a lot over the years in terms of skill. I did apply the formula to the USHL stats and will be doing a post in the future.
LikeLike