View Single Post
Old 09-28-2017, 09:49 AM   #16
Five-hole
Franchise Player
 
Five-hole's Avatar
 
Join Date: Jan 2003
Location: The C-spot
Exp:
Default

Quote:
Originally Posted by New Era View Post
Monahan being grossly under-rated again. Not surprised, but I see this is a good thing long term. Kind of surprised by the contents of the article, but not really though. Anything based on analytics is ironically bound to be short or reason. Things that jump out at me.

Jordan Staal. Proof that analytics are flawed. Good numbers, but just a bad player. Anyone that would take Staal over Monahan is nucking futz.

Leon Draisaitl. Should a guy play center before being identified as one of the best centers in game? Again, analytics are garbage because they are masking the fact that he rode the tails of the best offensive player in the game and was handed nothing but great opportunities to succeed. Will love to see him play the actual position he's lauded for being of the best in the game at.

Nathan MacKinnon. All that promise, but just not good at his position. The writer needs to get his nose out of Excel and start watching the games with his own eyes. The guy makes bad decisions like Sportsnet makes McDavid references.

Vincent Trochek. Really? In a game where strength up the middle is all but a guarantee of success, how is it possible that the Panthers could be so bad with the likes of Barkov and Trochek in the lineup? Again, anyone taking Trochek over Monahan should have their head examined.

Ryan O'Reilly. See Vincent Trochek. How can Buffalo be so bad with O'Reilly and Eichel in the mix? Something doesn't add up. O'Reilly continues to be one of those over-rated players for some reason.

The crowning glory, McJesus over Crosby, and the kicker is defensive play! If you're looking for proof that analytics are completely flawed - nee garbage - this beauty should tell you that. McDavid can't spell defense, let alone play it, and his numbers are superior to Crosby? Sorry, doesn't pass the eyeball test.
I don't disagree with your points regarding individual players, but I think it's a bit of a leap to suggest all analytics are garbage.

I think it's best to understand this article as presenting a hypothesis. That is, that the best centers in the league are those that are good on offense, defense, and transitioning, all while facing difficult competition. Sure, sounds good. It then goes further and states that those categories consist of a number of stats which are blended, and that there should be a weighted-three year average.

The "experiment" is then run by crunching the numbers, and you test the hypothesis by comparing the results to other "observational evidence" (i.e. the eye test, though I think there's a lot of "reputation" and "intuition" in there as well). As you point out, there are some results that don't cohere with other "observational evidence". As a result, the hypothesis should be tweaked, likely by adjusting the weighting of certain stats.

Throwing statistical analysis in the garbage would be like scrapping the scientific method because one scientist's hypothesis was proven incomplete or incorrect. That misunderstands the whole enterprise.

If nothing else, this kind of article enriches the debate. Otherwise we are simply down to "I think Trochek IS better than Monahan, because my eyes tell me so" vs "that is ridiculous, anyone who takes Trochek over Monahan needs their head examined". I mean, go ahead, but what's the point?
Five-hole is offline   Reply With Quote
The Following 5 Users Say Thank You to Five-hole For This Useful Post: