2017-18 stats and underlying metrics thread [Mod: updated season]

[Jimby]

For the record, I have said a couple of times on this board that I believe the reason for Scheif's lower corsi this year is because he is dragging around Wheeler who is having one of his classic slow starts (albeit with points this year thanks to Scheif) and not because he is waiting for better opportunities.

Here is Scheifele's shot chart from the 15-16 season where he had a very good CF% and a much higher rate of individual corsi attempts. It doesn't really look all that different in terms of shot location than it does for this season that you posted above. This should further raise doubts about the "Scheifele is becoming an anti- Corsi player" narrative

 

[Whileee]

Do the Jets' tend to concentrate their shot attempts to the center of the ice?

Do they tend to force opponents to take more shots from the edges and points?

 

[DoingItCoolKiwi]

20.11


28.11


Jets have been moving to the right direction slowly but surely. Good to see but still a long way to go, especially in the CF department

 

[Jetfaninflorida]

I am not a stats guy. I read on the post game thread that in the last 6 games we are top 10 in both RAW and Score close corsi and that the underlying numbers continuing to trend in the right direction. To me this looks like team performance in the standings is a future predictor of Corsi, not Corsi being a future predictor of performance. Am I wrong and has this been looked at?

 

[Aavco Cup]

I am not a stats guy. I read on the post game thread that in the last 6 games we are top 10 in both RAW and Score close corsi and that the underlying numbers continuing to trend in the right direction. To me this looks like team performance in the standings is a future predictor of Corsi, not Corsi being a future predictor of performance. Am I wrong and has this been looked at?

Wrong correlation. Our numbers for our bottom 6 have improved drastically with Lowry and Perreault back from injury. The Hendricks line is kicking ass with 4th line Jesus

 

[Maukkis]

#AdamForSelke... if he had the offense.

 

[Gm0ney]

I am not a stats guy. I read on the post game thread that in the last 6 games we are top 10 in both RAW and Score close corsi and that the underlying numbers continuing to trend in the right direction. To me this looks like team performance in the standings is a future predictor of Corsi, not Corsi being a future predictor of performance. Am I wrong and has this been looked at?

Is the Jets win-loss record is causing the team to outshoot and outchance their opponents more? Do wet streets cause rain?

 

[Whileee]

#AdamForSelke... if he had the offense.

Jets have plenty of offense. What Lowry brings is a presence that can keep the Jets' top lines away from some tough match-ups. He's a valuable player in a lot of ways.

Lowry's relxGF% is second on the Jets, trailing only..... Matt Hendricks???

 

[Rheged]

20.11


28.11


Jets have been moving to the right direction slowly but surely. Good to see but still a long way to go, especially in the CF department

I'm not sure if Tierny uses all-situations numbers or not but its interesting to me that Corsica has the Jets moving further over to the Dull quadrant while McCurdy's (5on5) has us moving more over to the Fun side of things haha.

 

[Jetfaninflorida]

Is the Jets win-loss record is causing the team to outshoot and outchance their opponents more? Do wet streets cause rain?

So, stats wise why has the team's performance in the standings not been regressing as the stats predicted? And in fact, the performance has been consistently good and now it is the stats that have been improving to come into correlation with the performance. The streets don't cause rain, but if you go to sleep and all the streets are dry, and you wake up and all the streets are wet, there is a good chance that it rained.

 

[mcpw]

So, stats wise why has the team's performance in the standings not been regressing as the stats predicted?

Because stats can only predict the future with certain probabilities, and those 5v5 stats most people are looking at don't predict special teams and goaltending, which we've been good at.

 

[YWGinYYZ]

The streets don't cause rain, but if you go to sleep and all the streets are dry, and you wake up and all the streets are wet, there is a good chance that it rained.

Or, you know ... incontinence.

Personally, I'm happy that certain things are trending in a more 'sustainable' direction in some key metrics. I was always happy with the points, but uneasy about when things might go sideways - recent trends make me a happy Jets fan on both fronts.

 

[Jimby]

Jets have also been very good 5 on 5 for the last month using GF% as a metric - 62.96%. The Jets played great 5 on 5 contrary to what 5 on 5 corsi suggested / predicted. Corsi just doesn't have much use over small sample sizes. Even a whole season can be a small sample size for corsi...

Because stats can only predict the future with certain probabilities, and those 5v5 stats most people are looking at don't predict special teams and goaltending, which we've been good at.

 

[Whileee]

So, stats wise why has the team's performance in the standings not been regressing as the stats predicted? And in fact, the performance has been consistently good and now it is the stats that have been improving to come into correlation with the performance. The streets don't cause rain, but if you go to sleep and all the streets are dry, and you wake up and all the streets are wet, there is a good chance that it rained.

Stats are improving because Scheifele's line is improving in shot metrics and the bottom 6 has solidified. The two go together somewhat because the Lowry line absorbs some tough matchups instead of Scheifele.

Interesting that the bottom 6 seems pretty good without Petan and Dano.

 

[Gm0ney]

Jets have also been very good 5 on 5 for the last month using GF% as a metric - 62.96%. The Jets played great 5 on 5 contrary to what 5 on 5 corsi suggested / predicted. Corsi just doesn't have much use over small sample sizes. Even a whole season can be a small sample size for corsi...

The primary reason for using shot metrics is because goals are relatively rare events and so they're not very useful for predicting the future. You need a sufficiently large sample so you can more easily separate the signal from the noise...and that's why people pay attention to shot attempts (Corsi). Add in things like shot distance, shot type, shooter skill, etc. and you get an even better measure (xG).

I'm not sure why some people are so hostile to the concept that there's a real relationship between goal scoring and outshooting/outchancing your opponents.

 

[Whileee]

The primary reason for using shot metrics is because goals are relatively rare events and so they're not very useful for predicting the future. You need a sufficiently large sample so you can more easily separate the signal from the noise...and that's why people pay attention to shot attempts (Corsi). Add in things like shot distance, shot type, shooter skill, etc. and you get an even better measure (xG).

I'm not sure why some people are so hostile to the concept that there's a real relationship between goal scoring and outshooting/outchancing your opponents.

I think it's because outshooting doesn't directly translate into "outchancing" opponents, and neither have a particularly strong predictive correlation with outscoring opponents. Yes, they are correlated, but there is a lot of "luck" and other unmeasured factors (talent?) that explain a greater proportion of the variance in outcomes than shot metrics by themselves. I expect that all the analytics folks working for various NHL teams are hard at work trying to come up with better parameters and better models to try to explain more of the variance in outcomes (goal share and wins).

 

[Jimby]

Saying that it is not very useful with small sample sizes is not being "hostile to the concept".

I'm not sure why some people are so hostile to the concept that there's a real relationship between goal scoring and outshooting/outchancing your opponents.

 

[garret9]

Saying that it is not very useful with small sample sizes is not being "hostile to the concept".

It's always useful... just the relative degrees change as randomness from noise increases.

 

[garret9]

I am not a stats guy. I read on the post game thread that in the last 6 games we are top 10 in both RAW and Score close corsi and that the underlying numbers continuing to trend in the right direction. To me this looks like team performance in the standings is a future predictor of Corsi, not Corsi being a future predictor of performance. Am I wrong and has this been looked at?

One example is not a trend.
A trend is not the entire population.

The trend is the opposite way as you pointed out, but the trend is not representative of every situation.

 

[Gm0ney]

Saying that it is not very useful with small sample sizes is not being "hostile to the concept".

Taken on its own perhaps not, but you've been consistently mocking the idea that Corsi has any value for months, so I think it's safe to put you in the "hostile to the concept" group.

 

[Joe Hallenback]

Stats are a useful tool for sure but Chayka probably answers the question on them the best

"The more you are in this game you realize the analytics argument is foolish. This is an information game. Decisions on character and other things are hard to quantify, but they matter." Chayka #Coyotes

 

[Whileee]

Stats are a useful tool for sure but Chayka probably answers the question on them the best

"The more you are in this game you realize the analytics argument is foolish. This is an information game. Decisions on character and other things are hard to quantify, but they matter." Chayka #Coyotes

I guess he's had his "Anthony DeAngelo moment"...

 

[Whileee]

I like it when the Jets dominant the shot metrics and still have a high shooting percentage.

This is how you do that....

 

[JetsFan815]

Jets rolling 5 game Corsi For % looking great lately:

So much for the "Corsi is not relevant to the Jets because of selective shooting" argument.

If this keeps up and the Scheifele line can be consistent, the 2nd line figure out whatever is going on with them, and something is done about that PK then this could be a pretty special team

 

[Whileee]

Jets rolling 5 game Corsi For % looking great lately:

So much for the "Corsi is not relevant to the Jets because of selective shooting" argument.

If this keeps up and the Scheifele line can be consistent, the 2nd line figure out whatever is going on with them, and something is done about that PK then this could be a pretty special team

You continue to ignore score effects. Jets have been one of the top teams in the "get a big lead and cruise" category. That is generally associated with lower CF% and higher shooting percentage. They have had a few games lately where they have started from behind, or in close games. Even against the Wild they started off behind, and then cruised past them with a dominating performance.

Scheifele is still a "selective" shooter. Wheeler isn't.

Much of the improvement in shot metrics is the turnaround of the Scheifele line, which I think is partly due to better match-ups now that Lowry is back and his line is absorbing more tough match-ups.

Yes, the Jets are a good team. I'm glad that Corsi is now convincing some folks of that fact.

Jets are still shooting pretty well, even with high Corsi...