Just recently, Carey Price became the first player with a cap hit of over 10 millions to win a playoff round. And he had to stand on his head like crazy to do it. They had a bit of luck to win another round quite easily thanks to the Scheifele suspension and key Jets injuries but I think you see where I'm coming from. Overpaying players isn't a receipe for success. If they don't buy into what you're trying to achieve, say goodbye. Might hurt at times and that's what the Bluejackets are experiencing now but overpaying just doesn't work. I think Kekalainen is well aware of it. Cap is flat and once you overpay for one guy, nobody is gonna be prepared to accept a discount. Gotta do it like the Bolts and have players buy into what you're trying to achieve. So no to Marner. Besides that, he's soft as f*** and the eye test says he's not at all good defensively. Anyway, the Leafs already said he's not gonna go anywhere so it's not worth talking about.
"Defensive" play gets evaluatated several different ways. He might not be your best choice in a scrum in the defensive zone, but solid positionining as a supporting forward and being so damn good at puck management and transition that the oppsoition hardly ever gets the chance to
be in the defensive zone makes up for a lot.
* * *
So all the stats are wrong and what you see is right?
More that the stats are measuring something else. It's bizarre to me how some hobbyist analytics folks seem to, when their models undervalue playmaking forwards as compared to triggermen, acknowledge this and recognize it as a shortcoming, and then they turn around and act like their models assigning exceedingly high value to highly-transitional mobile defensive play and greatly devaluing stay-at-home good-in-their-zone defensive play represents The True Way Of Things. There's an implicit bias in there, and while some folks do at least recognize that it's unavoidable because the latter doesn't have enough measurable events to properly quantify, some folks seem to be all "well, if it's not quantifiable, it doesn't exist" and come up with all sorts of strange rationalizations to justify it.
In Marner's case, his play style is one that hobbyist-level analytics favors - he's a mobile transition guy - and so analytics guys can easily conclude he's elite defensively. The toughness-biased eye test, on the other hand, is liable to see him not throwing his weight around and thereby conclude he's just not defensively capable. Both are almost certainly wrong, and playing into their own biases - it's just easier to attribute same with the eye test.
The key is recognizing what is and is not presently precisely measured with publicly available raw data. Crossing the line with the puck is. Shot attempts when a given defender is on the ice is. Heck, the location of the origin of a shot attempt when a given defender is on the ice is (and that was a huge recent advance; it's why the concept of shot quality is now being measured and taken seriously rather than written off as being "meaningless" like the hangers-on used to insist a few short years ago). Boxing out an approaching attacker isn't. (But I wouldn't be surprised if proprietary analytics has a way to approximate that, and I also wouldn't be surprised if the hobbyists are working on a way to make that happen.) Hits that successfully lead to puck turnover and a counterattack... well, I think they
could be, but I don't know if they actually
are. (I would
love to see hit location and resulting puck possession changes tracked as a means of gauging the effectiveness of one's physical play.) Stuff like that.