Behindthenet Blog

The value of keeping better statistics has finally moved past the NHL, with the top Finnish, Swedish and Russian leagues tracking and publishing ice time figures for every player. There's very little player movement between the NHL and Finland, but there's enough to generate an NHL Equivalency for the other two leagues that accounts for playing time differences.

Here are the Russian (KHL) and Swedish summaries:

     N  TOI OTH TOI NHL % TOI OTH/82 NHL/82 SAME TOI
KHL  23   18.0    14.5   80.8   45     24     30     
SEL  15   17.9    14.0   77.8   38     22     28

This gives equivalencies of 0.57 for the SEL and 0.54 for the KHL. Adjusting for icetime differences, the NHLEs are 0.78 (SEL) and 0.67 (KHL). The Russian figures are surprisingly low - in the past, the top Russian league was better than the SEL, and approached the caliber of the NHL in some cases.

Labels: Equivalencies

My attempts to analyze NHL Equivalencies have always been hampered by the question of usage. The NHL has published player ice time (TOI) data for over ten years, but other leagues have been slower to publish the data, either because they didn't track it officially or because they released only a very limited number of statistics to the public.

The Swedish Elite League has now published TOI for two seasons, which allows me to include ice time in the equivalencies. There aren't a huge number of players who've played significant time in both leagues the last two seasons, and because we only have one past season of SEL data, most of the players in question went from the NHL to the SEL, which is not the preferred direction for an equivalency study.

Here's the list:

                    Dir TOI SEL TOI NHL % TOI SEL/82 NHL/82 +22% TOI
Joe DiPenta         SEL   18.4    10.6   57.9   10     18     23
Kris Beech          SEL   18.5    11.5   62.1   62     36     46
Mathias Tjarnqvist  SEL   21.2    13.7   64.3   43     12     15
Ossi Vaananen       SEL   21.2    14.3   67.6   27      9     11
Jaroslav Hlinka     SEL   19.5    13.9   71.3   84     36     47
Petr Tenkrat        SEL   17.0    12.5   73.6   43     18     23
Ronald Petrovicky   SEL   10.5     8.3   79.1    9     16     20
Jan Hlavac          SEL   19.1    15.2   79.3   73     36     46
Josef Melichar      SEL   19.3    18.8   97.6   13     14     18
Michael Holmqvist   SEL   12.7    14.0  109.7   26     17     22
Patric Hornqvist    NHL   17.8    11.4   64.0   46     21     26
Josef Melichar      NHL   19.3    13.8   71.4   13     19     24
Ossi Vaananen       NHL   21.2    17.9   84.5   27     18     24
Fabian Brunnstrom   NHL   12.9    11.6   89.9   56     43     56
Jonas Frogren       NHL   17.3    13.4   77.5    2     14     18
       
                          17.9    14.0   77.8   38     22     28

Players retained approximately 78% of their ice time when they went from the SEL to the NHL, and 57% of their per-game scoring. However, if we compare scoring on a per-minute basis, giving each player 22.2% more ice time, they retain 72% of their scoring. In other words, given equal ice time, we would expect them to score 26% more than they do*. But it's not true for everyone. Victor Hedman played over 21 minutes already as an 18-year-old in Sweden, so it's difficult to imagine him playing much more than that as a rookie in the NHL. Obviously we expect huge skills growth for teenagers, but a 40-point season would be a tremendous achievement for Hedman in Tampa Bay.

*In a previous analysis, I found that when we adjusted AHL players for PP scoring, their NHL Equivalency was 32% higher than if we ignored usage. The usage factor in the SEL and AHL seems to be similar.

Labels: Equivalencies

http://puckprospectus.com/article.php?articleid=192

"Overall, Minnesota hockey translates to the NCAA (NHLE = 0.41) at approximately 0.18, giving an NHLE of 0.073. The translation to the USHL is 0.195; its translation to the NCAA is 0.65; the overall NHLE is 0.052. Via a similar process, the NHLE via the NAHL is also 0.052. This puts the difficulty level of Minnesota H.S. hockey somewhere between 5.2% and 7.3% - which is not very high: the leading scorer in Minnesota over the course of a decade might be good for 20 points as an 18-year-old rookie in the NHL."

Labels: Equivalencies, Puck Prospectus

The 2008-09 KHL season is about 14 games in, and we can take our first stab at producing a league equivalency. There are 23 skaters in the KHL who played at least 20 games in the NHL last year. Their cumulative stats:

	GP	G	A	PP82
NHL 2007-08	1191	131	228	25
KHL 2008-09	260	54	76	41

That would give a league equivalency of 25/41 ~ 0.60. However, the KHL has much lower scoring and a lower assist-per-goal rate than the NHL, so we need to adjust this figure by the overall PPG scoring rates of the two leagues = 0.41 (NHL) / 0.36 (KHL). This gives a league equivalency of 0.70, which is somewhat lower than Russian leagues have had in the past, albeit with much larger numbers of players and games played in the sample.

Labels: Equivalencies, KHL, Projections

I now have five-to-seven years of player data for Canadian Junior A leagues, so it's now possible to generate equivalencies for each league. I ran the data for British Columbia (BCHL), Alberta (AJHL), Saskatchewan (SJHL), Manitoba (MJHL), Ontario (OPJHL), Quebec (QJAHL) and the Maritimes (MJAHL).

The Quebec and Maritimes leagues did not send a significant number of players to the NCAA, and few players went from the SJHL to Major Junior. The Manitoba league was more likely to send players to other Junior A league than to have them step up to a higher-level of hockey.

Here are the equivalencies normalized to a player age of 18:

Junior A to NCAA	EQUIV	Mean Age	N
*USHL	0.66	19.41	384
AJHL	0.39	18.38	91
BCHL	0.34	18.34	190
OPJHL	0.32	18.33	94
SJHL	0.31	18.37	42

A huge drop from the USHL to the Junior A leagues - but not a huge difference among the Canadian leagues. The AJHL seems to have a higher level of play.

Junior A to Major Junior	EQUIV	Mean Age	N
AJHL	0.61	17.43	48
BCHL	0.57	18.70	34
QJAHL	0.51	18.51	30
MJAHL	0.47	17.62	31
OPJHL	0.44	18.44	217

Less of a difference this time, but again the AJHL is on top.

Major Junior to Junior A	EQUIV	Mean Age	N
BCHL	0.33	18.37	33
SJHL	0.27	19.17	29
AJHL	0.22	18.80	43
QJAHL	0.21	18.71	85
MJAHL	0.20	19.23	19
OPJHL	0.17	18.20	71

Note that the equivalency is in this table is actually for Junior A to Major Junior so that we can compare the relative performance of players stepping up a level of hockey versus those stepping down. Junior A leagues appear substantially less difficult for players coming from Major Junior than they did to players going in the other direction.

Overall, assuming that the level of play is identical in all three Canadian major Leagues (which has been shown elsewhere on this site), then the BCHL and AJHL appear to have the highest level of play of all the Junior A leagues.

Labels: Equivalencies, junior hockey, Projections

Kyle Okposo made a few headlines this week: after the World Junior Championships are over, he's quitting the University of Minnesota and joining the New York Islanders. He was drafted 7th overall in 2006, which puts him a bit behind the curve in jumping to the NHL: he's the only player in the top 15 who hasn't played professional hockey. His former teammates Phil Kessel and Erik Johnson are playing regularly in the NHL.

So the question is, if the Islanders keep him in the NHL, what can we expect from Okposo for his next 40 games? This chart shows the ratio of NHL points-per-game to NCAA points-per-game for all players who've jumped from college to the NHL since 1979:

	10%	25%	50%	75%	90%
NCAA	0.15	0.21	0.29	0.39	0.48

So 50% of players retained 29% of their scoring in the NHL; just 10% retained 48% or more. Okposo has had a pedestrian 2007 NCAA season (11 points in 18 games), and if we combine that with 2006 (40 points in 40 games), his middle 50% scoring range amounts to 7-14 points in the second half of the 2007-08 season.

That seems a little conservative, and we should take into account that Okposo isn't 20 years old yet. This chart shows the difference in PPG ratios by age:



So we were probably 25% too pessimistic with our projection. Okposo is a good bet to pick up between 10 and 18 points in his next 40 games. That's low enough that the Islanders might just assign him to their AHL affiliate, and we won't see Okposo until next year.

Labels: Equivalencies, NCAA, Projections

The short column I wrote the level of play in the USHL has generated a lot of interest in the last couple of days.


Marc Foster's junior hockey blog made reference to it:

Junior Hockey Blog

And the USHL President's office and Director of Scouting got in touch with me to express their gratitude at seeing these numbers available somewhere. A big thanks to James Mirtle for his blog post on the USHL, which made me think it was finally time to analyze the USHL.

Labels: Equivalencies, junior hockey, Projections, USHL