Leafs' Prospect/Marlies Discussion Thread!

[leafman101]

Looking at relative numbers there was little change though.

Faulk rel xGF/60
2019-20: 0.14
2018-19: -0.31
2017-18: 0.1
2016-17: 0.35
2015-16: 0.14

Rel xGA/60
2019-20: 0
2018-19: 0.24
2017-18: 0.07
2016-17: -0.06
2015-16: 0.36

Rel should account for differing systems, and are IMO much more valuable.

 

[leafman101]

But agree that looking at any one discrete stat alone is not enough information to make concrete evaluations. Or at least have a great amount of confidence in them.

 

[BeLeafer]

we've gone through examples of this before.

I can try again.

Here's one example of a fairly big minute defenseman who went from one good team to another good team, both that played very different styles, and played approximately the same minutes and role - Justin Faulk.

Last 3 years:

CAR: 53.5xgf% (#5) ----> 2.76xgf (#1), 2.40xga (#16)
STL: 52.4xgf% (#7) -----> 2.38xgf (#18), 2.16xga (#3)

So, 2 very good teams, with opposite styles of play.

Faulk went from similar role to similar role:

CAR '18: 17:46evtoi, -3.4dzdiff,
CAR '19: 18:48evtoi, -1.7dzdiff
STL '20: 18:23evtoi, -0.9dzdiff

And put up similar xgf%:

CAR '18: 51.6% (0.0rel)
CAR '19: 51.9% (-0.4rel)
STL '20: 51.2% (0.0rel)

But saw a significant change in the individual elements:

CAR '18: 2.67xgf, 2.51xga
CAR '19: 2.78xgf, 2.58xga
STL '20: 2.35xgf, 2.24xga


Are we actually supposed to conclude that Faulk became significantly worse offensively and better defensively in equal measure upon switching teams, or should we make the much more obvious conclusion?

The individual elements reveal that Faulk moved to a team that is tighter both defensively and offensively. xGF% only stayed about equal because both dimensions were more or less equally affected by the team. (Of course, you would never know this if all you use is xgf%.)

Faulk probably seems like a great example to use because it stayed so consistent between the teams. But the numbers (5v5 sva) on NST are quite different from yours and Faulk's xgf% fluctuated a lot.

13-14: 48.04% (-0.62 rel)
14-15: 53.25% (3.22 rel)
15-16: 49.38% (-2.47 rel)
16-17: 52.62% (4.61 rel)
17-18: 52.72% (0.27 rel)
18-19: 52.04% (-5.28 rel)
19-20: 51.51% (1.59 rel)

What the xga and xgf reveal for each of these seasons is that Faulk was very similar to all other defensemen on his team on xga and below avg on xgf. The latter fits well with him getting a good chunk of his points on the pp. Good pp guy, can hold his own defensively and not much of a play driver 5 on 5.

 

[BeLeafer]

giffin like a mouth breather too.

 

[zeke]

Looking at relative numbers there was little change though.

Faulk rel xGF/60
2019-20: 0.14
2018-19: -0.31
2017-18: 0.1
2016-17: 0.35
2015-16: 0.14

Rel xGA/60
2019-20: 0
2018-19: 0.24
2017-18: 0.07
2016-17: -0.06
2015-16: 0.36

Rel should account for differing systems, and are IMO much more valuable.

absolutely.

especially in this case where i picked a player in the ~same usage and role to keep other variables at bay.

there's entirely new levels of complexity when you start comparing players in significantly different roles in addition to those team differences.

 

[BeLeafer]

But agree that looking at any one discrete stat alone is not enough information to make concrete evaluations. Or at least have a great amount of confidence in them.

I agree. You would never use xga to evaluate a player. It can help guide you assessment of various dimensions of their play.

 

[zeke]

The individual elements reveal that Faulk moved to a team that is tighter both defensively and offensively. xGF% only stayed about equal because both dimensions were more or less equally affected by the team. (Of course, you would never know this if all you use is xgf%.)

Faulk probably seems like a great example to use because it stayed so consistent between the teams. But the numbers (5v5 sva) on NST are quite different from yours and Faulk's xgf% fluctuated a lot.

13-14: 48.04% (-0.62 rel)
14-15: 53.25% (3.22 rel)
15-16: 49.38% (-2.47 rel)
16-17: 52.62% (4.61 rel)
17-18: 52.72% (0.27 rel)
18-19: 52.04% (-5.28 rel)
19-20: 51.51% (1.59 rel)

What the xga and xgf reveal for each of these seasons is that Faulk was very similar to all other defensemen on his team on xga and below avg on xgf. The latter fits well with him getting a good chunk of his points on the pp. Good pp guy, can hold his own defensively and not much of a play driver 5 on 5.

so you're agreeing that its nonsense to look at a player who has the toughest usage on one of the worst xga teams in the league and pretend his raw xga number describes his defensive ability. good.

 

[leafman101]

I agree. You would never use xga to evaluate a player. It can help guide you assessment of various dimensions of their play.

For sure. Thats why evaluations have to be so detailed with caveats. IMO it is accurate to state that Rielly has not been a good at suppressing chances, while playing against some of the toughest matchups in the league, on a poor suppression team, with poor quality of teammates in his partners. Whatever that is worth.

And by signing a guy like Brodie, shifting some matchups to Muzzin and improving the Leafs team D system you address 3/4 problems there, which will likely impact the remaining problem.

 

[BeLeafer]

so you're agreeing that its nonsense to look at a player who has the toughest usage on one of the worst xga teams in the league and pretend his raw xga number describes his defensive ability. good.

Yeah, if you ignore everything I've said, you can draw that conclusion.

 

[BeLeafer]

Faulk was consistently below his blueliner teammates on xgf but average on xga (for both Carolina and St. Louis). His xgf% flucuated quite a bit though, particularly relatively.

Rielly, by comparison, is consistently above his blueliner teammates on xgf and well below on xga.

 

[zeke]

Faulk was consistently below his blueliner teammates on xgf but average on xga (for both Carolina and St. Louis). His xgf% flucuated quite a bit though, particularly relatively.

correct. his raw xga number in itself is actually meaningless.

Rielly, by comparison, is consistently above his blueliner teammates on xgf and well below on xga.

1. Rielly is consistently WAY above his blueliner teammates in quality of competition, while a guy like faulk is not.

2. Rielly could choose to our be told to play a different way, which could lower both his higher xgf and higher xga, without actually changing how good he is - i.e. how much he outproduces the opposition while on the ice.

 

[BeLeafer]

correct. his raw xga number in itself is actually meaningless.

I'm not sure what you even mean by raw, but any stat in insolation of other variables is quite useless.

1. Rielly is consistently WAY above his blueliner teammates in quality of competition, while a guy like faulk is not.

Rielly consistently struggles defensively and is probably over his head defending against top comp. His consistently team worst xga/60 clearly shows this. Keefe seems to agree and so has given the load to other players.

Who got the tougher usage this season -- Muzzin or Rielly?

2. Rielly could choose to our be told to play a different way, which could lower both his higher xgf and higher xga, without actually changing how good he is - i.e. how much he outproduces the opposition while on the ice.

If Rielly could choose to play well defensively, he would have done so long ago. Limiting scoring chances in your own end does not detract from production -- it improves it.

 

[zeke]

I'm not sure what you even mean by raw, but any stat in insolation of other variables is quite useless.

Rielly consistently struggles defensively and is probably over his head defending against top comp. His consistently team worst xga/60 clearly shows this.

Rielly's xgf% has always been comparable to the other dmen who face similar super-elite usage, actually.

Keefe seems to agree and so has given the load to other players.

Who got the tougher usage this season -- Muzzin or Rielly?

Rielly, by a lot.

Ceci a distant 2nd.

If Rielly could choose to play well defensively, he would have done so long ago. Limiting scoring chances in your own end does not detract from production -- it improves it.

Rielly plays very well - a positive impact player - against super elite competition. he has played the style and systems his coach has employed for the team. it has mostly been a very bad system.

you are choosing to say he has played "badly defensively" based on an isolated stat that is subject to all the things that have already been described above, and which you have already admitted doesn't allow that conclusion.

Rielly wouldn't have to get "better defensively" to improve his xga - he could simply play a much more conservative style, and take no offensive risks, and his xga would come down. his whole team could play that way and it would come down even further. and if he played normal matchups it would come down further still. none of that would mean he's actually improving defensively.

 

[Preston]

Solid debate fellas

 

[BeLeafer]

Rielly's xgf% has always been comparable to the other dmen who face similar super-elite usage, actually.

How does his xGA/60 stack up to them? Let's see who you are talking about and look at the 'raws' and then how these players fare against their teammates.

[Rielly, by a lot.

I guess that's enough to conclude that his ugly xGA/60 is explained simply by qoc.

you are choosing to say he has played "badly defensively" based on an isolated stat that is subject to all the things that have already been described above, and which you have already admitted doesn't allow that conclusion.

You get really inventive when you disagree.

I have never said he's bad defensively.

 

[zeke]

How does his xGA/60 stack up to them? Let's see who you are talking about and look at the 'raws' and then how these players fare against their teammates.


I guess that's enough to conclude that his ugly xGA/60 is explained simply by qoc.


You get really inventive when you disagree.

I have never said he's bad defensively.

unfortunately corsica is no longer and puckIQ is much harder to use.

but let me look at all the guys with a greater than +10% differential between toi vs elite/grit just this year for now (rielly's been getting this kind of usage and worse for years now):

(plus is good for offense bad for defense)


Schmidt +.21 (-.06)
Lindell +.20 (-.03)
Hronek +.18 (-.11)
Weber +.17 (+.22) <---
Morrissey +.17 (+.03)
Rielly +.16 (+.13) <---
McNabb +.15 (-.06)
Chara +.13 (-.09)
Chiarot +.12 (-.05)
Bouwmeester +.08 (-.38)
Parayko +.04 (-.08)
Provorov -.04 (+.06) <---
Niskanen -.05 (+.13) <---
Pelech -.06 (-.11)
McAvoy -.06 (+.26) <---
Pulock -.10 (-.06)
Greene -.14 (-.14)
Slavin -.14 (+.02) <---
Nemeth -.28 (+.11) <---

there's a few guys with beastly positive differentials here - McAvoy (who is awesome but I think gets very specific usage with chara and the big line), Nemeth (??? detroit's numbers are so wack I dunno what this is but maybe he's awesome?) and Slavin (pure beast). Also the Provo-Nisky pairing had a great year (after both being bad the previous year).

And thrn Rielly and Weber who managed to stay around par but in a much different way than the others. McAvoy and Weber the only ones with better offensve numbers in that kind of usage.

none of the rest of those guys comes close to staying even near even.

 

[BeLeafer]

What numbers are in those calculations? Specifically, what's elite/grit?

 

[zeke]

eh you'll have to read puckiq for that.

I dont love it but its actually pretty comprehensive and not bad .

 

[zeke]

PuckIQ

Hockey Analytics

www.puckiq.com

 

[BeLeafer]

Thanks. What does puckiq call these numbers?