[Enoch Root]

Originally Posted by Jay Random

You've made it clear that you don't.



Yes. Because there are some terrible teams in the NHL, and there are some players on those teams who shouldn't be there (and won't be once their contracts expire). At the other end, there are several teams that have very deep prospect bases and not enough NHL jobs for all of them, but they are not about to give away those prospects to bad teams out of the goodness of their hearts.

The other thing you don't seem to get is the distribution involved. It's not a bell curve; it's the extreme right-hand tail of a Pareto distribution. On the left you have masses and masses of beer leaguers and hockey-playing kids, who will never see the inside of a pro hockey game except by paying admission. On the right you have the tail of the distribution, which gets thinner as you reach the best players in the game. The difference in raw ability from the 82nd percentile to the 100th is much greater than from 0 to 18. McDavid, McKinnon, and Makar are miles better than the 100th-best player, but the 500th-best and 600th-best players (ignoring position) are pretty nearly interchangeable. In fact, a small change in the weighting of factors in the model could cause them to exchange places on the list. (This is why I wish JFresh included raw WAR scores as well as percentiles. You would then see just how close together the ratings of depth players are.)

The cutoff – that is, the worst player in the NHL, the poor schlub who rates 0% – is far out on that tail: miles better than rec hockey players, but he's only in the league because he fills a specific role, or because he used to fill it and is hanging on until his contract runs out. The face-puncher (a dying breed but not yet extinct), the fourth-line PK specialist, the 7th D who knows his principal job is to munch popcorn – these are your typical sub-replacement-level players.

There is also the fact that JFresh's data include players who aren't full-time NHLers. Players below replacement level can get a cup of coffee in the NHL as a reward for good play in the minors. They can be called up as injury replacements, and sometimes stay on the roster for quite some time. Prospects on the cusp go up and down frequently during the season. Some of those players are better than replacement value and are sent back down for contract reasons or because the team just doesn't need another player at that position. Some of them are worse than replacement value, and are, in fact, being used as replacements simply because they are already under contract and can be called up without complications.

Basically, the bottom 20% of JFresh's distribution are tweeners, and it's a matter of individual circumstance which ones get an NHL roster spot and for how long.

Engaging in a discussion with you is simply not worth the bother.