GDT: 2023-24 season Game 2 LA Kings vs Carolina Hurricanes @7:30pm 10/14/23

[Raccoon Jesus]

There is a "0-axis" which essentially shows the average. You can eyeball it by seeing essentially where all the bars consistently split, or you can look at the bottom of the chart and find the 0. Bars to the left of it say the contributions of that particular area (offense, defense, production, and miscellaneous) were below average. Bars to the right of it are above average. The scores then get "added".

Fiala's production of 3 assists definitely boost his "production" area, as well as Kopitar's 2 goals. So, you can see the lighter blue being much longer than what anyone else. I don't know exactly what Dom's model utilizes to measure each component, but it's to give examples.

The higher up on the chart, the more positive your contributions on the ice were overall.

So, this chart is saying Fiala's contribution of offense, defense, production, and miscellaneous was better than anyone else on the Kings. And that Jordan Spence was the worst.

It's based on data that is returned from some API call to the NHL statistics gathered. So, Spence keeping the puck on the powerplay multiple times because they were clearing attempts has zero impact on this model, because that sort of data isn't tracked. A bad pass to a teammate who can't control it may show that the teammate has a giveaway, instead of the original passer who set up the teammate poorly.

Then, of course, the model puts different weight on different things - for example, if Kevin Fiala gives the puck away on the powerplay that leads to a shorthanded goal against, it will have just as much weight as a player making an errant pass in the offensive zone that gets picked off. A "giveaway" without any additional context may have smaller impacts.

This is probably more than what you asked for. But the way this game went exposes some serious flaws in Dom's model of calculating a "raw game score" based on how statistics are logged in the game. It's part of why using raw analytics is a very unwise approach. BUT, it can be a good tool when used properly to evaluate and measure what you have seen, because implicit biases from the eye test can also be flawed.

A little bit of an aside maybe but kind of a furthering, this is also why Martinez-muzzin looked like one of the best pairings in the world by stats but worst in actual results

Some of these stats are good at capturing play volume but not catastrophic moments

Muzzin Martinez could possess the puck, play d, drive play forward. But then they’d pass the puck directly to the other team for a tap in as well. Erik karlsson is another great example.

Spence above is an example of the counter

But they’re useful as a mode of inquiry as long as you can void confirmation bias

 

[Frolov 6'3]

This whole GDT was filled with comments how much we sucked but a former elite player says otherwise.

 

[Sol]

I remember the GDT being hyper critical of the turnovers being made because the goalie couldn’t stop a thing.

It’s funny that those saying the Kings looked good but the goaltending sucked are also the same who think Blake has done nothing wrong

 

[johnjm22]

Carolina's expected goals was 2, but they scored 5.

10 goals given up in two games.

8th straight loss to Carolina, BTW.

 

[Axl Rhoadz]

I remember the GDT being hyper critical of the turnovers being made because the goalie couldn’t stop a thing.

It’s funny that those saying the Kings looked good but the goaltending sucked are also the same who think Blake has done nothing wrong

Get some new material, bud.

 

[kingsholygrail]

Turnovers are a problem. Goalie's gotta make a save at some point too.

 

[Telos]

That game was such a rollercoaster. There were a few small positives but the negatives are glaring at the moment and I am just hoping they can build upon them and regain form as the season goes. It may be early, but these points are going to matter big down the stretch.

I don't really want to harp on it all, but goaltending is easily the Kings' biggest issue that anyone and everyone can see from space. I think Blake and Luc's dream of just getting to the playoffs and rolling the dice to see if one gets hot enough to just be good enough alla Vegas is fantasy. I expect a deal before Christmas to be honest. It would likely be the most bent over backward, rage-inducing, trade of Blake's tenure, but he has to make it eventually this season.

Spence looked quite good out there, was very much in control and confident in his play, and looked like a natural out there 3v3. Byfield is trending up, but it won't matter if he doesn't develop the finish and can consistently produce. He will have the full season to put all doubters to shame though and hopefully he does it.

 

[johnjm22]

I think Blake and Luc's dream of just getting to the playoffs and rolling the dice to see if one gets hot enough to just be good enough alla Vegas is fantasy. I expect a deal before Christmas to be honest. It would likely be the most bent over backward, rage-inducing, trade of Blake's tenure, but he has to make it eventually this season.

I don't think that was ever the plan though. I think they've always considered Talbot/Copely place holders. They have 5 months to find something that works in net.

 

[SettlementRichie10]

The meme season is off to a good start.

 

[Trash Panda]

I don't think that was ever the plan though. I think they've always considered Talbot/Copely place holders. They have 5 months to find something that works in net.

5 months, and little-to-no cap space to make it work.

It’ll be another high-acquisition cost trade, because they simply do not have the time to throw this season away. Their roster construction dictates that they win before Doughty and Kopitar fall further off the cliff they are headed down. It’s a poor plan, and it’s fully in motion.

This years first is very much in play. Gonna be a low-event draft in June for the Kings, I wager.

 

[Telos]

I don't think that was ever the plan though. I think they've always considered Talbot/Copely place holders. They have 5 months to find something that works in net.

I think it was a bit of both. I think they were honestly hoping that they would placeholder their way into sliding by in the postseason while knowing they had the regular season to try to make a deal if they had to, which is sort of the situation they forced themselves into. Their interviews would look pretty stupid in hindsight if they had any other plan.

The problem is that at this rate, they are going to be in the same boat as last season, where you may have to make a deal sooner rather than later and I am not sure you can hope for a Copely 2.0 miracle to save your season just so you can even make the playoffs. If it continues, they are at risk of missing the playoffs which ends everything.

 

[Sol]

I think it was a bit of both. I think they were honestly hoping that they would placeholder their way into sliding by in the postseason while knowing they had the regular season to try to make a deal if they had to, which is sort of the situation they forced themselves into. Their interviews would look pretty stupid in hindsight if they had any other plan.

The problem is that at this rate, they are going to be in the same boat as last season, where you may have to make a deal sooner rather than later and I am not sure you can hope for a Copely 2.0 miracle to save your season just so you can even make the playoffs. If it continues, they are at risk of missing the playoffs which ends everything.

I don’t understand this olive branch you are extending. They got Korpisalo because they knew Copley was a flash in the pan. Most of us knew that. Talbot hadn’t been a competent goalie for a while. They are arguably in a worse position than last year because at least with last year the season started with us not really knowing that Quick had fell off the face of the earth completely.

What was even the point of getting Korpisalo if you weren’t going to stick to him? He’s easily better than what we have now. I feel like we’re in the same exact position as last year and what makes it worse is that we knew that this season our goaltending was going to be the worst. Last year you could argue we went into the season unsure of what to expect.


The inaction of this season is worse easily because we all knew the goaltending was going to be atrocious but they still went with it.

Any GM who knowingly puts him team in a position where they have to outscore their problems is a GM that needs to be fired. Blake got PLD knowing that the Kings were going to become that team for certain when he locked him in.

 

[Statto]

Get some new material, bud.

I’m sure you’ve posted this many times yourself

 

[Sol]

I’m sure you’ve posted this many times yourself

It be like that.

 

[King'sPawn]

A little bit of an aside maybe but kind of a furthering, this is also why Martinez-muzzin looked like one of the best pairings in the world by stats but worst in actual results

Some of these stats are good at capturing play volume but not catastrophic moments

Muzzin Martinez could possess the puck, play d, drive play forward. But then they’d pass the puck directly to the other team for a tap in as well. Erik karlsson is another great example.

Spence above is an example of the counter

But they’re useful as a mode of inquiry as long as you can void confirmation bias

Agreed. The cool thing about analytics is someone who develops a model and is responsible could reevaluate their model.

I was trying to work on a scouting app that could help contribute to things like this, but it's been many years since I've done any coding; I also just get distracted by writing and other projects.

I don't know if these models also pull the time of these events, or just the raw numbers. But I'd like to think if a player does a giveaway that leads to a goal or shot on goal within x seconds, that should have a bigger negative contribution to a defensive score, as opposed to a giveaway in the offensive zone.

But again, this always requires:
- a thoughtful model
- a reactive model designer who can reevaluate based on feedback
- the appropriate data to tell the story you're trying to tell
- responsible and honest interpretations of tbe results

 

[bland]

Carolina's expected goals was 2, but they scored 5.

10 goals given up in two games.

8th straight loss to Carolina, BTW.

They had two breakaways.

Are the statisticians trying to tell us that those were the only two "expected goals" in the entire game?

 

[johnjm22]

They had two breakaways.

Are the statisticians trying to tell us that those were the only two "expected goals" in the entire game?

One breakaway does not equal 1 expected goal.

Players don't score on 100% of breakaways.

 

[AbsentMojo]

Any GM who knowingly puts him team in a position where they have to outscore their problems is a GM that needs to be fired. Blake got PLD knowing that the Kings were going to become that team for certain when he locked him in.

Blake did this last year as well.. with giant holes at G and LD and hoping to slog through it. He got lucky that Copley was good. And to think you can have a half assed roster for 3/4 of the season and fix the problems at the deadline is a crap shoot. Because for one, you dont know how the new guy(s) will mesh and secondly having to change your system to a more playoff oriented approach - after needing to play run and gun due to bad goal tending. The more prudent and logical approach wouldve been to move Iafallo or RV for a goalie.

 

[Statto]

Any GM who knowingly puts him team in a position where they have to outscore their problems is a GM that needs to be fired. Blake got PLD knowing that the Kings were going to become that team for certain when he locked him in.

Isn’t out-scoring your opponent the only way to win?

 

[GoldenBearHockey]

Isn’t out-scoring your opponent the only way to win?

Dude, look who you are responding to.....

 

[tigermask48]

Needing 5-6 goals a game seems unsustainable over an 82 game season no matter how good the rest of the players are.

 

[BaileyFan]

5 months, and little-to-no cap space to make it work.

It’ll be another high-acquisition cost trade, because they simply do not have the time to throw this season away. Their roster construction dictates that they win before Doughty and Kopitar fall further off the cliff they are headed down. It’s a poor plan, and it’s fully in motion.

This years first is very much in play. Gonna be a low-event draft in June for the Kings, I wager.

Nah. 1st + 2nd to Ottawa for Tarasenko with retention as an Arvidsson replacement then run the most electric top 9 in the league when he returns. Goaltending doesn't sell tickets.

 

[Bandit]

In regards to goaltending… there is no goaltending.

 

[kinghock]

In regards to goaltending… there is no goaltending.

You are wrong. Kings have 3 old garbage goalies on the roster including Reighn

 

[kingsfan28]

There is a "0-axis" which essentially shows the average. You can eyeball it by seeing essentially where all the bars consistently split, or you can look at the bottom of the chart and find the 0. Bars to the left of it say the contributions of that particular area (offense, defense, production, and miscellaneous) were below average. Bars to the right of it are above average. The scores then get "added".

Fiala's production of 3 assists definitely boost his "production" area, as well as Kopitar's 2 goals. So, you can see the lighter blue being much longer than what anyone else. I don't know exactly what Dom's model utilizes to measure each component, but it's to give examples.

The higher up on the chart, the more positive your contributions on the ice were overall.

So, this chart is saying Fiala's contribution of offense, defense, production, and miscellaneous was better than anyone else on the Kings. And that Jordan Spence was the worst.

It's based on data that is returned from some API call to the NHL statistics gathered. So, Spence keeping the puck on the powerplay multiple times because they were clearing attempts has zero impact on this model, because that sort of data isn't tracked. A bad pass to a teammate who can't control it may show that the teammate has a giveaway, instead of the original passer who set up the teammate poorly.

Then, of course, the model puts different weight on different things - for example, if Kevin Fiala gives the puck away on the powerplay that leads to a shorthanded goal against, it will have just as much weight as a player making an errant pass in the offensive zone that gets picked off. A "giveaway" without any additional context may have smaller impacts.

This is probably more than what you asked for. But the way this game went exposes some serious flaws in Dom's model of calculating a "raw game score" based on how statistics are logged in the game. It's part of why using raw analytics is a very unwise approach. BUT, it can be a good tool when used properly to evaluate and measure what you have seen, because implicit biases from the eye test can also be flawed.

Thanks, very understandable now.