PDA

View Full Version : Super Advanced Awesomeness



zeke
10-15-2014, 01:33 PM
Glossary:

FR% = Fenwick Rel%: The player's on-ice Fenwick% minus the player's off-ice Fenwick%; off-ice Fenwick% is the percent of unblocked on-ice shot attempts taken by the player's team when the player is not on the ice; also known as FF Rel%

FC% = Fenwick Competition%: The TOI-weighted Fenwick% for a player's competition

CR% = Corsi Rel%: The player's on-ice Corsi% minus the player's off-ice Corsi%; off-ice Corsi% is the percent of on-ice shot attempts taken by the player's team when the player is not on the ice; also known as CF Rel%

CC% = Corsi Competition%: The TOI-weighted Corsi% for a player's competition

OZ% = Offensive ZS%: The percent of all non-neutral zone on-ice faceoffs taken in the offensive zone



Even Strength 5v5

N.Kadri: 12:56, +8.31fr%, 50.49fc%, +9.01cr%, 51.49cc%, +9.94oz%
J.Lupul: 14:27, +9.07fr%, 50.57fc%, +8.53cr%, 51.56cc%, +1.67oz%
D.Winnik: 12:20, +8.17fr%, 49.56fc%, +9.02cr%, 50.57cc%, -11.49oz%

Notes: Elite possession numbers against solid competition, but with heavily favorable zone starts, so not quite as impressive as it looks.

M.Santorelli: 11:22, +5.56fr%, 50.98fc%, +2.56cr%, 52.04cc%, +0.51oz%
D.Clarkson: 11:26, +10.55fr%, 50.73fc%, +7.21cr%, 51.71cc%, +9.53oz%
L.Komarov: 11:36, +8.18fr%, 51.37fc%, +5.24cr%, 52.32cc%, -0.65oz%

Notes: Solid possession numbers against solid competition with fairly neutral zone starts. Solid line.

T.Bozak: 13:38, -6.33fr%, 53.79fc%, -3.53cr%, 54.61cc%, -3.14oz%
J.Vanryk: 13:12, -9.20fr%, 53.68fc%, -3.74cr%, 54.57cc%, -3.14oz%
P.Kessel: 14:39, -11.01fr%, 53.68fc%, -8.11cr%, 54.55cc%, -0.85oz%

Notes: Horrific possession numbers, but against by far the toughest competition and with unfavorable zone starts, so not quite as awful as it looks. They still need to be much better, though.

P.Holland: 8:28, -10.00fr%, 45.74fc%, -11.84cr%, 47.25cc%, -7.36oz%
B.Kozun: 7:32, -6.58fr%, 46.68fc%, -10.08cr%, 48.14cc%, +9.41oz%
R.Panik: 7:20, -13.74fr%, 46.41fc%, -20.73cr%, 48.20cc%, -2.72oz%
M.Frattin: 9:20, -5.36fr%, 41.90fc%, +8.17cr%, 42.70cc%, -22.07oz%

Notes: Awful posession numbers against weak competition with neutral zone starts. 4th line has been really bad so far. Might be time to give some other players a look here.



D.Phaneuf: 15:23, +7.14fr%, 52.04fc%, +6.40cr%, 53.00cc%, +4.10oz%
S.Robidas: 15:17, -7.92fr%, 53.15fc%, -4.87cr%, 53.50cc%, -4.34oz%

Notes: Dion has very good possession numbers against very tough competition, with only slightly favorable zone starts. He's been great. Robidas has pretty bad possesion number, but against very tough competition and slightly unfavorable zone starts. Not as bad as it looks, but droppping him down to the third pair for now probably makes sense.

S.Percy: 15:30, -1.94fr%, 51.39fc%, -4.19cr%, 52.56cc%, -20.20oz%
R.Polak: 17:10, +1.00fr%, 51.70fc%, +2.28cr%, 52.82cc%, -19.04oz%

Notes: this pair has been out and out great. Good possession numbers against good competition, despite heavily unfavorable zone starts. No asterisks on this pair's great play so far.

J.Gardiner: 16:38, +6.58fr%, 49.34fc%, +6.57cr%, 50.34cc%, +17.50oz%
M.Rielly: 14:34, +0.00fr%, 48.73fc%, +0.15cr%, 49.71cc%, +25.25oz%
C.Franson: 13:38, -7.11fr%, 48.27fc%, -8.44cr%, 50.08cc%, +5.28oz%

Notes: Jake is actually the only one who comes out looking good here. Solid possession numbers against very tough competition, though with very favorable zone starts. Franson and Rielly both have poor possession numbers against pretty weak competition, even with extremely favorable zone starts. Jake is probably not the one who should be sitting.

zeke
10-15-2014, 01:35 PM
I might adjust those numbers to look only at "Close" situations but that's good for now.

number17
10-15-2014, 01:37 PM
Thanks Zeke ... how do they compare to other comparable players in the league though?

zeke
10-15-2014, 01:41 PM
Team hasn't been good so far. Ranked 24th in CR% and 26th in FR%.

zeke
10-15-2014, 01:47 PM
mind you, that's still a big step up from being last by a country mile in both last year.

number17
10-15-2014, 01:49 PM
Team hasn't been good so far. Ranked 24th in CR% and 26th in FR%.No surprise there ... yet we were to believe Randy would change?

Granted, it's been 4 games only, but I'm willing to bet we'll still have similar numbers, if not worse after another month.

zeke
10-15-2014, 01:54 PM
James Mirtle @mirtle
Since the start of last season, a span of 86 games, the Leafs have had a better possession rate than last night only five times.

James Mirtle @mirtle
Three of those games were against Buffalo.

BeLeafer
10-15-2014, 01:55 PM
Interesting stats. I'm still not completely sold on these type of stats, but I have a question about the offensive zone starts.

It seems you have suggested that this impacts the interpretation of the possession nos. (Corsi and Fenwick). But don't the players/line's play affect where the faceoffs are taken. Yes, I know the coach often decides who goes in for an offensive or defensive faceoff, but it is also affected by a line's play (e.g., icing will result in a defensive zone faceoff where the coach can't change it).

zeke
10-15-2014, 02:05 PM
hmm, i wonder.

it's really on icing that dictates a line has to take a certain faceoff though, no? nothing else comes to mind.

how many icings are there during a game?

zeke
10-15-2014, 02:09 PM
And I don't think you have to be "completely sold" on those types of stats to use them.

I for one think that its becoming clearer and clearer that they are just a descriptor of team performance, and not an actual measure of individual talent. This makes comparing players on different teams near impossible, though comparing teams to each other as a whole makes sense when judging how they've played.

These stats have their limitations but i think they're very good at describing a team's performance, and at least useful in comparing teammates to each other (though you have to kind of guesstimate how the various competition factors weigh in, which makes it very tough, and the lack of a clear way to do that probably indicates a fundamental weakness in the stats when measuring individual players).

BeLeafer
10-15-2014, 02:10 PM
I need to pay closer attention to this to get a better idea.

However, the key thing is to my mind, how often are the defensive/offensive zone starts the result of a line change? Otherwise, the player and his line have contributed to where the faceoff is taken.

zeke
10-15-2014, 02:13 PM
I think more and more the coaches change lines when the faceoff comes in the offensive or defensive zones. They won't let a "weak" line take a defensive zone faceoff. I could be wrong.

number17
10-15-2014, 02:18 PM
If you ask me how the team and its players did in possession, these stats very much reflect my impression of them. So it's not like the stats are a revelation (except in the case of Percy, I didn't know how good he was and how infavorable his starts were). Instead, they confirm what I believe my eyes told me.

And, given we've played 3 of 4 games at home, and Carlyle's had the last change, my first question for Carlyle, especially if I were Dubas, is "Why did you keep throwing your 1st line out there against opponent's top line? Why not use the 3rd line more as a checking line? and avoid the toughest matchup against your own top line?!"

BeLeafer
10-15-2014, 02:20 PM
And I don't think you have to be "completely sold" on those types of stats to use them.

I for one think that its becoming clearer and clearer that they are just a descriptor of team performance, and not an actual measure of individual talent. This makes comparing players on different teams near impossible, though comparing teams to each other as a whole makes sense when judging how they've played.

These stats have their limitations but i think they're very good at describing a team's performance, and at least useful in comparing teammates to each other (though you have to kind of guesstimate how the various competition factors weigh in, which makes it very tough, and the lack of a clear way to do that probably indicates a fundamental weakness in the stats when measuring individual players).

I think you are right about that. Coaching seems to me to have a big impact on these stats. Recall Maurice getting the team to fire any and everywhere ... lots of pucks directed at the opponents' net but not many wins.

And that's where I'm not sold. It has been a while and I have not spent much time researching this, but did when Mirtle was making claims about the relationship between possession and shooting percentage driving wins. I took a look at the sources he was using and they were incredibly crude regression modelling (simple linear regression). I have yet to take another look, but what I saw suggested that people were making pretty big leaps in concluding what these numbers mean. Not saying it's wrong, just that they may involve a leap of faith.

BeLeafer
10-15-2014, 02:22 PM
I think more and more the coaches change lines when the faceoff comes in the offensive or defensive zones. They won't let a "weak" line take a defensive zone faceoff. I could be wrong.

It does seem that way, but it would be good to have the numbers. For instance, does he always yank Kadri if his line ends up on the defensive zone? Carlyle seems more intent on doing this than other coaches, but it would be useful to know the frequencies of line changes for these zone starts.

zeke
10-15-2014, 02:24 PM
Actually the Maurice (and Wilson) Leafs' teams advanced stats actually confirm what this board had mostly been arguing the whole time - that they were pretty good teams, with scoring depth making up for lack of first line stars, and decent defenses, that were entirely undermined by league worst goaltending every year.

BeLeafer
10-15-2014, 02:31 PM
Actually the Maurice (and Wilson) Leafs' teams advanced stats actually confirm what this board had mostly been arguing the whole time - that they were pretty good teams, with scoring depth making up for lack of first line stars, and decent defenses, that were entirely undermined by league worst goaltending every year.

This only increases my skepticism!

zeke
10-15-2014, 02:37 PM
How are Toskala, Raycroft, and Gustavsson doing now, anyways?

MindzEye
10-15-2014, 02:39 PM
How are Toskala, Raycroft, and Gustavsson doing now, anyways?



Raycroft was pretty shit in an Italian league a few years after leaving the Leafs if I remember correctly

ForeverTML
10-15-2014, 02:46 PM
From wikipedia:
"On April 9, 2014, Raycroft announced his retirement from professional hockey"

BeLeafer
10-15-2014, 02:55 PM
How are Toskala, Raycroft, and Gustavsson doing now, anyways?

Univariate explanations are always the worst when dealing with a complex outcome. See record since these goalies left town.

MindzEye
10-15-2014, 02:57 PM
Univariate explanations are always the worst when dealing with a complex outcome. See record since these goalies left town.

How about just our records with other goalies? I mean, Reimer has had a pretty decent record overall in Toronto, and Bernier's record last year wasn't bad either.

zeke
10-15-2014, 03:09 PM
2006-07: 32.7shpg (3rd), 3.10gpg (8th) ---- 28.4shapg (6th), .888sv% (27th), 3.20gapg (25th)
2007-08: 31.0shpg (5th), 2.78gpg (11th) --- 29.3shapg (17th), .893sv% (29th), 3.12gapg (27th)
2008-09: 31.7shpg (9th), 2.98gpg (10th) --- 30.3shapg (17th), .885sv% (30th), 3.49gapg (30th)
2009-10: 32.6shpg (5th), 2.56gpg (25th) --- 29.8shppg (17th), .892sv% (30th), 3.21gapg (29th)
2010-11: 28.8shpg (25th), 2.60gpg (21st) -- 31.0shapg (18th), .904sv% (20th), 2.99gapg (25th)
2011-12: 28.3shpg (21st), 2.77gpg (10th) -- 30.8shapg (23rd), .898sv% (29th), 3.16gapg (29th)
2012-13: 26.3shpg (28th), 3.02gpg (6th) --- 32.3shapg (27th), .917sv% (7th), 2.67gapg (14th)
2013-14: 27.9shpg (25th), 2.71gpg (14th) -- 35.9shapg (30th), .914sv% (9th), 3.07gapg (26th)

It's pretty fricken hilarious that as soon as we finally got good goaltending, we switched to a system which killed our possession game.

unbelievable, really.

You'll also notice that the offense was great up until 2009-10 (though with a very unlucky SH% in 09-10).....and then Burkie came in, and decided that what we needed was a truculent bottom six, sent out all our scoring depth, which killed our shot creating ability. He did make up for it a bit by going out and getting the likes of Kessel and Lupul, though, which helped keep the offense afloat.

Now imagine a time where we had those elite offensive players AND scoring depth throughout? And a time where we had great goaltending AND a good possesion game?

hilarious that every time we've upgraded one we've gone directly out and sabotaged the other.

zeke
10-15-2014, 03:12 PM
Univariate explanations are always the worst when dealing with a complex outcome. See record since these goalies left town.

that's why it's a multivariate explanation - we went from a good possession team with no first line talent but deep scoring and awful goaltending, to an awful possession team with great goaltending and first line talent but no depth scoring.

record ends up being the same, for completely opposite reasons.

This year we might be putting more than half of those things in place for the first time since before the first lockout - looks like we could have first line scoring, depth scoring, and good goaltending all at the same time this year. Might be good enough to get us into the playoffs.

Now if we could possibly add the fourth pillar - possession - we could be even better than that.

zeke
10-15-2014, 03:14 PM
I like it.

"The Four Pillars of Hockey: Elite Scoring, Scoring Depth, Goaltending, Possession"

write that down.

worm
10-15-2014, 03:28 PM
cant i just copy and paste it?

zeke
10-15-2014, 03:40 PM
at least give it a nice font.

Montana
10-15-2014, 03:57 PM
It's basically the same as what me and PTBNL were saying for years (scoring, possession, goaltending) were the three key elements for predicting future team success.......with scoring split into two seperate categories.

Take those three things, give them proper weight, and you've got the single best predictive tool for teams ability, and probability of future success.

Leafin'
10-15-2014, 04:15 PM
So what do we do with Cody Franson? His point production is nice, and i'm not stat junkie, but his overall impact is not there. If we move out Gardiner in the Staal trade it probably signals that Percy will be with the big club all year. Percy has looked good, but he'll have growing pains as the season wears on.

Phaneuf -
Percy - Polak
Rielly - Robidas

In a somewhat realistic world we'd be able to find another version of Robidas/Gorges RHD veteran defender to play with Dion. Not sure what names are out there that would fit in cap wise as well as be available for a bargain. Unless Robidas can get his legs behind him and we promote one of the younger guys. Though having essentially 3 rookies on the defense spells problematic if we are making a push for the playoffs.

Any names that would fit?

I noticed that Adam Larsson has been a healthy scratch. He's another toolsy defender that has not been able to put it all together on New Jersey. Might be a guy that needs a fresh start somewhere else. Or maybe he's an overhyped swedish defender that busts. If he's available for cheap(say Holland type trade) it might be an avenue worth considering. Not a 1st pairing guy in the vet mold but maybe a Gardiner replacement if he gets moved.

LeafOfFaith
10-15-2014, 06:23 PM
Actually the Maurice (and Wilson) Leafs' teams advanced stats actually confirm what this board had mostly been arguing the whole time - that they were pretty good teams, with scoring depth making up for lack of first line stars, and decent defenses, that were entirely undermined by league worst goaltending every year.

Totally off topic, but wouldn't it be interesting if all of our problems could be laid to rest at the coaches' feet?

If Babcock had been our coach instead of Maurice, Wilson, and Fatass, wonder what might've happened with some of those teams.

I watch us flounder sometimes and wonder how it's possible that we can be so good for stretches and then so bad. 72 great games last year, 10 shitty ones, and out we go.

I've hated all our coaches since Quinn.

Bleedsblue&white
10-15-2014, 06:46 PM
I like it.

"The Four Pillars of Hockey: Elite Scoring, Scoring Depth, Goaltending, Possession"



And surprise.

blacksheep
10-15-2014, 11:08 PM
Totally off topic, but wouldn't it be interesting if all of our problems could be laid to rest at the coaches' feet?

If Babcock had been our coach instead of Maurice, Wilson, and Fatass, wonder what might've happened with some of those teams.

I watch us flounder sometimes and wonder how it's possible that we can be so good for stretches and then so bad. 72 great games last year, 10 shitty ones, and out we go.

I've hated all our coaches since Quinn.

Gee, can't be our coach we just fired all his assistants to force him to change his ways. That's why he didn't, and benched Gardiner when his stats prove Gardiner should be in the game...

Deckie007
10-15-2014, 11:21 PM
Totally off topic, but wouldn't it be interesting if all of our problems could be laid to rest at the coaches' feet?

If Babcock had been our coach instead of Maurice, Wilson, and Fatass, wonder what might've happened with some of those teams.

I watch us flounder sometimes and wonder how it's possible that we can be so good for stretches and then so bad. 72 great games last year, 10 shitty ones, and out we go.

I've hated all our coaches since Quinn.

72 great games last year? We haven't played 72 great games this decade.

JackBurton
10-16-2014, 12:17 AM
The Leafs need to hire a competent defensive coach. They haven't had that in 2 decades.

BeLeafer
10-16-2014, 07:22 AM
that's why it's a multivariate explanation - we went from a good possession team with no first line talent but deep scoring and awful goaltending, to an awful possession team with great goaltending and first line talent but no depth scoring.

record ends up being the same, for completely opposite reasons.

This year we might be putting more than half of those things in place for the first time since before the first lockout - looks like we could have first line scoring, depth scoring, and good goaltending all at the same time this year. Might be good enough to get us into the playoffs.

Now if we could possibly add the fourth pillar - possession - we could be even better than that.

I thought you said they were "entirely undermined" by goaltending? Sure seems like a univariate explanation to me.

number17
10-16-2014, 05:02 PM
The past decade has been the ultimate joke on Leafs fans with a clear mind.

Since Quinn, the Leafs have been on the lookout for a 'defensive coach' and we've been through the likes of Maurice, Wilson ... but meanwhile we've also continued to have the worst goaltending over the past decade. In trying to figure out how to 'fix' our defense when the biggest problem was our goaltending, we've traded bottom line speed and skill for 'defensive minded' goons ... and our GA still didn't improve.

When we FINALLY have goaltending, we also happen to have one of the worst possession coaches in modern NHL history.

I wonder what would happen if Wilson had Bernier and Reimer from the get-go. I would tend to think we'd have got somewhere in the playoff, because despite me not a big fan of Wilson's personality, he was a much better coach than his Leaf record showed.

zeke
10-29-2014, 10:51 AM
NHL Forwards Zone-Start Adjusted 5v5 Corsi Close Percentage (Minimum 50 minutes played):

1. M.Kruger CHI 68.2%
2. D.Winnik TOR 66.7%
3. N.Kadri TOR 65.5%
4. J.Lupul TOR 65.1%
5. C.Kunitz PIT 64.7%
6. P.Hornqvist PIT 64.4%
7. P.Sharp CHI 63.5%
8. S.Crosby PIT 62.5%
9. M.Granlund MIN 62.5%
10. M.Hossa CHI 61.9%


Basically, the three best possession lines in hockey this year have been:

1. The Sidney Crosby line
2. Chicago's second line* (they mix up lines so much that its hard to label them as one line)
3. Toronto's second line

And given how Pittsburgh and Chicago are much better possession teams overall than toronto, the Toronto trio's relative corsi is through the roof.

and Kadri moved to the top line last night and posted a 75% corsi close percentage, Lupul joined him at 69%, while Winnik dropped to 55% on the third line.

number17
10-29-2014, 11:02 AM
I think Lupul's been very good thus far this season, even if he may not have the stats to back him up in the shortened season so far. But I also think Lupul was terrible with or without the puck last season. What was Lup's Corsi close last season?

zeke
10-29-2014, 11:15 AM
Corsi has been a horrific Corsi player in the past.

His play this year has been a complete 180, and its a shame he isn't getting rewarded for it.

number17
10-29-2014, 11:21 AM
You meant Lup I presume.

Yes so Corsi supports what my impression of Lupul's possession game.

zeke
10-29-2014, 11:59 AM
yep.

Metalleaf
10-29-2014, 01:24 PM
Stephen Burtch
Just hoping to get news of this out there but - I have been informed that someone has plagiarized my work on dCorsi to market to NHL teams.

zeke
10-29-2014, 01:26 PM
heh

worm
10-29-2014, 01:51 PM
well now we know where PTBNL went

MindzEye
10-29-2014, 01:56 PM
well now we know where PTBNL went

That would actually not surprise me if PTBNL was working somewhere in the hockey analytics industry, and stopped posting here because of that.

Preston_Mizzi
10-29-2014, 02:16 PM
I'd put my money on suicide

BeLeafer
11-06-2014, 04:02 AM
Hey zeke ... have you checked this guy's work: http://mapleleafshotstove.com/author/evan-dorey/

He does game by game and cumulative scoring chance tracking, paying particular attention to ES and ES close scoring chances. This is the thing Mark Hunter likes to use and it seems to make good sense.

Any idea how to get this data rather than relying on this guy's work?

zeke
11-06-2014, 08:37 AM
looking at his articles he says he's using the Home Plate method (posts to bottom of circle, cut the circles in half from bottom to top through the dot, connect the top of the circles across).

that means they need their own guys to track them every game, which i don't think anyone does.

but he admits in his first article there that when he did it last year it ended up tracking within one percent of corsi so i'm not sure it's worth all the work, especially since that definition of scoring chance is iffy anyways - i don't like removing all point shots as scoring chances, for example, and any scoring chance analysis imo would probably want to account for screens, for whether the goalie is forced to move, and for second chance attempts.

that being said i don't really know those numbers so i could be talking out of my ass.

BeLeafer
11-06-2014, 08:45 AM
Yeah, I read that bit too. He says the reason he doesn't use things like point shots is because determining if they are scoring chances is subjective. The example he used is that it's a scoring chance if the goalie is screened but not so if the goalie can see the puck -- which would involve subjective judgement. That makes sense to me, but may penalize teams that have good point shooters.

The other thing I noted about his methodology is that "close" means within 1 goal in the first two periods or tied in the third and OT. I found that a bit strange, as within 1 goal in the third seems close to me.

Part of the reason I like this more is that scoring chances is much more transparently related to meaningful outcomes -- i.e., goals for/against -- than possession stats.

Damn shame if these stats aren't readily available.

number17
11-06-2014, 08:57 AM
speaking of stats, I am looking for 1 stat that I have no luck finding ....

I'm looking for penalties drawn and penalties taken, but without the roughing and fighting penalties (with behindthenet includes in their penalty taken stats ... cause the top penalty guys are all the goons).

I'm also looking for penalty draw that gets the team down a man (meaning, exclude all the off-setting minors)

Can't seem to be able to find a good site for it ... if you know one appreciate if you can point me to it

zeke
11-06-2014, 09:49 AM
Yeah, I read that bit too. He says the reason he doesn't use things like point shots is because determining if they are scoring chances is subjective. The example he used is that it's a scoring chance if the goalie is screened but not so if the goalie can see the puck -- which would involve subjective judgement. That makes sense to me, but may penalize teams that have good point shooters.

The other thing I noted about his methodology is that "close" means within 1 goal in the first two periods or tied in the third and OT. I found that a bit strange, as within 1 goal in the third seems close to me.

Part of the reason I like this more is that scoring chances is much more transparently related to meaningful outcomes -- i.e., goals for/against -- than possession stats.

Damn shame if these stats aren't readily available.

the beauty of the corsis and fenwicks is that they can be drawn from the official scoresheets. much easier.

but again its not even clear that a scoring chances analysis is more transparently related to outcomes - as even he admits, it ended up saying the same thing as Corsi did last year anwyas.

zeke
11-06-2014, 09:49 AM
speaking of stats, I am looking for 1 stat that I have no luck finding ....

I'm looking for penalties drawn and penalties taken, but without the roughing and fighting penalties (with behindthenet includes in their penalty taken stats ... cause the top penalty guys are all the goons).

I'm also looking for penalty draw that gets the team down a man (meaning, exclude all the off-setting minors)

Can't seem to be able to find a good site for it ... if you know one appreciate if you can point me to it

http://www.behindthenet.ca/nhl_statistics.php?ds=41&s=46&f1=2014_s&f2=5v5&f5=TOR&c=0+1+3+5+4+6+7+41+42+43+44+45+46#

number17
11-06-2014, 11:16 AM
http://www.behindthenet.ca/nhl_statistics.php?ds=41&s=46&f1=2014_s&f2=5v5&f5=TOR&c=0+1+3+5+4+6+7+41+42+43+44+45+46#But this is ALL the penalties they took, including offsetting minors (which doesn't get his team down a man) and offsetting fighting majors (which again, doesn't get his team down a man).

For example, we see Dion Phaneuf's 1.4 min of penalty drawn / 60 ranks among the highest in the NHL, but I think (at least i HOPE) a lot of that is fighting / roughing which is offseting and doesn't get his team into shorthanded situation.

leafman101
11-08-2014, 11:05 AM
Maybe I am late to the party on this one, I am not sure. But I was thinking about Corsi and I think the general public, the media in particular has misstated the importance of it.

Corsi is not an approximation for possession. I mean it is, you can infer possession from it. That just is not the importance of Corsi. Corsi is just shots directed at the net, relative to the shots allowed directed at the net. That is an approximation for scoring chances. Not just scoring chances, but scoring chance for/against differential.

That is why Corsi is such a valuable predicative tool. At the end of the day hockey is about scoring more goals than the other team. Teams with more scoring chances, on probability will score more, and teams that give up less scoring chances, on probability will give up less goals.

The problem with measuring scoring chances is it is subjective. What is a scoring chance, and what is not? Some measure it by proximity to the net. But we've seen tons of goal scored from everywhere, and on weird bounces in front. Shouldn't a Dion Phaneuf or PK Subban shot from the point be a scoring chance? Corsi overcompensates and captures everything directed at the net. I think over the course of the season and thousands of net attempts the quality of shots will even out, like was found with save percentage.

Of course teams that out chance their opponents likely possess the puck at a higher rate too. But Corsi gives you additional information not available from possession.

leafman101
12-06-2014, 06:44 PM
Really interesting interview with Steve Valiquette at about the 22:00 mark: http://iphone.tsn.ca/tsnpodcasts/TSNAnalytics9.mp3

He is doing some interesting work on goalie stats. He thinks save percentage is flawed because it doesn't capture the quality of shots. So he breaks down shots into green shots and red shots. Green shots being high quality scoring chances, and red the ones the goalies should stop.

The issue with breaking down shots like this is obviously the subjectivity involved in what is a green shot. However, they've broken it down in a way that is objective which is essentially shots where the goalie has less than half a seconds to see the puck clearly before the shot. These are plays such as screens, deflections, any play the crosses the middle of the ice, whether a pass or carrying it across the middle, one timers etc. Green shots go in ~33% of the time.

Even more interesting was he mentioned how most third lines produce a low amount of green shots, except for the Leafs. Komarov, Holland and Winnik apparently rank highly in green shot production.

Its interesting stuff. Valiquette acknowledges that its not perfect, but hopes it sparks discussion and produces something better in the end.

zeke
12-06-2014, 06:57 PM
if they can do it properly, it would work.

but it should be easy for them to prove it either way - if their stat ends up making better predictions it will be adopted. I'm thinking the best way to evaluate it would be determining how good a team's defensive play actually is - with that high number of events and enough variety in quality of competition, a better evaluation of shot quality should be easy to see.

I will say this - the leafs have posted an elite shooting percentage in each of the last 5 years (3rd - 6th - 1st - 7th - 4th) and that's pretty much impossible to do by random chance I think. Seems pretty clear that they are making offensive choices that result in a higher shooting percentage (though not necessarily a better offense).

MindzEye
12-06-2014, 07:15 PM
Leafs aren't the only team doing that either. I haven't gone through the numbers recently, but I remember Toronto, Pittsburgh, Tampa, and Washington ranking high in SH% more often that "luck" would suggest is likely.

zeke
12-06-2014, 07:57 PM
yeah, and to be specific that's 5v5 sh%, nothing to do with special teams.

zeke
12-08-2014, 04:13 PM
interesting to break down our possession by line.

zone start adjusted 5v5 corsi (all situations):

Bozak/kessel/jvr: 45.2/46.6/43.6 - ~45.1% avg (awful)
Kadri/Lupul/santo: 52.7/56.2/49.8 - ~52.3% avg (good)
Holland/Komo/clarkson: 45.7/48.4/47.4 - ~47.2% avg (bad)
Smith/Winnik/Panik: 47.7/50.6/48.3 - ~49.1% avg (average)

Phaneuf/Franson: 48.7/50.6 - ~49.7% avg (average)
Gardiner/Polak: 46.5/45.5 - ~46.0% avg (bad)
Rielly/Robidas: 49.4/46.6 - ~48.0% avg (bad)


zone start adjusted 5v5 corsi (score close situations):

Bozak/Kessel/JVR: 43.6/46.9/44.0 - ~44.8% avg (horrawful)
Kadri/Lupul/Santo: 54.5/60.5/47.0 - ~53.0% avg (good)
Holland/Komo/clarkson: 48.3/47.0/48.6 - ~47.7% avg (bad)
Smith/Winnik/Panik: 49.3/50.3/54.6 - ~51.8% avg (good)

Phaneuf/Franson: 48.4/50.4 - ~49.4% avg (average)
Gardiner/Polak: 47.1/48.2 - ~47.6% avg (bad)
Rielly/Robidas: 52.3/45.0 - ~48.5% avg (average)

so in short:

Great: none
Good: 2nd line
Average: 1st pair, 3rd pair, 4th line
Bad: 3rd line, 2nd pair
Awful: 1st line

obviously, the first line is killing us possession wise.

i also think rielly should be playing much more.

MindzEye
12-08-2014, 04:30 PM
The first line should have been broken up weeks ago.

and yes, Rielly should be playing more with Gardiner & Robidas playing less. Rielly looks ready for 21-22mpg.

leafman101
12-08-2014, 04:53 PM
Kadri is starting to take the steps forward you wished Bozak would have 3-4 years ago. PP production aside Kadri has been very, very good all around this year.

Now he is starting to get rewarded. Matched up against the Sedins last game, out there in the final minutes, moved to the top PP unit. Now if we could just get another good top 6 center so that one of the most potent lines in the league is no longer a liability...

zeke
12-08-2014, 05:00 PM
split the top line completely.


Komo - Kadri - Kessel
Lupul - Bozak - Santo
JVR - Holland - xxxxx

leafman101
12-08-2014, 05:20 PM
Sure, for now. At some point we're going to want a center that plays a little bit tougher than Bozak for those kinds of minutes though. His softness without the puck is just devastating. He's good on the draw, he can play the pk, he can produce. But if you are playing in a 7 game series for the cup, you are going to want a player that can win a hell of a lot more puck battles than Bozak does.

Those guys aren't easy to find though. Bozak is fine for now.

Its great that Kadri is turning into one of those players.

leafman101
12-09-2014, 12:38 PM
https://pbs.twimg.com/media/B4a_FMbCIAAumcf.jpg


Possession measured in time. Imagine that.

zeke
12-09-2014, 02:22 PM
wish they'd give us those numbers so we could test them.

leafman101
12-09-2014, 02:24 PM
Since 2002-03 Bozak is 29th worst center in the NHL (with over 200 gp) with a 46.39 ES Corsi. The only skilled player lower on the list than him is Tyler Ennis. Otherwise its filled with 4th liners like McClement, Marchant, Hall, Brodziak, Sillinger, Knopka, Wayne Primeau etc.

Since Bozak came into the league in 2009-10 he is 20th worst (out of 155).

zeke
12-09-2014, 02:26 PM
interestingly, his career shooting percentage is 17.0%.

which is off the charts, and would be dismissed as a completely unsustainable small sample size fluke.......except that its for his whole career.

leafman101
12-09-2014, 02:32 PM
Well evidently its sustainable. Would just be nice if he took a lot more of those shots, rather than giving them up seeing as how many of both of those are going in.

zeke
12-09-2014, 02:38 PM
active career shooting percentage (min 200 shots):

1. A.Tanguay 18.9 (1389shots)
2. S.Stamkos 17.5 (1424)
3. T.Bozak 17.0 (488)
4. B.Morrow 16.0 (1654)
5. M.Perreault 15.6 (346)
6. A.Henrique 15.5 (386)
7. B.Marchand 15.5 (647)
8. G.Nyquist 15.3 (287)
9. J.Toews 15.1 (1363)
10. M.Ribeiro 15.0 (1403)


that tanguay number is pretty crazy. especially since he's been bad for so long.

leafman101
12-09-2014, 08:06 PM
Stephen Burtch ‏

Here are Holland and Bozak's CF% graphs by game and probability distribution:

https://pbs.twimg.com/media/B4dLTWvCIAAQOfQ.png

Here is that same graph since Nov 1st... notice how much further to the right Holland's CF% prob dist is now:

https://pbs.twimg.com/media/B4dL1O7CAAA-hgE.png

Holland's possession results are significantly superior to Bozak's since Nov 1st. likely due to better line mates, but also more OZ starts

also worth noting that Bozak's CF% drop roughly coincides with the talk of Kessel being injured... weird how that would impact?!

Here is that same comparison since Nov 1st with Holland and Kadri - both on the upswing:

https://pbs.twimg.com/media/B4dMejoCIAAr0eO.png

More Holland and Kadri.

leafman101
12-20-2014, 03:56 PM
Travis Yost is tracking an interesting set of events, sustained offensive zone pressure measured by the number of multi shot possessions on a single zone entry.


Ideally, we want to see a team generate two or more shot-attempts upon entering the zone. In order to pull this data out, we must have time constraints. Here, any instance in which two or more shot-attempts were recorded within 10-seconds of one another counts as a 'multi-shot shift'. No neutral or defensive zone event must occur between these shot-attempts – if it does, we know the puck has left the offensive zone. Any data immediately following an offensive zone face-off must also be eliminated, as will any instance in which there is a stoppage of play following the first shot-attempt. Not only can we observe efficiency at generating multi-shot shifts upon entry, we can reverse it and see how well a team does at avoiding those multi-shot shifts against.


http://www.tsn.ca/polopoly_fs/1.162888!/fileimage/httpImage/image.jpg_gen/derivatives/default/yost-graph-eastern-conference-multi-shot-shifts.jpg

Pretty much confirms what we already know. The Leafs are very, very good with the puck, and maintaining possession when they have it, and using that possession to create scoring chances. But, they are worse without the puck.

Habspatrol
01-24-2015, 02:22 PM
Dan Rosen:
• Gary Bettman says the NHL will be testing a player tracking system for quantitative data tonight at skills and tomorrow at the all-star game.
• Commissioner Bettman said the goal of player tracking is to collect and maintain a digital record of everything in the game.
• The tracking will be done by Sportsvision chips embedded into pucks and player gear.

Was wondering when they'd start this. Pretty cool.

Hairnova posted this on the Habs board.

This has some awesome potential. Could potentially revolutionize advanced hockey statistics.

MindzEye
01-24-2015, 05:39 PM
I remember suggesting something like that (chips in the pucks and players) was technologically feasible when we were discussing advanced stats on here a few years ago. It makes a **** ton of sense.

Can't wait to see the data it generates.

Blueman
01-24-2015, 07:57 PM
I remember suggesting something like that (chips in the pucks and players) was technologically feasible when we were discussing advanced stats on here a few years ago. It makes a **** ton of sense.

Can't wait to see the data it generates.

Agreed - need to get more tech. Also, would be nice to have a puck sensor to digitally tell if it crosses the goal line. Or something like they have in Tennis at least.

hairnova
01-24-2015, 08:38 PM
Big news - NHL.com is going to start tracking advanced stats, beginning next month according to Chris Johnston.

hairnova
01-25-2015, 07:13 PM
Nirva Milord: Giving the NHL All-Star media a peek at the Sportvision prototype.

https://pbs.twimg.com/media/B8O9ExCIgAEGjiK.jpg

zeke
01-26-2015, 11:31 AM
it'll be interesting to see if possession time ends up being more important than shot attempts.

MindzEye
01-26-2015, 01:21 PM
it'll be interesting to see if possession time ends up being more important than shot attempts.

Was thinking the same thing.

I could see offensive zone possession time and multiple shot possessions ending up being the trump card.

hockeylover
02-23-2015, 05:30 PM
http://espn.go.com/espn/feature/story/_/id/12331388/the-great-analytics-rankings#

Great Analytics Rankings...

Chicago was the only team considered "All In" but here's the Leafs write-up:


There hasn't been a more dramatic analytics turnaround than in Toronto, as the Maple Leafs became a team that went from mocking analytics to making a push for the "all-in" category.

It's been just over a year since GM Dave Nonis revealed that the team's six-figure analytics budget went unspent year after year. Since Brendan Shanahan grabbed the reins as team president, the spending began.

The big hire was assistant GM Kyle Dubas -- based on his successful use of analytics to build the Canadian junior hockey Sault Ste. Marie Greyhounds. Dubas has since built an analytics team that clearly has influence, evidenced by the Leafs' late-summer free-agent additions and the midseason coaching change.

"As long as you have that interest and curiosity from the organization, and you can present it in a way that it's easily comprehended, it starts to have an impact," Dubas said earlier this season.

Another high-profile hire was Darryl Metcalf, who founded ExtraSkater.com, a site that provided in-depth player and game stats.

The on-ice results aren't there yet, but this is a team in need of a long-term build, not just a few tweaks.

Montreal and Detroit were in the "One Foot In" category. Ottawa was classified as "Skeptics", Colorado called "Non-Believers".

Metalleaf
02-27-2015, 02:31 PM
Stephen Burtch
The only goalies ahead of Bernier in 5v5 Adjusted SV% since Oct 2013 are Carey Price and Tuukka Rask. Semyon Varlamov ranks 4th.

leafman101
02-28-2015, 09:03 AM
‏@mc79hockey
Kyle Dubas dropping graphs.

https://pbs.twimg.com/media/B-79ok2W0AA4R0Q.jpg


A little different from the last few Leafs presentations about how stats are useless at the Sloan Conference.

leafman101
02-28-2015, 09:10 AM
James Mirtle
Here @kyledubas compares the Leafs (in blue) to the Greyhounds improvement during his three years there (in red)

https://pbs.twimg.com/media/B-8Acr5VAAE0rKr.jpg

leafman101
02-28-2015, 11:15 AM
James Mirtle ‏
Dubas said fixing an NHL team takes longer than an OHL one. Draft takes longer to affect roster and turnover is slower.

Dubas: "There’s teams that are so far ahead of us in our own league." Explains Leafs trying to play catch up using some new innovations.

Pucklosopher
02-28-2015, 12:09 PM
Why do the shot attempts have to go down before they go up? Just the players adjusting to the system?

leafman101
02-28-2015, 12:16 PM
Its measuring the total start differential from the start of the season. E.g, after game 1 you have a -1 shot differental, then game 2 is also a -1, overall you are at -2 now.
Its going down because they are a shitty team.

hockeylover
02-28-2015, 12:20 PM
Dubas on the "Eye Test": Your eyes and your mind are lying SOBs.

Dubas on inefficiencies in the marketplace: They don't last long. Eg. I don't think Daniel Winnik will be available as UFA again on July 29

hockeylover
02-28-2015, 07:30 PM
http://news.nationalpost.com/2015/02/28/toronto-maple-leafs-assistant-gm-kyle-dubas-promotes-analytics-while-brian-burke-stands-behind-old-school-scouting/

Toronto Maple Leafs assistant GM Kyle Dubas promotes analytics while Brian Burke stands behind old-school scouting

Metalleaf
02-28-2015, 07:50 PM
So this is in fact the Renaissance Age of the Maple Leafs? After years of living in the Stone Age.

leafman101
03-01-2015, 11:23 AM
The difference in the messaging between Leafs management, old and new, couldn’t be more stark.

“The notion that you can sit behind a computer and find athletes is bullshit,” Burke grumbled during his time on panel with Nate Silver, among others. Later, he added “this is still an eyeballs business.”

“What analytics taught me is your eyes and your mind are lying sons of bitches in the worst absolute way,” Dubas said earlier in the day, directly refuting Burke’s eyeballs adage.

http://www.theglobeandmail.com/sports/hockey/dubas-reveals-leafs-mindset-on-analytics-at-sloan/article23238104/?click=sf_globe

Burke may have been worse for this team than JFJ. Forget the moves he did make. He set this team back another 5 years by being an old, stubborn know it all that managed the team with an outdated method.

BeLeafer
03-01-2015, 11:31 AM
We still have a bunch of dinosaurs in the scouting department who, as recently as 12 months ago, were completely trashing analytics. Why Steve Kasper is still on staff is beyond me. Thankfully, they've been largely avoiding trades for pros.

hockeylover
03-01-2015, 11:34 AM
http://www.theglobeandmail.com/sports/hockey/dubas-reveals-leafs-mindset-on-analytics-at-sloan/article23238104/?click=sf_globe

Burke may have been worse for this team than JFJ. Forget the moves he did make. He set this team back another 5 years by being an old, stubborn know it all that managed the team with an outdated method.

Dunno about that. Rask, 1st, 2nd (packaged to move to 9th, picked Logan Couture), 4th for Raycroft and Toskala still haunts my dreams.

BeLeafer
03-01-2015, 11:36 AM
Burke was net neutral. Some good stuff and some awful stuff. But, no, he was nowhere near JFJ's orbit of calamity.

leafman101
03-01-2015, 11:38 AM
Burke just flat out built a bad organization. From management down. It was 5 wasted years, and who knows how long it will take to recover.

JFJ did some damage too. Its a tough one.

A trade can be good or bad and have long lasting implications. But poor management effects everything. Burke was a disaster.

BeLeafer
03-01-2015, 11:42 AM
Burke just flat out built a bad organization. From management down. It was 5 wasted years, and who knows how long it will take to recover.

JFJ did some damage too. Its a tough one.

A trade can be good or bad and have long lasting implications. But poor management effects everything. Burke was a disaster.

Well, there is that. He did bring in a lot of meatheads. Fair point.

leafman101
03-01-2015, 11:43 AM
I mean even if they just didn't completely ignore analytics, not necissarily become leaders in the industry, just not ignore it, the team would be miles ahead of where it is now. And the budget for analytics was there too. It makes no sense. Just pure arrogance, which is the absolute worst quality for the president of anything. That move is up there for the most ridiculous things a Toronto GM has done.

zeke
10-30-2016, 03:42 PM
OCT '15 STANDINGS (pts%) ----- Final Standings

1.MTL .833 ------- .500 (#22)
2.NSH .833 ------- .484 (#14)
3.DAL .800 -------- .665 (#2)
4.WSH .778 ------- .741 (#1)
5.MIN .750 -------- .530 (#17)


OCT '15 AdjusteD CF% ------- final standings

1.STL 55.6% -------- .652 (#3)
2.LAK 55.4% -------- .622 (#7)
3.WSH 54.2% ------- .741 (#1)
4.DAL 54.1% -------- .665 (#2)
5.CAR 52.7% -------- .524 (#18)

BKerr
10-30-2016, 03:51 PM
5v5 tied: 55.4% (6th)

Seems pretty good to me...

Habsy
10-30-2016, 04:27 PM
OCT '15 STANDINGS (pts%) ----- Final Standings

1.MTL .833 ------- .500 (#22)
2.NSH .833 ------- .484 (#14)
3.DAL .800 -------- .665 (#2)
4.WSH .778 ------- .741 (#1)
5.MIN .750 -------- .530 (#17)


OCT '15 AdjusteD CF% ------- final standings

1.STL 55.6% -------- .652 (#3)
2.LAK 55.4% -------- .622 (#7)
3.WSH 54.2% ------- .741 (#1)
4.DAL 54.1% -------- .665 (#2)
5.CAR 52.7% -------- .524 (#18)
You know what that means in Oct 2016? Jack shit.

BKerr
10-30-2016, 05:13 PM
When the Leafs had a great record under Carlyle to start the year but were getting solid goaltending and had really bad possession numbers, it was all you could talk about.

6th in Corsi while tied....
14th overall in corsi close....
8th in Fenwick tied
14th in Fenwick close

The Habs dont have "really bad" possession numbers.

That 14th number isn't ideal, but it doesn't qualify as "really bad" possession numbers equivalent to the Carlyle Leafs now does it?

hockeylover
10-30-2016, 05:18 PM
6th in Corsi stats while tied....
14th overall in corsi close....

The Habs dont have "really bad" possession numbers.

Not what I meant. When the Leafs have bad possession numbers and a good record, you go on and on about how they're going to regress. But now they have great possession and an average record and you seem to be in with the "points are all that matters" fans.

BKerr
10-30-2016, 05:20 PM
Not what I meant. When the Leafs have bad possession numbers and a good record, you go on and on about how they're going to regress. But now they have great possession and an average record and you seem to be in with the "points are all that matters" fans.

I think you missed the tongue in cheek nature of that post.

If you can't have fun trolling the fans of a rival a little bit after a win, well then....

rated
10-30-2016, 05:25 PM
Habs are better than some people are giving credit. With that said, its quite clear that Price is what separates Montreal from being a good team, to an elite team.

Deckie007
10-30-2016, 05:30 PM
Here is the thing, the Habs got 17 points in 9 October games.

Now, you're a leaf fan, so you might not understand why these points matter. See the teams with the most points (more or less, its a little complicated by conferences and divisions) but generally the teams with the most points in April qualify for the this little tournament the NHL holds called the Stanley Cup Playoffs. I know, the Leafs have only participated in one of these tournaments in a really long time, so you might not understand the concept, but its real. The points you get in october, they count the same as the points you get in January, February, March, and the early part of april.

Now how this tournament works is you play the same team as many times as it takes for one of you to win 4 times. The team that wins four advances in the tournament. The team that loses four gets to join the Leafs, who have gotten a head start on the golf course.

So much sass and condescending text from a fan of a team with no playoff success for a couple decades. The Leafs have been awful for years, we all know it too well. Doesn't change that you aren't a strong possession team and that it'll probably cause issues in the playoffs if you don't fix it. But hey, that record is sure spiffy.

BKerr
10-30-2016, 05:31 PM
Habs are better than some people are giving credit. With that said, its quite clear that Price is what separates Montreal from being a good team, to an elite team.

Thats a fair statement.

As I showed yesterday via the stats.... Habs with League Average (917) goaltending instead of what they got (901) after Price was injured... would have given up 28 less goals against and likely would have made the playoffs over Detroit.

Take that team, make Price healthy, make Petry and Gallagher healthy, and add Radulov.... swap Subban/Weber (even if you call that a downgrade in possession)... and they should be pretty good.

BKerr
10-30-2016, 05:33 PM
So much sass and condescending text.

https://ih0.redbubble.net/image.49696960.7899/ra,unisex_tshirt,x1000,101010:01c5ca27c6,front-c,240,146,225,355-pad,220x200,ffffff.u3.jpg

zeke
10-30-2016, 05:35 PM
You know what that means in Oct 2016? Jack shit.

it means that win/loss record in october is not a very good predictor of full year success.

MindzEye
10-30-2016, 06:07 PM
You know what that means in Oct 2016? Jack shit.

"Those who don't understand the past are doomed to repeat it."

- Some Wicked Smaht Guy

MindzEye
10-30-2016, 06:10 PM
6th in Corsi while tied....
14th overall in corsi close....
8th in Fenwick tied
14th in Fenwick close

The Habs dont have "really bad" possession numbers.

That 14th number isn't ideal, but it doesn't qualify as "really bad" possession numbers equivalent to the Carlyle Leafs now does it?

Has anyone actually identified a difference in score effects between "close" and tied? Isn't the theory behind score effects that when you're up or down by a lot, you change the way you play? What exactly is the rationale behind severing a sample into thinner slices here?

zeke
10-30-2016, 06:34 PM
use score adjusted corsi.

it doesn't get rid of any data, justt weights it.

Habsy
10-30-2016, 06:39 PM
Too bad no one actually said those things.
I was paraphrasing. :wink

Habspatrol
10-30-2016, 06:40 PM
Has anyone actually identified a difference in score effects between "close" and tied? Isn't the theory behind score effects that when you're up or down by a lot, you change the way you play? What exactly is the rationale behind severing a sample into thinner slices here?

There's a massive change in play between being tied and down/up a goal... especially later in games. There's no more aggressive team than a team down a goal in the 3rd.

zeke
10-30-2016, 06:41 PM
score

Adjusted

hockeylover
10-30-2016, 06:41 PM
use score adjusted corsi.

it doesn't get rid of any data, justt weights it.

That makes a lot of sense to me.

Habsy
10-30-2016, 06:55 PM
Thats a fair statement.

As I showed yesterday via the stats.... Habs with League Average (917) goaltending instead of what they got (901) after Price was injured... would have given up 28 less goals against and likely would have made the playoffs over Detroit.

Take that team, make Price healthy, make Petry and Gallagher healthy, and add Radulov.... swap Subban/Weber (even if you call that a downgrade in possession)... and they should be pretty good.
You forgot replace Eller with Shaw. I love Shaw's game. The points will come, he does everything else well.

Habsy
10-30-2016, 06:56 PM
it means that win/loss record in october is not a very good predictor of full year success.
I'd agree if the teams were identical. They're not so October 2015 means jack shit and you know it.

Habsy
10-30-2016, 06:57 PM
"Those who don't understand the past are doomed to repeat it."

- Some Wicked Smaht Guy
See my previous post then go pahk ye cah ovah dair.

BKerr
10-30-2016, 07:11 PM
Has anyone actually identified a difference in score effects between "close" and tied? Isn't the theory behind score effects that when you're up or down by a lot, you change the way you play? What exactly is the rationale behind severing a sample into thinner slices here?

I gave both slices.... and let you guys decide what you wanted to use.

I then used the worst of four numbers (which is also the close number) when i said, "That 14th number isn't ideal, but it doesn't qualify as "really bad" possession numbers equivalent to the Carlyle Leafs now does it?"

hockeylover
10-30-2016, 07:25 PM
I gave both slices.... and let you guys decide what you wanted to use.

I then used the worst of four numbers (which is also the close number) when i said, "That 14th number isn't ideal, but it doesn't qualify as "really bad" possession numbers equivalent to the Carlyle Leafs now does it?"

Yeah, I didn't actually say that last part and I've explained that reference already.

BKerr
10-30-2016, 07:28 PM
Yeah, I didn't actually say that last part and I've explained that reference already.
Fair enough. I just wasn't going to change my quote

Krustchev
10-30-2016, 07:39 PM
October champs!

Have we played any other months this season?

Krustchev
10-30-2016, 07:41 PM
I'm always amazed how much hab fans care what we think.

Yes, it's unusual that people want to discuss things on a discussion board.

Aberdeen
10-30-2016, 07:42 PM
Yes, it's unusual that people want to discuss things on a discussion board.
Read my other responses. Its how emotional you guys were getting that amused me.

hockeylover
10-30-2016, 07:44 PM
Was that the Bozak line again?... The hell?

Leafin'
10-30-2016, 07:48 PM
Embarrassing.

If we are going to shit the bed this year(i know its early) we might as well move all the vets and tank early. 4th last after this game.

Bozak has gotta go. GARBAGE

angelfish
10-30-2016, 11:00 PM
4274

worm
10-31-2016, 11:25 AM
I'm not sure that's it.

I think we're taking solace in the fact that this is a young team that's doing a lot of things right and not seeing wins.

Which, after years of watching teams that deserve to get blown out, seems a little unfair.

However, young players are just prone to some bad habits only experience grinds out of their game.

I'm okay with them doing a lot of things right, scoring a lot of goals and losing this year. But this is the last year I'll be okay with it.

yep

the big joke out in Edmonton is the lack of "moral victories" this year

worm
10-31-2016, 11:32 AM
Find me a Stanley Cup team that hasn't been a contender for the "Corsi Cup" over the last bunch of season.


I'll wait.

Yep. Habs need to improve and have Price be on a run in the post season.

What site do you use for it though?

I was doing a quick look at the habs were not as bad as I thought.

worm
10-31-2016, 11:38 AM
I'm always amazed how much hab fans care what we think.

meh

we have been chatting with you guys for years

MindzEye
10-31-2016, 01:58 PM
Yep. Habs need to improve and have Price be on a run in the post season.

What site do you use for it though?

I was doing a quick look at the habs were not as bad as I thought.

There's a few.

Stats.hockeyanalysis.com

Corsica.hockey

Puckon.net

Montreal is 19th in Corsi Close according to Puckon, 11th in score adjusted Corsi according to Corsica. Even the kinder of the two doesn't put Montreal anywhere near where you'll find almost all Cup winning teams.

Pittsburgh (3rd), Chicago (2nd), LA (1st), Chicago (2nd), LA (1st), Boston (14th), Chicago (1st)

So the rule of thumb is, win the Corsi cup or have a goalie put up a .940 with Chara and Bergeron in front of him in all key situations.

Habsy
10-31-2016, 10:04 PM
Itís hinted above that some of Montrealís record is the result of unstable percentages, and thatís true, but itís important not to overlook the fact that in addition to those wicked percentages, the Canadiens are doing a good job in terms of possession. Montreal has played just over three hours of score-tied hockey at five-on-five, and have a 54 percent Corsi rating in those situations. Thatís an elite number (only Los Angeles topped it last year), and when itís combined with goaltender Carey Price and a bit of finish, it may well be insurmountable.

http://m.bleacherreport.com/articles/2672966-nhl-power-rankings-br-experts-week-4-poll/page/15

MindzEye
10-31-2016, 10:28 PM
I ask again why being tied is important? We know that being up or down by a lot can skew Corsi because teams purposefully change the way that they play. We already adjust for this in "close" or "adjusted" metrics. So why should anyone take a thin slice of the close or adjusted corsi seriously? Is there some work that's been done by anyone to suggest that "tied" corsi is specifically important, or is it just a smaller sample that the Habs currently look good in?

I know the answer. You won't like it.

Habspatrol
10-31-2016, 10:34 PM
I ask again why being tied is important? We know that being up or down by a lot can skew Corsi because teams purposefully change the way that they play. We already adjust for this in "close" or "adjusted" metrics. So why should anyone take a thin slice of the close or adjusted corsi seriously? Is there some work that's been done by anyone to suggest that "tied" corsi is specifically important, or is it just a smaller sample that the Habs currently look good in?

I know the answer. You won't like it.

I gave my reasoning, you didn't like it?

BKerr
10-31-2016, 10:42 PM
I think time left in the game matters too.

a 1 goal lead in the first period or in the second likely doesn't change how teams play.

a 1 goal lead with 10 minutes left in the third? I've seen enough hockey over the years to know that possession is often controlled by the team trailing.

I like corsi close, but i think it can even be modified further as there are more score effects out there than just a two-or-more goal lead.

zeke
10-31-2016, 10:46 PM
score

adjusted

Habspatrol
10-31-2016, 10:46 PM
Only 8 teams in the league are currently over 50% CF with a 1 goal lead.
Last year only 3 teams were over 50% and those percentages were 53.3%, 50.7% and 50.4%

So yeah, having a lead definitely has a big affect on Corsi.

Habsy
10-31-2016, 10:47 PM
Only 8 teams in the league are currently over 50% CF with a 1 goal lead.
Last year only 3 teams were over 50% and those percentages were 53.3%, 50.7% and 50.4%

So yeah, having a lead definitely has a big affect on Corsi.
Ya think? It doesn't play into their narrative though. Damn those Habs for leading most of their games.

Habspatrol
10-31-2016, 10:48 PM
score

adjusted

Can you give me a link to where I can find it?

I like the theory, just skeptical of the weighting.

zeke
10-31-2016, 10:57 PM
the historical number of events for either team is counted for each possible score state from the perspective of the home team. For instance, when trailing by one goal, home teams have recorded 51,921 unblocked 5v5 shots while their opponents (the away team) have recorded 43,075 between 2007 and 2014. The adjustment coefficients, or weights, are selected in order to satisfy this ratio while producing a total quantity of weighted shots that is equal to the original unadjusted total. The coefficients are given by: coef(team) = [average # of events]/[# of Events for team]. I calculated these coefficients for each of shots, unblocked shots, shots on goal and goals using the complete (at the time) data set since 2007, staying true to Micah’s original formula. These weights are used for “Score and Venue Adjusted” stats.


http://www.corsica.hockey/blog/2016/06/19/adjustments-explained/

MindzEye
10-31-2016, 10:57 PM
I gave my reasoning, you didn't like it?

I didn't see it

Habspatrol
10-31-2016, 11:00 PM
I didn't see it

There's a massive change in play between being tied and down/up a goal... especially later in games. There's no more aggressive team than a team down a goal in the 3rd.

It's supported by the fact that very few teams are over 50% CF when leading.

Habspatrol
10-31-2016, 11:13 PM
Figured this discussion made more sense for this thread.

zeke
10-31-2016, 11:29 PM
I may start using xGF% instead of corsi.

leafman101
11-01-2016, 08:33 AM
Yeah I really like expected goals for. I think that is probably getting closer to what teams are using.

I really liked war-on-ice's WAR and high danger scoring chance stats, but I suspect those are a couple key reasons the Pens bought them out.

As mentioned in the other thread Corsi's large sample size provides great value for approximating possession. And it has been shown that a team that possesses the puck more is more likely to win in the future. But really I think the more important tool is dangerous shots directed at the net. Or at least opportunities for dangerous shots. While possession is a good predictor of team goals for%, ultimately you are trying to predict goals themselves, so I would think scoring chances does a better job of that, particularly for individual players.

I think factoring in the quality of the shots, as xGF does, is important. All shots are not created equal. So where Corsi captures a larger sample size, it also captures a a larger sample size of less meaningful events. And now that we have years of data, the sample size is less of a concern.

number17
11-01-2016, 09:03 AM
How is xGW calculated though? Sorry haven't caught up with the new advanced stats for a while.

Corsi is a good stat, but it's shots directed at the net and it can also be a shot from the blueline (which, unless you are Bernier or Vesa, is not gonna go in ... it doesn't mean much) but given the large enough sample size, it is representative of how good your team is offensively.

I like Scoring chances but my only gripe is it is somewhat subjective. Plus as you said, it has a smaller sample size and is less reliable at this point

leafman101
11-01-2016, 09:17 AM
xGF/GA is a measure of shot quality from Corsica. The actual formula for these things never gets revealed, but basically they take into account the different categories of shots and the odds of scoring on each of those types of shots. So its a measure of the quality of shots a player/team takes and gives up.


Here’s what my model does account for:

Shot type (Wrist shot, slap shot, deflection, etc.)
Shot distance (Adjusted4 distance from net)
Shot angle (Angle in absolute degrees from the central line normal to the goal line)
Rebounds (Boolean – Whether or not the shot was a rebound)
Rush shots (Boolean – Whether or not the shot was a rush shot)
Strength state (Boolean – Whether or not the shot was taken on the powerplay)5

Each of the six shot types6 provided by the NHL forms its own category, and these are further subsetted by rebound and non-rebound. Only unblocked shots are used due to the unfortunate fact blocked shot coordinates are unavailable.7 These twelve shot sub-types are regressed independently according to the remaining variables. The rationale here is that each shot subset should respond to the variables differently. Namely, distance and angle do not influence slap shot quality in the same manner as they do, say, deflections. In addition, the relationship between goal expectancy and distance or angle are not assumed to be linear. That is to say, the model is not bound by the idea that shot quality changes at a constant rate along the scales of distance and angle. I found these modifications improved the model’s ability to assess goal expectancy.8

According to their math its xGF% is not a better predictor of future goals than CF%. However xG/GA is maybe the best predictor of future goals out there. Better than goals/60.


This idea of projecting future outcomes is of great importance in analyses relating to hockey and indeed a great number of fields. In its present condition, 5v5 xGF% is not a better predictor of future 5v5 GF% than CF% at the player level. Regular skaters’ 5v5 xGF% in one >400 TOI season did not yield a higher correlation with the next season’s 5v5 GF% than the same test performed with CF%.12 The same variance observed in early shot quality analyses prevents on-ice xG from predicting real goals, or itself for that matter, in any practical way. Though descriptive of shot quality, the xG model has not yet shown to be appreciably predictive of future shot quality or goals at the on-ice level.13

xG does, however, have predictive value at the individual skater level. 5v5 ixG60 is a better predictor of future 5v5 G60 than G60 itself (0.152 for forwards and 0.128 for defence vs. 0.140 and 0.076 respectively).

http://www.corsica.hockey/blog/2016/03/03/shot-quality-and-expected-goals-part-i/