Minnesota vs Massachusetts High School Hockey

I've wondered periodically if we had an easy way to evaluate the difference between various high school hockey leagues. Most high school players jump to one of the USHL, NAHL or NCAA, but those leagues tend to not be too concerned with collecting scoring data from high school. And while states tend to post scoring leaders for the entire state each year, I found that players often play under slightly different names in college and junior, which makes matching up the two databases a huge pain.

What we can do, however, is look at players who ended up playing in the NHL to see how they did in high school at Age 17 or 18, and then how they did the next season in another league. Because I'm looking only at NHL players, that next league has historically been the NCAA - while many high school players go to the USHL today, it has yet to catch up to the college numbers over the last 25 years.

At any rate, I have data for only two states that produced a lot of NHL players - Minnesota and Massachusetts. (I could probably find the Michigan data, but I wanted this to be easy!) I ran the standard NHL Equivalency analysis on each group, which gave me the following equivalencies from each high school league to the NCAA:

Minn: 0.31, stdev = 0.30

Mass: 0.30, stdev = 0.17

These figures both point to an NHLE of approximately 0.12 for each league, which is almost as good as the ECHL! This is substantially higher than the 0.05 or so that I previously predicted for Minnesota High School Hockey, but my previous projection used every high scorer over a five-year period, not just players who made the jump to the NHL. It should be obvious that future NHL players are better than the typical guy who scores 30 goals as a senior in high school, and we would expect to see them retain a higher percentage of their performance when they go to the NCAA.

So in picking only future NHL players for the dataset, I engaged in selection bias that skewed the final result. However, there is one important conclusion that can be drawn from this data, and that is that Minnesota and Massachusetts HS hockey have been roughly equivalent in quality over the last 25 years. But we can also see from the much larger standard deviation on the Minnesota estimate that the quality of hockey varies much more around the state than it does in Massachusetts. This isn't surprising - Minnesota is a huge state, geographically, and high school teams are isolated from each other in a way that isn't possible in a small state. So the relative caliber of opposing teams can vary tremendously in a way that it wouldn't if you drove in to Boston for a couple of games every other weekend.