Actually incorrect; his model is focused strongly on puck possession and elements of same, since that's what's readily measurable with publicly available information. Players who do their best work without the puck aren't really measurable, and they try to fake it by inferring measurements based off of what happens when they're on the ice, and attempt to isolate that from everyone ELSE on the ice by paying closer attention to those moments when different Other Folks are on the ice.
I'd be less grumpy about him and others Doing The Best They Can With What They Have (anything more detailed would be extremely labor-intensive and costly, and that's likely a large part of why teams pay their own in-house statisticians) if they were more willing to acknowledge the weak points of those models rather than speak as though they're gospel.
I trust that you know the formula better than I do. My main point though wasn't about what's in the formula, or even that its criteria is "bad" in a vacuum, but that it doesn't evaluate diversity / playing styles that complement each other well; it just looks pretty much at all players under the same criteria in a vacuum instead of how all the players's varying strengths fit together in a puzzle.
Yeah, a line of three Patrik Laine clones is probably not going to be a great line (and probably be an especially bad first line compared to a league-average first line). But having a guy like Laine with players that can complement him 5-on-5 (not to mention having a guy like Laine on the PP) is extremely, extremely valuable and dangerous.
However, on the flip side, also acting like it's some sorta weird "gotcha" that Dom's model has the better teams higher-up seems strange to me. I'm not sure why it's so odd / problematic that the teams that he has higher on his better contracts list tend to be the better team; if anything, that shows his model is pretty good and we should perhaps be a bit more concerned than we are about this upcoming season. But we'll find out soon enough.
rsport.ria.ru
