Post-Game Talk: Canucks Lose 4-3 (OT) vs Wild

[PuckMunchkin]

He missed one week on ice work. Shouldn’t have affected his skating or cardio

We have different understanding of what surgery does to a pro athlete.
Edit.

Sorry I dont know why I took this bait.

Lets just not do this f***in dance.

 

[mossey3535]

Another good effort, another poor performance by our 'star' goaltender. The Canucks have played well as a team, they just need some consistency in net. Last year Demko started really slowly and this year it has been the same. Similarly, the Canucks have reverted (albeit less so) to trying to be a more "defensive-minded" team. Know your personnel. Their key defensive pieces: Demko, Hughes, Myers (and Rathbone) are all much better in open, end-to-end play than when tasked with structured defense.

Here is something I posted at the start of last year when the Canucks were off to a similar slow start, especially with special teams, that I think still applies today.

"[Demko] has been a middling goalie this year in terms of goals saved above expectations. Very good at ES, but very poor on the PK. I am tired of him being touted as a top goalie because of highlight reel saves when the underlying results just don't back it up. It is a classic Vancouver media dogpile both for Demko and anti the rest of the team.

No I am not saying he sucks overall, or that he has been costing the team games, just trying to bring everyone back to reality a little.

Demko, like Markstrom before him and DiPietro to come, is an athletic goaltender that thrives on persistence and instinctual saves and plays his best hockey when he is able to reign in the movement and settle his game. He is best suited for a run-and-gun style of play where he is involved in the game, faces lots of shots and typically overperforms when facing prototypical "high danger" chances, but fares far more poorly (to the point of being below average) on "low-danger" chances like point shots and shots with minimal pre-shot movement. Both he and Marky seemed to struggle with rebound control, often making dramatic saves, only to punt rebounds into the goal-front, then make another dramatic save again. Both Demko and Markstrom have had a penchant for making ridiculous saves and later giving up soft goals, often in the same game.

A goalie who is good against high danger plays will still concede mostly high danger goals since that is how most goals are scored (NST has it as 55%, or 6463/11854 since Demko has entered the league). He will just concede fewer than he should. Prior to this game, Demko had conceded 112 high danger goals out of 203 total since 2018, for 55%, so he is normal in terms of counts there.

Vancouver gives up lots of high danger opportunities against and Demko saves more than his share. The stats bear that out, his HDGSAA is +5.96 in that time (18th for GP>20) and he ranks as one of the better goalies in the league in goals saved above expected over the last three seasons. Thing is, his LDGSAA is a paltry -0.92 and 43rd in that time. These aren't earth-shattering differences, but given the hypothesis that he is a good goaltender in general, one would expect his low danger goaltending to be just as strong as his high and medium (where he truly shines at +11.07) danger goaltending, which it just isn't. He has shown an improvement in all facets of his game over time, yet his LDGSAA numbers still lag far behind his strong play against high danger chances (he sits 64th at -1.59 all situations).

I think it likely a combination of athleticism over poise, Clark's coaching techniques and the Canucks style of play that lead to that.

It is one thing to have great GSAx numbers based on saving all the difficult shots, but it seems more valuable to have a less athletic, but more consistent goaltender who makes the easy saves and gives up fewer rebounds and therefore does not artificially inflate their own GSAx.

Also, if the Canucks have recognized that their goaltender is elite at making saves off of odd man rushes and those sorts of "high danger" chances, it absolutely behooves them to really open up play, forecheck aggressively and try to trade chances, especially if they figure that the opposing netminder is better suited to in-zone defense. Not doing this has been one of Green's biggest coaching failings this season.

For the record, I think the Canucks' management has absolutely targeted players that fit that style of play. Tyler Myers is not good if you want low-event, safe hockey, but great if you hope to trade chances and create off the rush since he excels in neutral zone lateral puck movement and at facilitating controlled entries on offense. The tradeoff is that he is poor defensively off the rush as well as in-zone, but I would imagine that their analytics team liked him for his rush-offense profile enough to gamble on him anyway.

In short:
He is a frustrating player to watch, because he clearly has the potential to be a great (and maybe Vezina calibre) goalie, but is not doing the easy stuff as well as a player with his talents should. Like Markstrom, who has cut out some of the inconsistencies from his game, so too may Demko evolve and become more poised in the future, and like his former mentor, legitimately deserve to be in Vezina contention."

I think you're basing too many conclusions on stats in this case. Although rush chances resulting in, say, a shot from the slot are rated as "higher danger", there's no granularity to it. Overall HDSV% can't tell you that the goalie is good at odd man rushes. Saying that the team in response to high HDSV% should trade chances, and that is a rationale for having Tyler Myers is misguided in my opinion.

Also, it's been shown that HDSV% is a better predictor of overall traditional Sv%, not LDSV% so your general premise regarding "steady" goaltenders has some flaws.

If the team on the whole prevents scoring chances, this means the goalie faces less HD shots. HD shots are more likely to score, so it doesn't make sense to give up chances because your goalie is like 1-2% better than his peers on average.

Also, IMO you misread his strengths. When he is set and on angle, he basically can't be beat on a straight up shot. As you pointed out, he is great at athletic improvisational plays. But I wouldn't want the team to give up lots of rush chances, let alone odd-man rush chances since his CONTROLLED lateral movement has oscillated wildly.

So why does he give up LD shots?

-Our defence is bad and has a long history of screening their own goalie
-Demko right now is probably below average at fighting through screens
-He has a bad tendency to lean towards his posts and also to go down into post integration too often and too early. This means he gives up too many bad angle shots, which are definitely LD.
-His play reading is average at best, so he is "tricked" sometimes
-He has a known "blind spot" glove side mid-range bad angle where he instinctively blocks and doesn't seem to react.

I do agree that Clark has a lot of work to do in order to calm down his game further and address some glaring technical issues. IMO Clark emphasizes close range and desperation practice too much, and it has led to Demko being very confident and over-reliant on down movement in his crease. But Clark has also said he wants an athletic goalie with tons of tools and that he can work on their technique easily. While I dispute this assertion, it's well known that his MO is to take raw, big goalies and calm down their games. So that's another flawed premise of yours.

Overall I do think Demko is somewhat frustrating, so I agree with your conclusion. I just disagree with how you got there.

 

[tyhee]

Don't know what to say because what most people are seeing is that they played well 5on5, not dominant just better than their opponent, and the numbers have backed that up.
...

I'm not so sure that "the numbers have backed that up."

Raw shot attempt numbers are in the Canucks' favour so that the CF% and FF% at 5 on 5 looks favourable, but they're getting beaten in high danger chances to the extent that right now looking at the Team Analytics at 5 on 5 section for the Canucks on hockey-reference.com, it has at 5 on 5 an expected goals deficit, 9.9 for and 10.5 against.

Evolving Hockey is slightly kinder to the Canucks, having them dead even for expected goals for at 5 on 5 after 5 games of the season with an expected goals for of 0.5.

I don't see that Hockey Reference has those numbers for 4 on 4 or overall even strength. Evolving Hockey does and while it shows the Canucks dead even at 5 on 5, they show an expected goal deficit at 4 on 4 and at even strength.

Those numbers don't suggest the Canucks have been better than their opponents overall at 5 on 5, being slightly outplayed or at best even. When all even strength situations are taken into account the numbers have the Canucks with a small deficit.

It would be interesting to take score effect into account. I don't know where to find those figures but with the Canucks having had leads as much as they have, it wouldn't be surprising if stats suggest the Canucks have been the better team at even strength when the score is even.

 

[1440]

I think you're basing too many conclusions on stats in this case. Although rush chances resulting in, say, a shot from the slot are rated as "higher danger", there's no granularity to it. Overall HDSV% can't tell you that the goalie is good at odd man rushes. Saying that the team in response to high HDSV% should trade chances, and that is a rationale for having Tyler Myers is misguided in my opinion.

Also, it's been shown that HDSV% is a better predictor of overall traditional Sv%, not LDSV% so your general premise regarding "steady" goaltenders has some flaws.

If the team on the whole prevents scoring chances, this means the goalie faces less HD shots. HD shots are more likely to score, so it doesn't make sense to give up chances because your goalie is like 1-2% better than his peers on average.

Also, IMO you misread his strengths. When he is set and on angle, he basically can't be beat on a straight up shot. As you pointed out, he is great at athletic improvisational plays. But I wouldn't want the team to give up lots of rush chances, let alone odd-man rush chances since his CONTROLLED lateral movement has oscillated wildly.

So why does he give up LD shots?

-Our defence is bad and has a long history of screening their own goalie
-Demko right now is probably below average at fighting through screens
-He has a bad tendency to lean towards his posts and also to go down into post integration too often and too early. This means he gives up too many bad angle shots, which are definitely LD.
-His play reading is average at best, so he is "tricked" sometimes
-He has a known "blind spot" glove side mid-range bad angle where he instinctively blocks and doesn't seem to react.

I do agree that Clark has a lot of work to do in order to calm down his game further and address some glaring technical issues. IMO Clark emphasizes close range and desperation practice too much, and it has led to Demko being very confident and over-reliant on down movement in his crease. But Clark has also said he wants an athletic goalie with tons of tools and that he can work on their technique easily. While I dispute this assertion, it's well known that his MO is to take raw, big goalies and calm down their games. So that's another flawed premise of yours.

Overall I do think Demko is somewhat frustrating, so I agree with your conclusion. I just disagree with how you got there.

I made conclusions and then looked to the stats to back them up. If you don't think the picture they paint leads to the conclusions that I have drawn that is fair. That said, I put more trust in what I intuit from watching every minute of Demko's play than what the stats say (as opposed to most other goalies, for whom I would trust the stats) - as you say the granularity in the stats isn't there.

One example of the stats not matching my intuition is that I feel that Demko gives up more scoring chances against on in tight and low angle (as in close to the goal line shots) rebounds than the average goalie would. He then saves a lot more of these rebounds (which are both legitimately dangerous opportunities and high danger chances against based on how most models weigh rebounds and shots proximal to the goal) than the average goalie would, which ultimately inflates his apparent value as a goalie both in the eye and stats tests. Perhaps your point about post integration would tie into this. Unfortunately for me, this just doesn't match the stats and most suggest that he does not give up more HD rebound chances than the average goalie. I would also suggest that Demko is above average in dealing with other types of shots, which has the effect of normalizing his rebound stats overall. Of course, the capabilities of the defenders in such situations are a factor, and better defenses would arguably suppress these rebounds regardless of how many are produced. Intuition here is that the Canucks' defense is lackluster in this capacity, which is another wrinkle in my theory.

I will admit I know very little about goalie mechanics so I will defer to you on those sorts of topics. That said, I think we agree more on some things than you suggest.

For one, I don't think a team should lay out "trading chances" as an organizational strategy and acquire Tyler Myers to fulfill that strategy. That is putting the cart before the horse. Tyler Myers has his strengths (his ability to generate/suppress east-west passes as a defenseman off the rush is exceptional [I can't provide a reference here because the link is dead, but Alex's Novet's 2019 xG model saw him as one of the largest risers in XGF% when it added pre-shot movement to the model], and makes sense given his reach and skating ability), but is overall a niche/specialist player - not a build around (I still think he should be deployed as a forward, but that is another story). That said, when the previous regime has handed you a bunch of "lemons", a good coach should figure out how to deploy them to make "lemonade". Demko fits that because as you point out, his ability to fight through screens is poor and there aren't many screens happening off of odd-man rushes. Ideally the Canucks would just acquire and deploy more well rounded defensemen and not have to live and die with their goaltending, or play some risky style to succeed.

Maybe it is another case of cart before horse in my blaming Clark's coaching for some of Demko's technical issues. It would make sense for Markstrom too that the perception that he was constantly bailing the Canucks out has more to do with Clark's desire to start with big athletic goalies and calm down their play, therefore producing an environment where chances against the Canucks more often generate athletic-looking, flashy saves and everyone in media/social networks goes "wow the Canucks goalies just keep bailing them out". It is difficult to tease out the points that reject the popular premise that the Canucks are/were with Clark as coach, a team with good goaltending and bad everything else, but I do think that the philosophy behind Clark's coaching generates an environment that over-accentuates goalie value and makes them look good at the expense of the defense (which, incidentally is a great thing to do if you want to keep your goalie coaching job).

The Canucks market is such a echo-chamber that it generates this feedback loop where a premise is dreamed up, gets talked about in the media, gets echoed in the play-by-play (John Shorthouse has a lot more influence than he himself might realize and he needs to be more careful with his wording), and then gets regurgitated by fans, media, broadcast that it is exceptionally difficult to dissuade the average fan that they should be critical of what seems like consensus. Thanks for your nuanced approach to this topic.