What does the fate of Griffin Reinhart teach us? An understanding that projection requires us to think in terms of Probability

Griffin Reinhart was waived today by the Vegas Golden Knights.  This signals probably the end of Griffin's NHL chances - maybe not for good (as Vegas should cool down real fast as their underlying numbers are bad and they're bound to regress at some point, leading them to possibly recall Griff in desperation), but it is at the very least near the end of the line for Griffin Reinhart.  Reinhart was of course drafted by the Isles with the number four pick.  He then was traded to the Oilers for the picks that eventually became Matt Barzal and Mitchell Stephens.  Reinhart then bounced between the Oilers and their farm club for a while before becoming a member of Vegas in the expansion draft this year.


I want to use Reinhart's seeming end as an opportunity to look back quickly at a debate regarding him and how we should talk about prospects in general:



There has always been a lot of talk both positively and negatively about Reinhart.  He has the build of an ideal big defenseman and played in juniors for a topnotch junior team in a big time hockey city - the Edmonton Oil Kings.  But by his first year and second year out of the draft, there were a great number of scouting and statistical-minded people who viewed him quite negatively, even if there were a number of still believers.

An old post by Daniel Friedman, an Isles blogger, was retweeted by a friend on twitter today, taking aim at a column by yours truly about Reinhart.  My column on Reinhart can be found here and the main point of the column was that while there were a few reasons to still be high on Reinhart in his Draft+2 year, there were concerns due to his drop in scoring.  As I explained in that post, a lack of scoring for juniors-playing Defensemen prospects has in general (to oversimplify a bit) been found to be a negative marker for success in the NHL - such players have been successes of course, but at a lower rate than guys who have better scoring numbers.

Friedman took issue with this post (although, annoyingly, he doesn't link it in his piece for some reason).  I'm going to ignore the parts of his post that are basically yelling "WATCH THE GAME" because well, cmon now, you know why such arguments are bad if you're reading me at this point.  But in general, a large gist of his post is that a bunch of low scoring D-men in juniors have found success in the NHL, so a lack of scoring by Reinhart is not worrisome and can be explained by other factors.  To use one quote from that article:

Odds don’t really exist here, if you think about it. Reinhart can either play or he can’t, and we’ll have a much better understanding of which one applies to him once, you know, he actually plays in the NHL. To sit there calculating the odds of success or failure based on some trend is absolutely ridiculous.

Reinhart's failure isn't actually proof that Friedman was wrong, nor is the fact that my pick for a Reinhart's best comparable, Duncan Siemens, now seems incredibly apt.  But it's another opportunity to explain why Friedman's view, which isn't different from that of many people, is wrong:  probability is basically all we have to go with when projecting the future - whether that be of prospects or of  actual NHL players and teams.

My post on Reinhart was a post about probability - how often prospects with Reinhart's statistical features succeed in the NHL.  Those probability numbers are by definition not absolute projections - a player more likely to bust than succeed will succeed some number of times also by definition.

Indeed, statistical markers have been negative for some successful Isles players in the past before.  For example, Anders Lee's #s in his draft year (as an overager) and then in his last year of college certainly didn't predict that Lee would become anything more than an average NHLer, if even that - his junior year suggests that he was more likely to fail at becoming an NHLer than not.  Now of course, Anders Lee is a well above average NHL player.  Similarly, Josh Ho-Sang's draft +1 and draft +2 #s were not that of a player one could expect to be an above average NHLer, while JHS's performance in the NHL so far suggests he may very well be that type of player.

Reinhart's failure is not proof that all low probability events are going to turn out for the worse, just like the success of gambles isn't proof that the odds involved with the gamble are wrong.  But it should hopefully make you understand what I was talking about in terms of the probability of his success back then and why making such a gamble with a high pick or with high draft picks in trade (Edmonton) is a risky move that probably should've been avoided*.

*Post-Post Addition: It should be noted that the low probability that Griff might succeed may have been a worthy gamble for VEGAS to take, as the assets they were spending on Griff was much much lower and the benefit of a win for Griff was much higher than Edmonton/NYI.  Again, probability explains their optimal choice here.