Jets Advanced Stats thread

[Whileee]

CF% and their so-called predictive value of future success

Let me understand the logic of this. Having a high historical CF% is predictive of future success because:

1. If you have a high CF% historically, it makes sense that you can continue to have a high CF% in the future
2. If you outshoot your opponents over a period of time, you will likely outscore your opponents over a period of time.
3. If you outscore your opponents over a period of time, you will likely have a higher victory % over time.

Is this the logic? If so, to prove out the predictive value of CF%, we should be going through each of these bullets rather than jump from having a high current CF% to future wins. The logic first needs to go from having a high current CF% to a high future CF%, right?

I may take a look at the CF% at the 20 game mark and see how well the logic holds but first, let's agree on the logic beforehand.

I looked at this a couple of years ago, and others have looked at it. CF% in the first 25-30 games is correlated with CF% and GF% for the remaining 55 games, or so.

Here are the analyses of CF% up to December 8 (about 25-30 games) vs. CF% and GF% for the rest of the season for all teams, 2014-2019 seasons. Interestingly and perhaps not surprisingly, the best predictor of future success (from December 8 until the end of the season) is points%. Point% at December 8th is more highly correlated with points% for the rest of the season (data from 2016-19).

Data are from Natural Stat Trick. I think they are 5v5 SVA, but I could be mistaken about the selection.

 

[Whileee]

The NHL tracking data is interesting. JFresh has been looking at speed bursts.

Jets have a number of forwards that rank highly, but the Jets' have 5 D in the bottom 100, and only 1 D in the top 100. Interestingly, the only D with a relatively high rate of fast bursts is Samberg, perhaps because he's very fast in his puck retrievals, giving him more time to make a play. Jets D need to play faster, I think, especially getting back in puck retrievals.

 

[Buffdog]

The NHL tracking data is interesting. JFresh has been looking at speed bursts.

Jets have a number of forwards that rank highly, but the Jets' have 5 D in the bottom 100, and only 1 D in the top 100. Interestingly, the only D with a relatively high rate of fast bursts is Samberg, perhaps because he's very fast in his puck retrievals, giving him more time to make a play. Jets D need to play faster, I think, especially getting back in puck retrievals.

View attachment 761448

O

I didn't realize they tracked this. My first thought when it came to D is that a major reason for needing to skate very fast is because you got out of position and you're chasing the play

I could also see if you're jumping into the rush

Edit: obviously you'd need the ability to reach the threshold speed while trying to get back into position. The D who get out of position and don't skate well enough to re-establish it are obviously the worst option

 

[KingBogo]

The NHL tracking data is interesting. JFresh has been looking at speed bursts.

Jets have a number of forwards that rank highly, but the Jets' have 5 D in the bottom 100, and only 1 D in the top 100. Interestingly, the only D with a relatively high rate of fast bursts is Samberg, perhaps because he's very fast in his puck retrievals, giving him more time to make a play. Jets D need to play faster, I think, especially getting back in puck retrievals.

View attachment 761448

View attachment 761449

It is an interesting stat but what does it really tell us about a player’s impact. For example Adam Fox a Norris trophy winner is near the bottom and you have fringe 6/7 defensemen much higher. For example Bogosian is at 44 compared to Fox’s 190.

 

[surixon]

I didn't realize they tracked this. My first thought when it came to D is that a major reason for needing to skate very fast is because you got out of position and you're chasing the play

I could also see if you're jumping into the rush

Edit: obviously you'd need the ability to reach the threshold speed while trying to get back into position. The D who get out of position and don't skate well enough to re-establish it are obviously the worst option

It's one of the things that came with the puck and player tracking chips the NHL implemented a few years ago.

It's interesting but imo on its own doesn't say much.

It is an interesting stat but what does it really tell us about a player’s impact. For example Adam Fox a Norris trophy winner is near the bottom and you have fringe 6/7 defensemen much higher. For example Bogosian is at 44 compared to Fox’s 190.

On its own it says nothing as players play at different paces. You can get your cerebral players like Fox, JoMo, Perfetti etc that read the play ahead of their peers and as such aren't chasing the play often. They don't need to burn energy to be effective. Others like Ehlers need to push the pace to be effective. Then there are the bad NHL players who are out of position and have to chase the play and bust due to that.

 

[KingBogo]

It's one of the things that came with the puck and player tracking chips the NHL implemented a few years ago.

It's interesting but imo on its own doesn't say much.

You could almost make a counter argument that as a defensemen the less times he have to accelerate to a certain speed the more likely you are to be in an optimal defensive position.

It's one of the things that came with the puck and player tracking chips the NHL implemented a few years ago.

It's interesting but imo on its own doesn't say much.



On its own it says nothing as players play at different paces. You can get your cerebral players like Fox, JoMo, Perfetti etc that read the play ahead of their peers and as such aren't chasing the play often. They don't need to burn energy to be effective. Others like Ehlers need to push the pace to be effective. Then there are the bad NHL players who are out of position and have to chase the play and bust due to that.

I just made a similar comment.

 

[Guffman]

I looked at this a couple of years ago, and others have looked at it. CF% in the first 25-30 games is correlated with CF% and GF% for the remaining 55 games, or so.

Here are the analyses of CF% up to December 8 (about 25-30 games) vs. CF% and GF% for the rest of the season for all teams, 2014-2019 seasons. Interestingly and perhaps not surprisingly, the best predictor of future success (from December 8 until the end of the season) is points%. Point% at December 8th is more highly correlated with points% for the rest of the season (data from 2016-19).

Data are from Natural Stat Trick. I think they are 5v5 SVA, but I could be mistaken about the selection.

View attachment 761437View attachment 761445View attachment 761444

Thanks, Whileee. It is VERY interesting that you mention point % is the best predictor of future success.

I am not sure if you are aware of the ground breaking study I did a few years ago, which still holds up today. The study established a new metric called The Guffman, which is essentially goal differential. Standings, which is based on point %, is highly correlated to The Guffman statistic. If you look at league standings at any point in time, you will see the top teams having the highest Guffman, and the worst teams having the lowest Guffman.

If point % is the best predictor of future success, and Guffman/goal differential runs lock step with point %, isn’t it fair to say that the Guffman is an excellent barometer of future success? Does it share the same correlation? Is it a bit better/worse?

When I raised that years ago, Corsi proponents exclaimed that Corsi was a predictive stat, while goal differential was not. Were they lying or stating that out of ignorance?

If you can conjure up the graphs/analysis, I would appreciate it.

 

[Whileee]

Thanks, Whileee. It is VERY interesting that you mention point % is the best predictor of future success.

I am not sure if you are aware of the ground breaking study I did a few years ago, which still holds up today. The study established a new metric called The Guffman, which is essentially goal differential. Standings, which is based on point %, is highly correlated to The Guffman statistic. If you look at league standings at any point in time, you will see the top teams having the highest Guffman, and the worst teams having the lowest Guffman.

If point % is the best predictor of future success, and Guffman/goal differential runs lock step with point %, isn’t it fair to say that the Guffman is an excellent barometer of future success? Does it share the same correlation? Is it a bit better/worse?

When I raised that years ago, Corsi proponents exclaimed that Corsi was a predictive stat, while goal differential was not. Were they lying or stating that out of ignorance?

If you can conjure up the graphs/analysis, I would appreciate it.

Goal differential can be expressed as a percentage: goal for %. It's easier to work in percentages than actual differences because those depend on the actual number of games and goals.

So, here is a graph showing goal differential (as goals for %, GF%) up to December 8 vs. points % for the rest of the season (i.e. approx. first 28 games vs. remaining 54 games).

There is a correlation, which is a bit lower than the correlation between the early-season CF% vs. the Points% for the rest of the season. So, the Corsi and expected goals metrics are somewhat more correlated with future points % (within a season) than the Guffman %.

But the additional issue is the stability and utility of the estimates for assessing performance at the level of the individual player and/or D-pairs and line combinations. Because goals are much less common than shot attempts (Corsi, etc.) they are less statistically reliable at the individual or pair or line level, so goal differentials aren't as useful for analyzing the performances of individuals. As an example, so far this season the even-strength goal differentials for some Jets players...

Dillon -4
Schmidt -1
DeMelo +7

Gustafsson +2
Kupari -3

You could argue that Dillon has been worse than Schmidt or DeMelo, but has he been 4-times worse than Schmidt (as indicated by his EV goal differential)? Probably not.

Similarly, Gus has been good, but not that much better than Kupari, perhaps (5 goal differential).

So the main reason for using shot metrics isn't so much that goal differentials less important, but that they are less statistically reliable (stable) for analyzing players and combinations because of the smaller numerators.

 

[JetsFan815]

The NHL tracking data is interesting. JFresh has been looking at speed bursts.

Jets have a number of forwards that rank highly, but the Jets' have 5 D in the bottom 100, and only 1 D in the top 100. Interestingly, the only D with a relatively high rate of fast bursts is Samberg, perhaps because he's very fast in his puck retrievals, giving him more time to make a play. Jets D need to play faster, I think, especially getting back in puck retrievals.

View attachment 761448

View attachment 761449

Would be interesting to plot this against xGF% (or maybe RelTM xGF% would be better) to see how much this actually matters and if there is any correlation between this and final results. Also wonder how much of this impacted by systems that the teams run. For example only 2 Carolina d-men are in the top-100 and one of them is #6 and a few surprising names in the Calgary guys are up there (Zadorov, Tanev, Andersson) who I wouldn't expect to be in the top half.

 

[gojetsgo]

 

[BobTheSolarsystem]

#firebradlauer

 

[Whileee]

I still see quite a few comments about the Jets having mediocre defense.

So far, the Jets are trending towards their best 5v5 defensive shot metrics in the 2.0 era. They are currently 2nd in the NHL in CA/60 and 6th in xGA/60.

I think the strong shot metrics are being driven by effective implementation of a very high-pressure system, using speed and aggressiveness to pressure the puck all over the ice. Overall, it's creating more puck possession for the Jets. The issue is that this system increases the risk of odd-man rushes when the reads are wrong or delayed, or the execution isn't good enough. That along with some soft goaltending here and there is leading to some goals against.

If the Jets become more consistent in execution and get a few more saves, they are going to be a very tough opponent. The forward depth and talent is really good, and the D is actually starting to mesh well. As a result...

 

[surixon]

I still see quite a few comments about the Jets having mediocre defense.

So far, the Jets are trending towards their best 5v5 defensive shot metrics in the 2.0 era. They are currently 2nd in the NHL in CA/60 and 6th in xGA/60.

I think the strong shot metrics are being driven by effective implementation of a very high-pressure system, using speed and aggressiveness to pressure the puck all over the ice. Overall, it's creating more puck possession for the Jets. The issue is that this system increases the risk of odd-man rushes when the reads are wrong or delayed, or the execution isn't good enough. That along with some soft goaltending here and there is leading to some goals against.

If the Jets become more consistent in execution and get a few more saves, they are going to be a very tough opponent. The forward depth and talent is really good, and the D is actually starting to mesh well. As a result...View attachment 765044

The only thing they need to do is to start opening the slot more in the offensive zone. They do that and they will be scary offensively as well.

 

[MrBoJangelz71]

I still see quite a few comments about the Jets having mediocre defense.

So far, the Jets are trending towards their best 5v5 defensive shot metrics in the 2.0 era. They are currently 2nd in the NHL in CA/60 and 6th in xGA/60.

I think the strong shot metrics are being driven by effective implementation of a very high-pressure system, using speed and aggressiveness to pressure the puck all over the ice. Overall, it's creating more puck possession for the Jets. The issue is that this system increases the risk of odd-man rushes when the reads are wrong or delayed, or the execution isn't good enough. That along with some soft goaltending here and there is leading to some goals against.

If the Jets become more consistent in execution and get a few more saves, they are going to be a very tough opponent. The forward depth and talent is really good, and the D is actually starting to mesh well. As a result...View attachment 765044

Good post!

I think last year we saw a similar start with similar results.

Team was playing fast and aggressive, high pressure forechecking and aggressive defensive puck pursuit. Problem is last year's team was not well equipped to maintain the style through an 82 game season.

To maintain this style you must be running 4 lines and your team needs to be conditioned for it.

Last years team started off well because we were able to play our bottom 6 a bit more. Then injuries hit, which weakened our bottom 6 which prevented us from playing that high tempo style.

Add in that I think our team lacked conditioning needed to support the pace resulting in a massive drop off in the second half of the season.

This year we have 4 solid lines and probably enough depth to coverup a couple of injuries without dropping off in play. The question is can our conditioning maintain this pace throughout an 82 game season.

 

[BobTheSolarsystem]

I still see quite a few comments about the Jets having mediocre defense.

So far, the Jets are trending towards their best 5v5 defensive shot metrics in the 2.0 era. They are currently 2nd in the NHL in CA/60 and 6th in xGA/60.

I think the strong shot metrics are being driven by effective implementation of a very high-pressure system, using speed and aggressiveness to pressure the puck all over the ice. Overall, it's creating more puck possession for the Jets. The issue is that this system increases the risk of odd-man rushes when the reads are wrong or delayed, or the execution isn't good enough. That along with some soft goaltending here and there is leading to some goals against.

If the Jets become more consistent in execution and get a few more saves, they are going to be a very tough opponent. The forward depth and talent is really good, and the D is actually starting to mesh well. As a result...View attachment 765044

Have to credit bones on this one... they have buy-in from everyone this year and at least partial buy-in from KFC and scheif - let's hope it lasts for the whole season

 

[BoneDocUK]

Good post!

I think last year we saw a similar start with similar results.

Team was playing fast and aggressive, high pressure forechecking and aggressive defensive puck pursuit. Problem is last year's team was not well equipped to maintain the style through an 82 game season.

To maintain this style you must be running 4 lines and your team needs to be conditioned for it.

Last years team started off well because we were able to play our bottom 6 a bit more. Then injuries hit, which weakened our bottom 6 which prevented us from playing that high tempo style.

Add in that I think our team lacked conditioning needed to support the pace resulting in a massive drop off in the second half of the season.

This year we have 4 solid lines and probably enough depth to coverup a couple of injuries without dropping off in play. The question is can our conditioning maintain this pace throughout an 82 game season.

Agree.

Though this year's Jets are bolstered by the well-trained duo (soon trio) of well trained ex-Kings, who have played a similar style for years, and a more experienced and stronger Perfetti and Barron, etc., so maybe they'll be more durable.

 

[ps241]

Good post!

I think last year we saw a similar start with similar results.

Team was playing fast and aggressive, high pressure forechecking and aggressive defensive puck pursuit. Problem is last year's team was not well equipped to maintain the style through an 82 game season.

To maintain this style you must be running 4 lines and your team needs to be conditioned for it.

Last years team started off well because we were able to play our bottom 6 a bit more. Then injuries hit, which weakened our bottom 6 which prevented us from playing that high tempo style.

Add in that I think our team lacked conditioning needed to support the pace resulting in a massive drop off in the second half of the season.

This year we have 4 solid lines and probably enough depth to coverup a couple of injuries without dropping off in play. The question is can our conditioning maintain this pace throughout an 82 game season.

Have to credit bones on this one... they have buy-in from everyone this year and at least partial buy-in from KFC and scheif - let's hope it lasts for the whole season

Agree.

Though this year's Jets are bolstered by the well-trained duo (soon trio) of well trained ex-Kings, who have played a similar style for years, and a more experienced and stronger Perfetti and Barron, etc., so maybe they'll be more durable.

Since the trade deadline last season we have done a 40% makeover/change to our forward group. Namestnikov, Nino, Valarid, Iafallo, and Kupari are all really good responsible defensive systems players (in all three zones). Chevy has very much upgraded the forward assets to fit the system.

 

[gojetsgo]

 

[Buffdog]

Crazy to think we're sitting where we are and helle is still rocking a .892 save%

 

[tbcwpg]

Lots of stats from Jfresh today. As it relates to the Jets, some above average finishing has counterbalanced some poor goaltending and then some. Strong defensively 5v5 but also not generating a ton of offence 5v5. Special teams not good either.

What I take from it is that, if Hellebuyck starts to rebound to form, even if the Jets finishing drops a bit, that should balance things out. If they can figure out the PP and PK they should be comfortably in the top 10-12 of the league.

The downside is Dallas and Colorado are also very good, so someone would be disappointed in the 1st round if the top 3 maintain.

Also standing out was Vancouver is at like +29 goals above expected (scored and saved) whereas Edmonton is -22 goals above expected.

 

[DRW204]

Jets being one of the top finishing teams sure is great to see after the past couple years.

 

[DRW204]

Currently 5th in goals for per 60

Through 18 gp, their running gf/60 never dipped below 3.

Its at the highest its been at this point of the season vs the previous two years.

IMO lot was said about the teams defense in the offseason but the bigger weak point In my view was goal scoring.
This year they've certainly started out real strong on both ends but the bigger improvement in my eyes has been the goals for

 

[Whileee]

Another look at the Jets using the Evolving Hockey RAPM model...

Based on shot metrics (i.e. shot volume and quality) the Jets have been very good defensively (at 5v5 and on the PK). Their 5v5 offense has been less impressive, and their PP offense has been bad. Vilardi might juice the offense. The PK's problem was sub-par goaltending early in the season, which has been coming around.

But it makes you wonder whether the Jets' main need is on D. They certainly don't seem to have problems defending with this D core, and the D has been generating plenty of goals / points. It will be interesting to see if the Jets' 5v5 and PP offense gets to a higher level with Vilardi back in the top-6 and on the PP.

 

[JetsFan815]

Another look at the Jets using the Evolving Hockey RAPM model...

Based on shot metrics (i.e. shot volume and quality) the Jets have been very good defensively (at 5v5 and on the PK). Their 5v5 offense has been less impressive, and their PP offense has been bad. Vilardi might juice the offense. The PK's problem was sub-par goaltending early in the season, which has been coming around.

But it makes you wonder whether the Jets' main need is on D. They certainly don't seem to have problems defending with this D core, and the D has been generating plenty of goals / points. It will be interesting to see if the Jets' 5v5 and PP offense gets to a higher level with Vilardi back in the top-6 and on the PP.

View attachment 773495

A big part of the lack of EV offense is due to the 2nd line (and the 4th line but they don't play as many minutes so impact this less). They have a pretty low 2.18 xG/60 right now which is well below NHL average. Ehlers coming back to form would go a long way. Also Vlad is a good stop gap in case of injuries but maybe not a long term solution there.

 

[Whileee]

A big part of the lack of EV offense is due to the 2nd line (and the 4th line but they don't play as many minutes so impact this less). They have a pretty low 2.18 xG/60 right now which is well below NHL average. Ehlers coming back to form would go a long way. Also Vlad is a good stop gap in case of injuries but maybe not a long term solution there.

The 2nd line has been scoring at a high rate, perhaps because it generates a lot of its offense off the rush, rather than sustained offensive zone time and shot pressure. I think that's okay.

I agree that Ehlers getting back on form will increase the shot metrics for line 2, as well.

Interesting decisions on line combos when Vilardi is back. I think the top line probably needs a boost, so he should end up there. The question is whether to keep Namestnikov on line 3, or shake that up by moving Iafallo there. Right now, I don't think I'd mess with lines 2 or 3, but do you put Iafallo all the way down to line 4? Good dilemma to have.