This is my mailbag, where I answer your questions posed to me on Twitter (@JFreshHockey). Visualizations used in these answers are available to Patreon subscribers.
John: Why are the Jets ahead of the curve when it comes to forbidding their Dmen to do anything, and why hasn’t the rest of the league copied them?
It really is remarkable. I would recommend that readers check out Jack Han’s system analysis of the Jets because it really puts a fine point on just how little defencemen are allowed to do on that team and how much of an extra burden it places on the forwards (and goaltenders). It’s a copycat league, so maybe if they make a run we’ll see teams will crummy defencemen run a similar system.
Jack: I've noticed that Barkov's RAPM charts actually reflect his reputation in terms of defense for once. Have you gotten the chance to look at some film, and if so what has or hasn't changed to reflect his improvement defensively?
Not in too much detail, but I suspect that being separated from Jonathan Huberdeau and embedded in an apparently more stingy system has allowed him to refocus his clearly strong defensive talents and put them to proper use. That Barkov - Huberdeau line was the “scoring” line in previous years, and no longer having to carry that load has paid off early on for my Selke favourite.
Versace: Any teams that you thought would be good before the season started that have turned bad for you? And vice versa?
Almost disappointingly, this season has been mostly chalk. No huge surprises either way so far. I think most people will get their pre-season predictions mostly right. The biggest disappointments for me so far are the Blue Jackets and the Penguins, both of whom I thought could be at the top of their respective divisions and both of whom will have to claw their way into a playoff spot.
Doooberry: Where do you rank McAvoy among the best defensemen?
#1 or #2. He’s for real. He was my Norris pick last season and I have him just behind Makar so far this year. If he scored a few more powerplay points per season he would be viewed as the top defenceman in the league.
DaavLuuv: Why don’t we see Marner at centre?
The main answer is that the Leafs have two centres making $22M+. The second is that he hasn’t played any centre since junior, right? Not sure if the Leafs feel they can afford to try out their top winger at a new position where he might struggle.
Ned: What player has impressed you most this season defensively?
Jack Eichel has the best defensive numbers in the league. No idea how, but that’s just the way things are right now. So I guess it’d have to be him.
MiroHeiskanenFan: Miro Heiskanen has yet to score a goal this season. Is this something that is chalked up to just bad luck, or is there something within his analytics that show he’s playing differently than previous years?
Of course Miro rewarded you by scoring a goal after you asked me that question. His stats still aren’t very good (especially offensively) but considering how boringly the Stars play maybe there’s a system thing going on here. Macro-level analytics have never been as strong on him as his eye test or reputation.
Rye: Why doesn't the NHL just vaccinate their 600 or so players so they can keep them safe & stop postponing games? Seems like a trivial number considering tens of thousands of doses are given out daily.
I believe there was some PR issues surrounding a rumour that the NHL was looking to secure doses for all of itss players before the season started and they had to promise that they weren’t going to hop the line.
Luke: Colorado doesnt seem to be playing at the level they should, is this just a regression or is this more of the normal?
Last year the Avs had a 52% xGF% (8th) and a 59% GF% (1st by far). They were always going to regress at least a bit, and from what I can tell their plan this offseason was to improve enough that the blow would be cushioned. That plan is partly working, as their xGF% is higher (54%) but they’re not finishing like they should and their non-Grubauer goaltending hasn’t held up, giving them a 52% GF%.
They’ll be fine.
AJ: The consensus has been that Wahlstrom has become a much better two-way guy than when he had been drafted. Do his underlying numbers so far this season show that he is good on both sides?
That kind of thing can be tough to tell for a player who plays on a system as strong as the Islanders’ is defensively (only Dallas and Florida have a lower xGA/60). That being said, early results look pretty good for Wahlstrom!
Kleven: Artem Zub, how good is he and what should his extension be?
So far so good.
If he can stick to being a low-event king then the Sens should be very happy with him. As for an extension I can’t really help you there because I’m not sure what the expected price tag is.
Goofball: What would you offer Ovechkin on his next contract?
Whatever he wants. It’ll be a long road back for the Caps once he retires and IMO keeping him chasing 800 is more valuable than whatever else they’d spend that money on from a marketing perspective. Even if it’s not the “analytically sound” decision, I think they should sacrifice a little to make sure he retires a Capital.
CMac: What are your thoughts when players say this:
Mark Scheifele full of fun turns of phrase today. Called analytics "hogwash", saying he doesn't have time for them and in fact doesn't understand them. He'd much rather watch a game and tell you who played well.
It doesn’t bother me in the slightest. As Garret Hohl said yesterday, analytics are useful at the decision-maker level when it comes to team construction and roster management. If Mark Scheifele looks at a heat map that says he’s bad defensively that does nothing to help him improve. It’s up to the coaches to interpret statistical information they’re given and translate that into actionable training.
Put it this way: if there was a model that measured how good you are at your job and it said you were overrated and not very good at it, you’d think it was bullshit too.
Way Below: What do the numbers say about Elias Pettersson's season? At the start of the year, he seemed to settle on the perimeter in the O-zone and not be as involved down low in the D-zone. The last ten or so games, he looks determined to get to his spots. Also, how unlucky with posts?
Pettersson’s numbers have dramatically improved. They’re not elite in any category right now but he has definitely taken a big leap like you said to have a more positive impact on the game. Those posts have definitely hurt his finishing numbers and I would imagine that more goals are coming.
Another factor relevant to this model is that as new information and combinations played out, it became clear that J.T. Miller was the major weak link on the first unit and he is now the one mostly getting dinged for its struggles early on.
RockyHockey: Makar is great, duh. Is Girard elite too?
No. No he is not.
In the tweet linked, the EvolvingWild twins named their Norris finalists based on what I presume are the early results of their xWAR model - this is a fun thing they do that shows off their models and simultaneously pisses a whole lot of people off who don’t understand that they’re doing that. Girard right now is basically the perfect example of a player who would be “overrated” by that particular model, because
He is finishing and scoring at a crazy rate (13% sh% compared to a 4% career average)
He has been impacting on-ice quantity of shot attempts to a far far greater extent than on-ice quality.
Since xWAR mostly factors in a combination of impact on shot rates and finishing above expected, it’s no surprise that he ranks at the top for them. I personally see his average expected goal driving at both ends and luck-inflated finishing and see no reason to change my view which is that he’s a perfectly fine top four defenceman who’s overrated because he does some very flashy and fun to watch things on the ice.