The bell curve analogy is wrong, not because it's the wrong distribution (I haven't seen any data on the distribution of athletic talent and skill) but because pro athletes are all on the long tail to the right of the distribution, and so far out it's irrelevant what the distribution actually is.
That is, the top players are way out at the end of the tail, and most players are toward the fatter portion of the tail, up to a hypothetical point which would be "replacement level." To the left of replacement level are the AHL "lifers" and ECHL players and all the kids in college and juniors with no NHL future. So there are far more players who are replacement level or just below (since actual performance varies around their core value) than are solid NHL starters. Once all the solid starters have jobs, the remaining spots are filled with replacement players each year in a NHL version of "musical chairs."
Now what complicates this picture is that players rise up to solid starter or even allpro, but eventually decline to replacement level or below, so the names are constantly changing.
I think you literally described what I said up there lol.
The distribution is talent... now there’s not a nominal valuation for talent or even numerical that clearly depicts talent. That could be your argument for its relevance, however, replacement is synonymous with talent in this regard. What you’re arguing on behalf of “replacement level” is, in fact, not statistically correlated, but correlated on the talent or play of the player replacing.
The bell curve analogy is for NHL players. You can use any type of valuation for it as long as it’s normally distributed /60.
To reiterate this... the correlation isn’t dependent upon the statistics. This goes with my previous argument of just because Sam Gange scores 20 goals... it doesn’t make him a great goal scorer.
Andrew MacDonald is not a replacement level player. He’s regressed to a bottom pairing Dman, but that’s based on his play, his career stats and play hardly make him a replacement level player.
If the career of the player, is similar to sub par NHL metrics, you could say they’re a replacement level player. These stats can help you make an argument, but they can’t be your argument.
Look at the careers of Michael Haley, Jody Shelley, Colton Orr, Zac Rinaldo... these players are under this description. Their stats show that their only productions was in intangibles (fear, grit, forechecking... stuff like that). Their play can easily be replaced by the same basis of talent (other players similar in talent level, but that can produce more).
In today’s NHL... Danick Martel, Corban Knight, JT Brown, Mark Foligno, and players of such are who are defined as replacement NHL players... they can come in and fill a role (PK, checker, net front, or whatever) and not be totally out of place, but still play below the “required” NHL talent level.