Toggle sidebar

The Advanced Stats Thread Episode VIII:

  • Xenforo Cloud has upgraded us to version 2.3.6. Please report any issues you experience.
Status
Not open for further replies.
Blue Blooded said:
Disagree with the last sentence, his projections heavily suggest that he's been carried and his pre-NHL record does not indicate significant NHL upside. I wouldn't take him over DeAngelo right now or for the future. Don't get me wrong, a player that can hang with a good partner is valuable. But unlike Toronto we don't have one, let alone two, LDs who can carry a player like him.

View attachment 143271 View attachment 143273
Click to expand...
How is this guy being carried? He's the one carrying everyone else...

TOR 1718 Carrick With-or-Without You
 


That's a shame. Should we claim Gagner and call it a day? 2 more years at 3.15. Could be a useful flip at the deadline given his history of scoring on the PP.
 
Blue Blooded said:
At least I'm not at 900.0% yet!
Click to expand...
My reaction to ADA sitting instead of McQuaid:
4a4.png
 
  • Like
Reactions: Ail and Blue Blooded
silverfish said:
How is this guy being carried? He's the one carrying everyone else...

TOR 1718 Carrick With-or-Without You
Click to expand...
Delved into the microstats to see why he rates out low:
  • Doesn't score points
  • High number of shot attempts with very low ixFSh% (this is an underrated red flag according to my model)
  • Plays very soft xGF/xGA QoC
  • Low TOI% (basically a non-factor though)
The fact that his possession metrics are great, he gets offensive deployment, that he gets hit and that he draws penalties couldn't make up for it.

Basically he is no Adam Clendening ;)
 
Blue Blooded said:
Delved into the microstats to see why he rates out low:
  • Doesn't score points
  • High number of shot attempts with very low ixFSh% (this is an underrated red flag according to my model)
  • Plays very soft xGF/xGA QoC
  • Low TOI% (basically a non-factor though)
The fact that his possession metrics are great, he gets offensive deployment, that he gets hit and that he draws penalties couldn't make up for it.

Basically he is no Adam Clendening ;)
Click to expand...
I'm surprised that ixFSh% is important at all for defenders. Such a huge component of any xG model is going to be shot distance and shot type. My guess is most defenders are shooting from 55+ feet and mostly slaps or wrist shots, so they're never going to grade out too well their on the aggregate.
Blue Blooded said:
Another indicator that his possession metrics may be circumstantial:
View attachment 143475
Click to expand...

Interesting and disheartening.
 
I think that’s a byproduct of how Babcock runs his transitions. Zaitsev has poor numbers too, while Gardiner and Rielly are absolute transition studs. Seems odd that all your RD have bad transition stats but all of your LD are kings.
 
Mac n Gs said:
I think that’s a byproduct of how Babcock runs his transitions. Zaitsev has poor numbers too, while Gardiner and Rielly are absolute transition studs. Seems odd that all your RD have bad transition stats but all of your LD are kings.
Click to expand...
This is a good point. I really love the granularity we're getting with entry and exit data, but they are extremely system based. Just look at Smith on Detroit vs Smith on NYR. I'm assuming they're different. Could be wrong.

I'd love to see entries/exits relative to team.
 
silverfish said:
This is a good point. I really love the granularity we're getting with entry and exit data, but they are extremely system based. Just look at Smith on Detroit vs Smith on NYR. I'm assuming they're different. Could be wrong.

I'd love to see entries/exits relative to team.
Click to expand...
I mean it makes sense from a tactical standpoint, and it matches what I see them try to do.

They try to feed it to Marner, Nylander, and Kapanen (not to the same degree) off zone entries. Marner and Nylander are both incredible playmakers off the RW and are both in position to either shoot or generate royal road passes.
 
silverfish said:
This is a good point. I really love the granularity we're getting with entry and exit data, but they are extremely system based. Just look at Smith on Detroit vs Smith on NYR. I'm assuming they're different. Could be wrong.

I'd love to see entries/exits relative to team.
Click to expand...
Quite similar actually:
upload_2018-10-2_16-12-59.png
 
silverfish said:
I'm surprised that ixFSh% is important at all for defenders. Such a huge component of any xG model is going to be shot distance and shot type. My guess is most defenders are shooting from 55+ feet and mostly slaps or wrist shots, so they're never going to grade out too well their on the aggregate.
Click to expand...

I should clarify, it is only important (implicitly important I should add) if you're trying to project certain things. You have a negative coefficient for iCF/60 and a positive one for ixGF/60 if you're trying to project P/60, GF/60, GD/60, xGF/60, xGD/60, xGF%Rel, while both are positive for G/60, CF/60, CD/60 and CF%Rel. The balance between the coefficients has the consequence that many shot attempts with high ixFsh% is the best, though many attempts with a decent ixFsh% is likely better than very few attempts with a high ixFsh%. Firing away low-danger attempts or not getting your shots through (blocked shot = 0% "danger") counts as a negative.

I've got a few off the cuff rationales as to why:
  • A shot attempt is something that has a likelihood of ending possession - especially if blocked. I can see how taking a low-danger shot instead of creating a better opportunity would be detrimental to xGF%.
  • Being able to get into positions to put up high ixGF numbers indicates talent, especially if the player is able to do so without tanking his CF%/xGF%. Any defenceman can stand at the blue line and throw pucks at the net as soon as he gets them.
 
  • Like
Reactions: silverfish
Blue Blooded said:
I should clarify, it is only important (implicitly important I should add) if you're trying to project certain things. You have a negative coefficient for iCF/60 and a positive one for ixGF/60 if you're trying to project P/60, GF/60, GD/60, xGF/60, xGD/60, xGF%Rel, while both are positive for G/60, CF/60, CD/60 and CF%Rel. The balance between the coefficients has the consequence that many shot attempts with high ixFsh% is the best, though many attempts with a decent ixFsh% is likely better than very few attempts with a high ixFsh%. Firing away low-danger attempts or not getting your shots through (blocked shot = 0% "danger") counts as a negative.

I've got a few off the cuff rationales as to why:
  • A shot attempt is something that has a likelihood of ending possession - especially if blocked. I can see how taking a low-danger shot instead of creating a better opportunity would be detrimental to xGF%.
  • Being able to get into positions to put up high ixGF numbers indicates talent, especially if the player is able to do so without tanking his CF%/xGF%. Any defenceman can stand at the blue line and throw pucks at the net as soon as he gets them.
Click to expand...
I would've been able to calculate the threshold ixFsh% where a shot attempt turns from a detriment to a positive, but my coefficient CSV that R printed for me is broken and any coefficient that starts with anything other than "0," is just a long string of numbers without a decimal point.

Going by which decimal placement produces reasonable results, a shot from a defenceman must have an ixFSh% of at least 1.067% to contribute positively to a projected xGF%Rel. I.E. as long as it's not blocked it's basically good, but the higher the quality the better.

An example of how it comes out; Brent Burns' 22.17 iCF/60 and 0.38 ixGF/60 gives his projected xGF%Rel a +2.55 while Anthony DeAngelo's 11.41 iCF/60 and 0.3 ixGF/60 gives him a +3.17. However, the model has no idea that DeAngelo has an airsoft while Burns has a 7.62 cal SIG516.
 
Status
Not open for further replies.

Latest posts

Ad

Upcoming events

Ad

Top Bottom