My Thoughts‎ > ‎

Updated Translation Factors

posted Nov 8, 2015, 8:34 AM by Robert Vollman
Translation factors are used to create high-level estimates of how a player’s scoring rate will change as they come to the NHL from other leagues. To create a player’s NHLe (NHL equivalency), multiply their scoring rate in the other league by the factor below.

Updated Translation Factors, as of 2014-15 NHL season
.80 Kontinental Hockey League (up .02)
.60 Swedish Hockey League (up .05)
.47 American Hockey League  (up .02)
.44 Western Collegiate Hockey Association (defunct, up .02)
.41 National Collegiate Hockey Conference (new league)
.40 Switzerland NLA (up .04)
.37 Hockey-East (up .04)
.35 Big-10 (new league)
.32 Ontario Hockey League (up .02)
.32 Central Collegiate Hockey Association (defunct) (up .02)
.29 Finland SM-Liiga  (down .01)
.27 Western Hockey League  (up .01)
.26 Quebec Major Junior Hockey League (no change)
.23 Eastern College Athletic Conference (down .02)

These translation factors are based on everyone to play at least 20 games in the given league before playing at least 20 in the NHL, from the 2005-06 season through 2013-14 (for the target league; add one to each for the NHL seasons), while ignoring the 2012-13 lockout season, since so many NHLers competed in other leagues.

This system was first designed for baseball by Bill James back in 1985, and introduced to hockey 20 years later by Gabriel Desjardins. Since then, I’ve invested a lot of time into the topic, including separating goals and assists, adjusting for age and ice time, considering the slope, and updating the translation factors every year. It’s funny, but a lot of people still use Desjardins’ original 2006 translation factors, which have fallen seriously out of date.

There is a spreadsheet under “Download Statistics” where you can grab these results for yourself, including the formulas, and including separate sheets with all the players that went from that league to the NHL, all throughout history, no matter how many games they played. For years, Desjardins would send me the data, but since he stepped back from hockey, the latest batch of data is courtesy of Josh Weissbock of CHLStats. If you’re interested in this metric, I highly recommend grabbing all the data for yourself.

What changed?

The translation factors haven’t been updated since the 2012-13 NHL season. There was no point updating them in 2013-14, since so many NHLers competed in other leagues in the lockout-shortened 2012-13 season. In fact, that data isn’t even included in the translation factors above. So really, this was merely updated for those who spent 2013-14 somewhere else, and 2014-15 in the NHL.

There were a couple of errors in my spreadsheet when I first announced these new translation factors on Twitter. My first error was to include players who played in fewer than 20 NHL games. It didn’t change the results dramatically in most cases, but some leagues literally have only about one player a year who qualify for the calculation. 

My next error was a typo in the NCAA Division 1 leagues where it would look at the wrong row to figure out which U.S. College league the player was in. It was easy to catch, because it produced a crazy result (i.e. that the ECAC is a stronger league than Hockey East – this was caught by Sarah Connors and Sean Hathaway). Also, every time I re-sorted that page, the translation factor would change – obviously an error!

I include all the formulas in my spreadsheets because I want everyone to know how to calculate everything I do for themselves, and also so they can spot and correct errors, or make improvements and refinements of their own. I’m a very thorough and detail-oriented individual, but I do occasionally make mistakes. I’m not scared of mistakes, and neither should you, provided you check your work, publish your formulas, and set your pride aside when receiving feedback.

What about other leagues? 

Some leagues just don’t have enough players going straight to the NHL on which to base the calculation. I stopped including the Czech Extraliga for just that reason, but the same logic applies to leagues like the USHL or ECHL.

In these cases, I would use the “Wilson Method.” When calculating the NHLe for Mikael Backlund of the Allsvenskan league many years ago, FlamesNation’s Kent Wilson calculated the translation factor from that league to an intermediary league (in this case, Sweden’s Elitserien), and then used that league’s translation factor to convert it to the NHL. It’s a lot of work, but the same method can be used on the Czech Extraliga and the USHL.

Anything else?

Nope, that’s it. For more detail, check the original Hockey Abstract (the blue book), which has a long and detailed chapter on translation factors, including the history, a break down of each league, and an analysis of things like ice time, goals vs assists, age, and more.