2018-19 stats and underlying metrics thread
- Thread starter garret9
- Start date
JetsWillFly4Ever
PLAY EHLERS 20 MIN A NIGHT
MoneyPuck twitter account basically said the Jets have been getting significantly lucky all year.
NHL Playoff Odds -MoneyPuck Analytics Based Playoff & Cup Odds
They also say the St. Louis Blues have a much better chance at winning the Stanley Cup than the Jets and Predators do. Pass the salt?
Blues and Avs with a higher probability of the Cup than the Jets or Caps. Okay....
Jets at #10 in PDO 5v5. That's not too "lucky". Special teams have been an advantage for the Jets.
If they mean that the Jets are lucky because they are not at the centre of their model's probability distribution, then they don't understand statistical probability. All models are wrong.
JetsWillFly4Ever
PLAY EHLERS 20 MIN A NIGHT
I haven't looked too much into it but they basically just posted a graph of the Jets outperforming their xG differential by a fairly significant amount.
It's not clear how much random error exists in their xG model predictions. The point is that you should always expect a certain proportion of teams to deviate from the model prediction. That doesn't necessarily mean they are "lucky" or "unlucky", but might mean that the model hasn't precisely estimated their expected results due to imprecision and/or inaccuracy in the models.
winnipegger
Registered User
- Dec 17, 2013
- 8,223
- 6,600
Are these predictions ever right? In my experience the arrow of time only moves forward and nobody knows what will happen in the playoffs. VGK's odds of going to the conference finals last year were probably -6343%. That number doesn't even make sense it's so crazy.
- Mar 10, 2010
- 34,899
- 31,348
Interesting post Whileee.
I always like reading and gaining knowledge from people who have much more expertise in areas than I do.
pucka lucka
Registered User
What? These aren't prediction ffs. they are probabilities. Those are not the same thing.
I referenced their xG predictions, from which they derived performance probabilities. Two different things. Lack of precision or accuracy in their xG predictions will affect their probability calculations.
JetsWillFly4Ever
PLAY EHLERS 20 MIN A NIGHT
Yep, and that's only at 5v5, I read his GSAA while we are shorthanded is positive.
2017-18 NHL Goalie Statistics | Hockey-Reference.com
The GPS stat; explained as Goalie point share, an estimate of the number of points contributed by a player due to his play in goal
Last year Hellebuyck led the league in that category. How reliable/usable is that stat, and how does it work exactly? Does it matter if you win (or lose) 12-0 or 1-0?
The GPS stat; explained as Goalie point share, an estimate of the number of points contributed by a player due to his play in goal
Last year Hellebuyck led the league in that category. How reliable/usable is that stat, and how does it work exactly? Does it matter if you win (or lose) 12-0 or 1-0?
Are these stats surprising to anyone?
From Evolving-Hockey (expected goals plus/minus per 60 minutes)... The methodology they use is based on a model that adjusts for teammates and opponents.
From Evolving-Hockey (expected goals plus/minus per 60 minutes)... The methodology they use is based on a model that adjusts for teammates and opponents.
Here's a "stats" based suggestion for getting Laine going. How about putting him with the Jets' two best players in terms of adjusted expected goals for (Copp and Ehlers), instead of two of the poorer players in that regard (Little and Connor)?
Ducky10
Searching for Mark Scheifele
- Nov 14, 2014
- 19,809
- 31,386
Based on what I think I have seen and lacking hard data, I'd say only marginally.
Maukkis
EZ4ENCE
- Mar 16, 2016
- 10,576
- 7,279
Little and Connor being among the worst might just be the consequence of them being shackled to Laine. Partially because Laine has sucked this season, but also because Laine produces little xGF with his shots (he just is capable of overperforming his expected totals over a large sample size).
Connor has brought down both Little and Scheifele lines, whereas Ehlers elevates them. Copp has been a net positive despite being shackled to Lemieux.
I'm going to keep pushing for trying Copp with Laine until Maurice relents and gives it a try.
Gil Fisher
Registered User
Here's some more context on 5v5 play so far this year. P/60 ranking among league forwards (>400 minutes). Added a few others in for colour. Copp has really moved up since slotting in between Lemieux and Appleton. Roughly, 1-90 are 1st liners; 91-180 are 2nd liners; 180-270 are 3rd liners.
13. Matt Duchene (2.93)
66. Mark Stone (2.17)
69. Blake Wheeler (2.16)
78. Mark Scheifele (2.10)
98. Brian Little (2.01)
100. Sean Couturier (2.00)
115. Andrew Copp (1.90)
116. Nikolaj Ehlers (1.90)
158. Nikolaj Tanev (1.66)
160. Braden Schenn (1.65)
167. Patrik Laine (1.62)
171. Kyle Connor (1.62)
197. Mathieu Perreault (1.48)
217. Jack Roslovic (1.39)
278. Adam Lowry (1.10)
Lost Appleton (2.42 in 198mins) and Lemieux (2.14 in 253mins) due to filters.
13. Matt Duchene (2.93)
66. Mark Stone (2.17)
69. Blake Wheeler (2.16)
78. Mark Scheifele (2.10)
98. Brian Little (2.01)
100. Sean Couturier (2.00)
115. Andrew Copp (1.90)
116. Nikolaj Ehlers (1.90)
158. Nikolaj Tanev (1.66)
160. Braden Schenn (1.65)
167. Patrik Laine (1.62)
171. Kyle Connor (1.62)
197. Mathieu Perreault (1.48)
217. Jack Roslovic (1.39)
278. Adam Lowry (1.10)
Lost Appleton (2.42 in 198mins) and Lemieux (2.14 in 253mins) due to filters.
garret9
AKA#VitoCorrelationi
garret9
AKA#VitoCorrelationi
winnipegger
Registered User
- Dec 17, 2013
- 8,223
- 6,600
What the hell is the matter with Carolina. Every year they are in this spot. They even have a bonafide elite center now with Aho and some decent scoring talent on the wings. It really boggles the mind that they can't make the playoffs. Consistently good shot metrics.
