The Advanced Stats Thread Episode VI: RIP To Our Databases

[DanielBrassard]

KZB were outstanding last night but I am a bit concerned about the rest of our lines. GHM has not worked for a long time, Chytil might not be ready and the 4th line might be OK once Fast comes back. But we really need another top-6 center if we wish to compete and we need to disband the GHM line.

 

[Machinehead]

If we need another top six center, how about try Miller instead of insisting he be with Hayes?

It's literally never worked and AV thinks it's magic.

 

[DanielBrassard]

If we need another top six center, how about try Miller instead of insisting he be with Hayes?

It's literally never worked and AV thinks it's magic.

That would be ok with me but AV loves Miller and Hayes together for some reason.

 

[Oscar Lindberg]

Fast will need time to get back up to speed

Lindberg needed time, everyone wanted him traded when he first came back

Hayes should be with Zucc and Nash

They need another top 9 winger. I don't see Chytil sticking past the 9 games (which I think is the right call in the long run)

Miller can center Grabner, but they need another winger to play with them

 

[Machinehead]

That would be ok with me but AV loves Miller and Hayes together for some reason.

A lot of posters around here also think they're Perry and Getzlaf together, and I'm not sure what everybody is seeing.

They're hard to watch.

 

[Levitate]

A lot of posters around here also think they're Perry and Getzlaf together, and I'm not sure what everybody is seeing.

They're hard to watch.

Isn't it basically because they somehow put up a significant amount of points despite barely shooting and having poor shot analytics?

I mean they generally create some real high quality scoring chances but because they never shoot the goddamn puck except when creating a high quality scoring chance, they're all kinds of screwed up on the analysis scale. Both guys want to be the uber playmaker and look for the play no one else sees rather than shooting the puck. When it starts working out it is magic, but if not then it's watching guys throw the puck away a lot trying for a perfect pass, not taking shots that are available, and playing "meh" defense

 

[Machinehead]

All of that is true, but to me, 65-70 points makes it ok to be an analytics black hole. 45-50 points does not.

 

[TheDirtyH]

All of that is true, but to me, 65-70 points makes it ok to be an analytics black hole. 45-50 points does not.

It seems more important that they continue outscoring their opposition, and considering that neither player plays on the PP, 45-50 points almost entirely at 5v5 or 4v5, is much better than you're coloring it.

I'm also against the line, but mostly because I hate having Grabner in the top 9. You might consider it luck that they performed so well last year, and I'd probably agree. But it's been uncomfortable watching 'PDO' and 'Luck' suddenly become interchangeable terms. Seems a dangerous reduction...

 

[silverfish]

Never mind. I saw everything from 2.2 to 4.5 for xGA on Corsica for a minute there. Will just wait 'til the end of the game

8-5 loss to the Maple Leafs in charts:

 

[DanielBrassard]

That was ugly. All Situtations xG for Toronto was 6.03. 5v5 was 4. Lost the xG battle in both games so far.

 

[Machinehead]

That was ugly. All Situtations xG for Toronto was 6.03. 5v5 was 4. Lost the xG battle in both games so far.

I'm pretty sure we beat Colorado in xG.

Ours was nearly 7 in that game.

EDIT- All situations we did. Even strength, Colorado won.

Powerplay was swarming that night.

 

[Filthy Dangles]

Is there a site that lets you see and sort by team pace? (Cf+ca/60)?

Corsica used to but doesn't look like you can on revamped site.

 

[silverfish]

Our xGA through two games is 9.4 (corsica.hockey).

As usual, under AV, our defensive scheme is completely broken to start the season. Some really bad turnovers yesterday from Smith (kind of surprising) and DeAngelo (not surprising at all), but the forward help on rush shots against has been non-existent.

 

[ManUtdTobbe]

Yeah, while some players have been bad defensively, the biggest issue has definitely been because of the system (or lack there of)

 

[Mac n Gs]

I’ve been a broken record about neutral zone defense for three years and now I’m going to harp on the zone pressure system. It helps you keep numbers down low and have people in position to run a counter attack. No one in this league runs man-to-man anymore. With how teams play swarm offense and use the old Russian tactic of offsetting pressure to open up seams, unless you’re perfectly in sync, you’re gonna get burned. You need forward support to play defense, and with how we’re always trying to get the puck up ice the other way, we’re extra prone to HDSCA.

 

[Filthy Dangles]

Our xGA through two games is 9.4 (corsica.hockey).

As usual, under AV, our defensive scheme is completely broken to start the season. Some really bad turnovers yesterday from Smith (kind of surprising) and DeAngelo (not surprising at all), but the forward help on rush shots against has been non-existent.

Besides a few giveaways what is it? Obviously the stats are bad and we can see with our eyes they are giving up way too much but here's where stats can't help you.

You mentioned forward help on rush shots That's a part of it. Stems from bad offensive zone play/positioning with forwards caught down too low in the zone.

Zib is probably our best center in coverage and breakout. He's really the only center i 100% trust to cover for a defenseman trying to step up on a stretch play or the like

 

[silverfish]

Besides a few giveaways what is it? Obviously the stats are bad and we can see with our eyes they are giving up way too much but here's where stats can't help you.

You mentioned forward help on rush shots That's a part of it. Stems from bad offensive zone play/positioning with forwards caught down too low in the zone.

Zib is probably our best center in coverage and breakout. He's really the only center i 100% trust to cover for a defenseman trying to step up on a stretch play or the like

What is it that you think the stats can't help you with?

We can use stats to:

1. Dissect the quality of shots against, with a probability of being a goal model (ie. xG).
2. We can diagnose shot location; where are we bleeding shots against from?
3. We can use the two above to solve for where we need to plug holes on the ice

No?

 

[silverfish]

DeAngelo out here busting my charts

 

[Mac n Gs]

Free Ryan McDonagh is my favorite Ryan McDonagh.

 

[Machinehead]

I've discovered some wild inaccuracies in HERO charts

Pretty similar, right?

Actual corsiREL over that time period: Kreider - +4.31, Oshie - -0.86. Not in the same league.

 

[Revel]

Silverfish - Would be cool if we could keep your charts together all by themselves in chronological order for the season. Would make a handy, quick reference for evaluating player performance over the course of time.

 

[ManUtdTobbe]

I've discovered some wild inaccuracies in HERO charts

Pretty similar, right?

Actual corsiREL over that time period: Kreider - +4.31, Oshie - -0.86. Not in the same league.

Don't think Dom uses raw CorsiREL, he should have some sort of write up somewhere though.

 

[silverfish]

Silverfish - Would be cool if we could keep your charts together all by themselves in chronological order for the season. Would make a handy, quick reference for evaluating player performance over the course of time.

http://nerdhockey.com/game-summaries-1718/

My God these takes in the PGT are ****ing horrendous. Hank posts a shutout in a 2.53 xGA game == "He was okay".

Nash is attempting 24 shots per 60 at 5v5 & all situations, but he's been "invisible".

Brutal.

 

[silverfish]

Cracknell last year:
50.99 CF% and +1.63 relCF%
48.83 xGF% and -0.72 relxGF%
1.33 p/60 [roughly third line production by my gorilla math]

Overall, it's not a bad claim. He should be a good option for the fourth line.

[Corsica.Hockey]

Would've rather just kept Pirri.

 

[Mac n Gs]

Cracknell last year:
50.99 CF% and +1.63 relCF%
48.83 xGF% and -0.72 relxGF%
1.33 p/60 [roughly third line production by my gorilla math]

Overall, it's not a bad claim. He should be a good option for the fourth line.

[Corsica.Hockey]

Would've rather just kept Pirri.

Eh, Cracknell should help us establish cycles and grind down other teams. This means Desjardins is gone, so this is a win in my books