garret9
AKA#VitoCorrelationi
Penguins are 4th in points despite being near the end in goal differential due to 3 blow outs.
Small samples and rarity of goals makes weird stuff happen.
Small samples and rarity of goals makes weird stuff happen.
2017-18 stats and underlying metrics thread [Mod: updated season]
- Thread starter garret9
- Start date
- Status
Interesting. What previous seasons are included?
Includes last season.
Corsica xGoals are the 1.0 version.
So, r-squared is about 0.15 for Corsi, and less for Fenwick and other metrics?
Corsi at 20 games explains 15% of the variance in goals for the final 62 games?
garret9
AKA#VitoCorrelationi
Aye. Raw. No adjustments (arena, venue, score).
Theoretically speaking, about 1/3 of the variance* is luck.
So, we have a **rough** estimate of raw Corsi predicting under 23% of the non-luck variation in goal share distribution.
Not surprising.
- no adjustments for confounding variables (score, venue, arena, etc.)
- no shot quality
- no finishing talent
- no goaltending talent
- teams are not the same roster for games 1-20 as they are games 21-82
*The graph is record with OT and shootouts ignored. This artificially reduces the role of luck but it gives us a rough guideline.
Last edited:
JetsFan815
Registered User
- Jan 16, 2012
- 19,258
- 24,483
Every Jets game that I watch I think to myself if this is going to be the beginning of regression. I want to believe that we are anyway near as good as the record shows but I can't help but worry that just like the subprime mortgage crisis things are about to go belly-up any moment. Helle bought the coaching staff some time to make some changes, they have to right the ship soon or else I am not sure how long this can last.
winnipegger
Registered User
- Dec 17, 2013
- 8,287
- 6,803
I feel like when the game is inevitably played by robots in the coming decades they will simply program the robots with algorithms highly correlated with long term success. Game will be entirely decided by luck and a human being's role in the game of hockey will simply be to maintain the robots and lubricate their mechanical parts. It must be a nightmare trying to teach people with no higher education (players, coaches) the benefits of all of these things. But with robots it will be quite simple.
I actually really enjoy trying to understand what the hell you are all talking about. There are some graduate degrees in statistics floating around in HF jets I had no idea.
I actually really enjoy trying to understand what the hell you are all talking about. There are some graduate degrees in statistics floating around in HF jets I had no idea.
- Jun 10, 2014
- 57,479
- 29,341
It is easier to teach robots to lube and maintain robots than it is to teach them to play hockey.The role of humans will be to consume the advertised products.
mcpw
WPG
- Jan 13, 2015
- 10,024
- 2,072
2014-18, >140 minutes 5v5TOI:
Most productive lines with Scheifele (alternative title: only sub-50%-Corsi lines with Scheifele) --
Ehlers-Scheifele-Laine, 5.1 GF/60 (49.6 CF%, 60.5 GF%)
Connor-Scheifele-Wheeler, 3.7 GF/60 (44.6 CF%, 60.0 GF%)
Stafford-Scheifele-Wheeler, 3.5 GF/60 (48.7 CF%, 69.2 GF%)
The Chef himself seems to be trolling.
Most productive lines with Scheifele (alternative title: only sub-50%-Corsi lines with Scheifele) --
Ehlers-Scheifele-Laine, 5.1 GF/60 (49.6 CF%, 60.5 GF%)
Connor-Scheifele-Wheeler, 3.7 GF/60 (44.6 CF%, 60.0 GF%)
Stafford-Scheifele-Wheeler, 3.5 GF/60 (48.7 CF%, 69.2 GF%)
The Chef himself seems to be trolling.
5.1 GF/60...
It's so weird looking at these numbers from last year:
There was no reason to separate Laine and Scheifele whatsoever. They were pretty even Corsi-wise but were just scoring like crazy. It's one thing to think that they were getting lucky but Maurice should know better by watching Laine by now.
Wheeler, Scheif, and Ehlers were doing well in the Corsi department but they could not get it done offensively. Scheifele's as well as Laine's production falls off a cliff when separated. Just look at TOI for #1 and #2 lines on that list, then look at the goal differentials.
Maurice, could we please have nice things again? Y u do dis?
Just dropping in to point out that Hendricks has a 6.2 rel Fenwick% while getting slammed in the dzone at 63.6%. Seems like people are blasting how slow he looks on the ice while ignoring the results...
Ducky10
Searching for Mark Scheifele
- Nov 14, 2014
- 19,809
- 31,386
[
Laughing usually isn't my reaction, it's more like this.
Our PK is getting laughed at
Laughing usually isn't my reaction, it's more like this.
WPGChief
Registered User
I think this should be taken more as a "every other Jets player is getting out-shot all the time" rather than "Hendricks' 9GP and 62TOI (5v5) of sample size outperforming his past ~500 GP of sample size means the numbers were wrong!!!"
Remember when Tanev used to be regarded in the same manner around 5-7 games ago?
Gil Fisher
Registered User
where are we getting our fancy shot and xG stats from these days? I can't find squat anymore (besides Corsica, which seems sub-optimal).
Gil Fisher
Registered User
You sure? I think we're around 12th/13th in GF% (GA vs GF), but near 22nd or 23rd in xGF%.
Gil Fisher
Registered User
Yeah. Been doubting it since I posted...just came back to the data and I have no idea where I got it from.
So in conclusion. We blow chunks.
Sean Tierney's got some interesting graphs - he posted a whole whack of 'em recently, if you're interested. IE:
Dayofthedogs
Bettman's hammer
Regarded Tanev around here like what?
We would say he's actually played well when he has instead of referencing last years stats and calling him an ECHLer?
The post you replied to said nothing of how he expects Hendricks to play it said he hasn't been doing too bad. Has he or has he not performed well in relFenwick?
This kind of stuff bugs me a bit. I get the distinct impression some posters would rather watch certain Winnipeg Jets fail to reaffirm their position. Tanev has been one of the 12 best forwards this year and probably deserves to be in the group of 14....
I'm under no illusion about what Tanev or Hendricks are but I'm going to cheer for them as long as their wearing jets uniforms and I'm happy point out when they've had a good game or done something well.
Last edited:
Aavco Cup
"I can make you cry in this room"
- Sep 5, 2013
- 37,630
- 10,440
Would love to know what they are using. Whether it's contracted or internal etc.
Board Bard
Dane-O-Mite
- Jun 7, 2014
- 7,890
- 5,059
Player --- Expected goals (Maurice version)
Lowry --------20
Tanev --------15
Hendricks -- 10 plus grit
Dayofthedogs
Bettman's hammer
I don't watch a lot of other coaches pressers but is there any others who actually reference fancy stats as much as PoMo does? He seems to be way more open to discussing this kind of stuff in the media. It kinda flies the the face of those whom believe PoMo has no idea abut whats actually going on with regards to the poor shot metrics and PK.
I get the sense he puts out a lot of fires in the media when he can and is well aware of the teams short comings.
Aavco Cup
"I can make you cry in this room"
- Sep 5, 2013
- 37,630
- 10,440
Moved this here since it would get buried in the GDT
Since there are several versions of xG out there, it's not surprising that teams are pushing envelopes in this area. Pick any of the publicly available methods and the team can probly end up with a more refined end product. The fact that they can employ video breakdowns of every shot/scoring chance and refine shot quality more subjectively than just shot location from NHL.com would make one think at least in theory that this sort of system would better reflect reality
Since there are several versions of xG out there, it's not surprising that teams are pushing envelopes in this area. Pick any of the publicly available methods and the team can probly end up with a more refined end product. The fact that they can employ video breakdowns of every shot/scoring chance and refine shot quality more subjectively than just shot location from NHL.com would make one think at least in theory that this sort of system would better reflect reality
garret9
AKA#VitoCorrelationi
It's no secret that I have talked to *members* of the Jets organizations. I've had solid conversations with some of these, and have heard good things about other members of the organization (supposedly Larry Simmons is really smart and actively looking to improve things in many areas).
I know that Maurice is interested in things from a team level perspective.
Now, the real question is about HOW one uses statistics and, most importantly, how do they attack situations where their own notions are counter to what statistics may indicate.
Another question is how do they deal with things that may be more explanative than predictive.
I know that Maurice is interested in things from a team level perspective.
Now, the real question is about HOW one uses statistics and, most importantly, how do they attack situations where their own notions are counter to what statistics may indicate.
Another question is how do they deal with things that may be more explanative than predictive.
Board Bard
Dane-O-Mite
- Jun 7, 2014
- 7,890
- 5,059
My guess for all of those questions is "poorly," based partly on who they sign and how they arrange the roster, and partly on me thinking they're pretty close to neanderthal.
- Status
