Has countering puck possession been successfully figured out?

[Bounces R Way]

Good goaltending>>>Corsi

every single time

 

[RageQuit77]

Corsi-centered approach to the game brought to it's logical ultimate edge would mean that amounts of shots are maximized, ragardless how low quality changes they would be. That would mean shooting always when having actual puck possession in O-zone, in which case you could use team's collective GVA stat as a proxy of "possession".

In fact maximal corsi count cannot be reached by maximal low quality shooting (= high total GVA), as more a team lose pucks immediately after a shot, more they lose also possible rebounds, and thus also potential corsi happenings.

Optimal corsi-centered "strategy" is somewhere in between "shot-asap-no-matter-how-low-change" -approach and "try-at-at-least-ensure-some-quality-of-shooting" -approach. Either case, worse corsi team with better high-quality high-danger scoring change production, with high percentage quality shooting snipers and goal scorers with low SOG totals will wash the corsi-first team in a long run.

 

[Machinehead]

Pittsburgh didn't systematically counter Washington, Fleury just played out of his ****ing mind.

Even more so Holtby was drunk out of his mind.

I wish we got that Holtby. Would have saved ten years off my life.

 

[Machinehead]

Good goaltending>>>Corsi

every single time

The opposite is true if you look at the recent history of the Stanley Cup.

Lundqvist, Luongo, Bobrovsky, and Price put together - 0 Cups

Top 5 corsi teams - 7 Cups since the lockout and 3 in the last 5 years.

 

[Whileee]

...don't forget score effects.

 

[JetsFan815]

Awful thread.

 

[TomasHertlsRooster]

In all honesty, I've used last year's Pittsburgh team in the past as an example of how tanking dominated but they are probably not the best model for sustainable success. The 2016 Pittsburgh team is one of the best teams ever and their Corsi reflects that.

That 2017 team, on the other hand, was dominated in 3 of the 4 series that they played and in the one series that they were mostly the better team, they still went to game 7 OT. They had elite Conn Smythe goaltending from a guy with an inconsistent, spotty playoff history when their starter went down with an injury. Then, when that backup started to struggle, the starter got healthy and played at an elite, Conn Smythe level. They also have 3 of the very best finishers in the world who manage to score a lot of goals and assists and maintain a high on-ice shooting percentage. Most teams don't have even one of those players and I can't think of very many that have two forwards with the offensive talents of Kessel, Crosby, and Malkin. The Penguins have 3.

On top of that, even superstar goalies usually don't put together the playoffs that Fleury+Murray did. Playoff performances from goalies can be pretty random and difficult to predict and a guy with a .908 career playoff SV% suddenly putting up a .924% isn't something you can consistently count on from anybody. Then Murray came back and put up a .937%, when his career playoff SV% is .928%.

Then you look at the teams they played against - Pekka Rinne was having a magnificent playoff run and then completely fell apart in the finals. Braden Holtby had the highest playoff SV% of all time and he was very bad against Pittsburgh. Bobrovsky won the Vezina then sucked against them. Yes, part of this comes from the Penguins having good shooters, absolutely, but they struggled to score on Craig Anderson. Given that, you have to admit that at least part of their domination of elite goaltenders could be credited to luck. If they really were so good that they made elite goalies look like AHL scrubs, they should have made a good goalie like Anderson look like an ECHL scrub.

Their #1D was Brian Dumolin. They won the Cup without their #1 defenseman and got dominated in 3 of the 4 series. They had a #1D playing way way over his head and were lucky to head into the finals and play against a team whose #1C was playing way over his head because their actual #1C was injured. They played Columbus who doesn't have an elite #1C., Ottawa who doesn't have a #1C, Nashville whose #1C is a decent borderline guy but was injured, and Washington whose #1C is a legit guy in the regular season but is a huge playoff choker who is so underrated as a regular season player that he doesn't really get flack for his playoff performances. How many teams play 4 rounds and don't play a single round against a team with a #1C? The Penguins did and they were able to win because their #2C is a top-5 center in the league, and they were extremely lucky to get those centers.

Teams are not going to be able to model themselves after Pittsburgh and disregard CF%, or having a single true top-pairing defenseman because they will not have the luxury of having two elite top-5 players who are both centers and they will not have the luxury of having an elite top-10 winger and a Conn Smythe goaltending performance from your backup and another Conn Smythe goaltending performance from your starter just as the backup was falling apart when the starter recovers from injury. Nobody can prepare for that, it takes a ton of luck. The closest thing would be if Edmonton won the Matthews lottery. That would have given them two superstar elite #1Cs and maybe they could have won last year when Talbot and Draisaitl were also playing at Conn Smythe level, but even Edmonton was already so lucky to get McDavid.

Pittsburgh's 3 MVPs were guys drafted 1st, 2nd, and 1st overall in 3 back to back seasons. Their #1D was a guy who was a part of a package that they traded their #2 overall pick (the year after those 3 back to back seasons) for. A lot of teams can't be a bottom-2 team for 4 straight years because it's not financially sustainable and could lead to relocation. Yeah

 

[Pookie]

Look - the simple truth is the team that possess the puck more tends to outscore their opponents. Corsi is a baseline structure that measures shot attempts for versus shot attempts against. Anything can happen in a seven-game series, and the two teams from last year, Pittsburgh and Ottawa, both rode impressive goaltending to their postseason positions. Both teams are nowhere near as dominant this season. Every other team in your short sample size is languishing respectively this season.

You say that number 2, 3, 4, and 9 in possession metrics won't make the playoffs this season. That means everybody else in the top 10 are making the playoffs, as well as several in the teens. Given that fully half the NHL makes the playoffs, your odds are relatively good to make the playoffs. Not everyone with a top 10 possession metric will make the playoffs. That's just how it goes. That those teams are currently out of the playoff picture speaks of underlying issues beyond their metrics. The only thing anybody said about corsi was that it was a probability, not a certainty.

Take last night's Jets game, for instance. The Jets absolutely crushed Anaheim in possession time. It's not even close. Same with LA the previous game. They outshot and out possessed both teams by a ridiculous margin, yet needed overtime in both games to win it. That's just the way sample sizes go.

The team that possesses the puck more tends to outscore their opponent. It's just the simple truth. That's all corsi is a quantifier of.

Except that’s not the case.

More games are won while being outshot.

So, enter the made up construct of Score Effects to help duct tape this theory together. This beauty of a smudge factor suggests that shots taken while winning count differently than shots taken while ahead.

But if total shots over a 60 minute segment of time where really a factor, score effects wouldn’t be an issue.

The simple truth is that puck “possession/shot attempts” is only weakly correlated with results.

 

[TomasHertlsRooster]

2017 Cup Final: #16 CF% beats #5 CF%
11th FF% close beats #16 FF% close
#6 xGF% beats #8 xGF%
#2 points beats #18 points

2016 Cup Final: #2 CF% beats #8 CF%
#6 FF% close beats #4 FF close
#1 xGF% beats #2 xGF%
#4 points beats #11 points

2015 Cup Final:#2 CF% beats #4 CF%
#7 FF% close beats #4 FF% close
#18 xGF% beats #5 xGF%
#7 points beats #5 points

2014 Cup Final: #1 CF% beats #8 CF%
#1 FF% close beats #5 FF% close
(The WCF, which everybody considered the real SCF had #1 CF% and #1 FF% close beating #2 CF% and #2 FF% close)
#4 xGF% beats #7 xGF%
#9 points beats #12 points

2013 Cup Final: #4 CF% beats #3 CF%
#2 FF% close beats #6 FF% close
#6 xGF% beats #8 xGF%
#1 points beats #5 points

2012 Cup Final: #2 CF% beats #13 CF%
#4 FF% close beats #11 FF% close
#6 xGF% beats #9 xGF%
#13 points beats #9 points

2011 Cup Final: #14 CF% beats #6 CF%
#14 FF% close beats #4 FF% close
#16 xGF% beats #20 xGF%
#8 points beats #1 points

2010 Cup Final: #1 CF% beats #14 CF%
#1 FF% close beats #9 FF% close
#1 xGF% beats #8 xGF%
#3 points beats #18 points

CF%: 13 of 16 Cup Finalists in the top-10
6 of 8 Cup winners in the top-5
One Cup Finalist (winner, Pittsburgh 2016) not in the top half, but was 16th.
Average SC winner rank: 5.25
Average SC loser rank: 7.625
Average SC Finalist rank: 6.4375

FF% Close: 12 of 16 Cup Finalists in the top-10
4 of 8 winners in the top-5 (1 other winner #6 one winner #7)
One Cup Finalist (loser) not in the top half, but was 16th
Average SC winner rank: 5.75
Average SC loser rank: 7.375
Average SC Finalist rank: 6.5625

xGF%: 13 of 16 Cup Finalists in the top-10
Three Cup Finalists (two winners, one loser) in the bottom half, winners ranked 16th and 18th, loser ranked 20.
3 of 8 winners in the top-5 (3 winners at #6)
Average SC winner rank: 7.5
Average SC loser rank: 8.375
Average SC Finalist rank: 7.9375

(Remember, a team must be at least 24th in points to make the playoffs, and that assumes that the #1 team in their conference is below the worst team in the other conference.)

Points: 11 of 16 Cup Finalists in the top-10
4 of 8 winners in the top-5
Two cup finalists (two losers, 2017 Nashville, 2010 Philadelphia) not in the top half
Average SC winner rank: 5.875
Average SC loser rank: 9.875
Average finalist rank: 7.875


What jumps out:

- There appears to be a higher correlation between all 3 of these advanced stats than points, despite the fact that almost every single playoff team is in the top half of points. It also appears pretty clear that when it comes to Stanley Cup winners and finalists, these advanced stats matter.

-Every Cup Finalist has finished top-5 in at least one of these stats except for New Jersey, Philadelphia, Boston (2012), and Pittsburgh (2017). All of these teams except for Philadelphia finished top-10 in at least one of these statistics.

-xGF% has blemishes when it comes to a very good 2011 Canucks team and a very good 2015 Blackhawks team being ranked in the bottom half of the league. I've heard people talk it up as the best of these stats but maybe not. Those teams were clearly very very good, and the other advanced stats back them as being pretty good as well.

-Not a single team was in the bottom 3rd of the league for any one of these stats except for Vancouver, who was exactly 20th in xGF%.

Tomorrow I will post data for every single playoff series in the same time frame. I'll probably post it in here unless people really like this research and consider it different enough/its own topic to the degree that I should make a different thread

 

[WJCJ]

"League's Corsi leading" doesn't cancel the absence due to injury of 4 of the top 6 D & one's #2C.

I was going to say the same thing, The only thing last year's playoffs will tell us if the Bruins are the model is that you can't win with that many injuries to key players.

 

[supsens]

In all honesty, I've used last year's Pittsburgh team in the past as an example of how tanking dominated but they are probably not the best model for sustainable success. The 2016 Pittsburgh team is one of the best teams ever and their Corsi reflects that.

That 2017 team, on the other hand, was dominated in 3 of the 4 series that they played and in the one series that they were mostly the better team, they still went to game 7 OT. They had elite Conn Smythe goaltending from a guy with an inconsistent, spotty playoff history when their starter went down with an injury. Then, when that backup started to struggle, the starter got healthy and played at an elite, Conn Smythe level. They also have 3 of the very best finishers in the world who manage to score a lot of goals and assists and maintain a high on-ice shooting percentage. Most teams don't have even one of those players and I can't think of very many that have two forwards with the offensive talents of Kessel, Crosby, and Malkin. The Penguins have 3.

On top of that, even superstar goalies usually don't put together the playoffs that Fleury+Murray did. Playoff performances from goalies can be pretty random and difficult to predict and a guy with a .908 career playoff SV% suddenly putting up a .924% isn't something you can consistently count on from anybody. Then Murray came back and put up a .937%, when his career playoff SV% is .928%.

Then you look at the teams they played against - Pekka Rinne was having a magnificent playoff run and then completely fell apart in the finals. Braden Holtby had the highest playoff SV% of all time and he was very bad against Pittsburgh. Bobrovsky won the Vezina then sucked against them. Yes, part of this comes from the Penguins having good shooters, absolutely, but they struggled to score on Craig Anderson. Given that, you have to admit that at least part of their domination of elite goaltenders could be credited to luck. If they really were so good that they made elite goalies look like AHL scrubs, they should have made a good goalie like Anderson look like an ECHL scrub.

Their #1D was Brian Dumolin. They won the Cup without their #1 defenseman and got dominated in 3 of the 4 series. They had a #1D playing way way over his head and were lucky to head into the finals and play against a team whose #1C was playing way over his head because their actual #1C was injured. They played Columbus who doesn't have an elite #1C., Ottawa who doesn't have a #1C, Nashville whose #1C is a decent borderline guy but was injured, and Washington whose #1C is a legit guy in the regular season but is a huge playoff choker who is so underrated as a regular season player that he doesn't really get flack for his playoff performances. How many teams play 4 rounds and don't play a single round against a team with a #1C? The Penguins did and they were able to win because their #2C is a top-5 center in the league, and they were extremely lucky to get those centers.

Teams are not going to be able to model themselves after Pittsburgh and disregard CF%, or having a single true top-pairing defenseman because they will not have the luxury of having two elite top-5 players who are both centers and they will not have the luxury of having an elite top-10 winger and a Conn Smythe goaltending performance from your backup and another Conn Smythe goaltending performance from your starter just as the backup was falling apart when the starter recovers from injury. Nobody can prepare for that, it takes a ton of luck. The closest thing would be if Edmonton won the Matthews lottery. That would have given them two superstar elite #1Cs and maybe they could have won last year when Talbot and Draisaitl were also playing at Conn Smythe level, but even Edmonton was already so lucky to get McDavid.

Pittsburgh's 3 MVPs were guys drafted 1st, 2nd, and 1st overall in 3 back to back seasons. Their #1D was a guy who was a part of a package that they traded their #2 overall pick (the year after those 3 back to back seasons) for. A lot of teams can't be a bottom-2 team for 4 straight years because it's not financially sustainable and could lead to relocation. Yeah

You should look at Craig Anderson’s playoff save%
The guy seriously ups his game in the playoffs and is hard to score on.

 

[TomasHertlsRooster]

You should look at Craig Anderson’s playoff save%
The guy seriously ups his game in the playoffs and is hard to score on.

I'm a Sharks fan, I'm fully aware. I watched the Dan Boyle own goal game when I was 13 years old and that series taught me a lot of what I know about hockey. We had like 50 shots in regulation and 0 goals on Anderson lol

Anderson is actually a very good goalie with elite playoff numbers but that doesn't really disprove my point. Why did Pittsburgh completely thrash Holtby (The reigning Vezina winner at the time,whose regular season and playoff numbers are superior to Anderson's), Bobrovsky (that year's Vezina winner who had won before), and Rinne (the Conn Smythe front runner before the series started, this year's Vezina front runner, and a former Vezina finalist), and then suddenly struggle to score on Anderson?

 

[ps241]

I am the oldest guy posting in this thread. I lived in the era when broadcast teams thought hits, blocked shots, and faceoff % were all important enough to talk about.

Corsi is a breath of fresh air.

 

[SladeWilson23]

It's not about just getting pucks to the net. It's about getting quality chances now. High danger shots are the new trend.

Every team in the league typically generates and allows roughly the same number of chances from the high danger areas of the ice. That number is between 10 and 13 high danger chances per 60 minutes.

What separates the best offensive / defensive teams from the worst offsensive / defensive teams is the number of low danger chances.

 

[SladeWilson23]

The opposite is true if you look at the recent history of the Stanley Cup.

Lundqvist, Luongo, Bobrovsky, and Price put together - 0 Cups

Top 5 corsi teams - 7 Cups since the lockout and 3 in the last 5 years.

The top 4 teams in Corsi last year missed the playoffs as well. Also 7 times in 12 years is not exactly that convincing.

Like I said I agree 100% with you philosophically. I just don't agree with focusing on the Corsi standings because they're way too tight, and there's too much fluctuation in them throughout the season.

 

[lomiller1]

Well, xGF% is a stat with higher correlation to future goals than CF%.

The version on xGF% that outperformed Corsi is no longer public domain. The guy who created it had to go dark when he was hired by the Avs.

 

[lomiller1]

Again, "shooting is good" shouldn't be a crazy idea.

One of the weirdest elements of the discussion of Corsi in the mass media is that the most outspoken people against it also tend to lecture you continually about the importance of getting pucks to the net, and say things like "there is no such thing as a bad shot".

 

[lomiller1]

Braden Holtby had the highest playoff SV% of all time and he was very bad against Pittsburgh. Bobrovsky won the Vezina then sucked against them.

Side note, but Holtby has/had been bad since Feb 2017. He was among the worst starters for any playoff team from Feb1 onwards, played at the same low level in the playoffs and that has more or less continued though this year. I'd be interested to hear of there was some injury or concussion involved.

 

[lomiller1]

What separates the best offensive / defensive teams from the worst offsensive / defensive teams is the number of low danger chances.

True, though I suspect not quite in the way you intend. The teams that score the most almost invariable take the most low danger shots. They also take the most high danger shots because those low danger shots set up tips, deflections, rebounds and general havoc around the net.

 

[madmike77]

The Flames are one of the top Corsi teams this year. Tells you about all you need to know about the stat.

Their high danger chances are solid too. Advanced-stats-wise the Flames should easily be a playoff team. Fact is they’re not even close. I’ve pretty much jumped off the advanced stats bandwagon at this point.

 

[SladeWilson23]

True, though I suspect not quite in the way you intend. The teams that score the most almost invariable take the most low danger shots. They also take the most high danger shots because those low danger shots set up tips, deflections, rebounds and general havoc around the net.

Exactly. What typically leads to more HD chances is taking a lot of low danger shots. Tips, deflections, and rebounds are in while pretty passing plays are out.

 

[Spawn]

Goodhart's Law

"Goodhart's law is an adage named after economist Charles Goodhart, which has been phrased by Marilyn Strathern as: "When a measure becomes a target, it ceases to be a good measure."[1] One way in which this can occur is individuals trying to anticipate the effect of a policy and then taking actions which alter its outcome."

Seems like teams started 'gaming' corsi when it's 'value' was revealved.

I think you're on bang on with the bold. It's what Dallas Eakins tried to do with the Oilers. He saw that good teams tended to have good corsi so he assumed that good corsi would equal being a good team. So he coached for corsi. Seems like the Flames coach has done the same the last few years.

Seems like a lazy way to coach. Turns out just throwing the puck on net the minute you get inside the blueline isn't the be all and end all of hockey.

Not that these stats don't have value. They just don't have as much value as initially thought, and they have even less value when it becomes the goal.

 

[DK59]

I am just wondering if this may just be a bizarre outlier year because for both corsi and Xgoals there are so many teams that have positive numbers (for both of these stats)that will miss the playoffs. This year it looks very possible that half the teams that currently have positive corsi numbers (some dramatically so) will miss the playoffs. The same is the case for Xgoals which surprises me because I would have expected differences in shot quality to explain more about what are seeing. In past years there have always been anomalies which is to be expected but nothing like we are seeing this year.

 

[Machinehead]

The top 4 teams in Corsi last year missed the playoffs as well. Also 7 times in 12 years is not exactly that convincing.

Like I said I agree 100% with you philosophically. I just don't agree with focusing on the Corsi standings because they're way too tight, and there's too much fluctuation in them throughout the season.

Fair enough, I just think saying that goaltending beats it is patently false.

There's a long history of good corsi teams winning in the playoffs and a long history of teams who lean on their goaltender hitting a wall in the playoffs.

 

[Pavels Dog]

Hockey is way too complicated to be boiled down to any one stat.

Corsi is like Sex Panther by Odeon, 60% of the time, it works every time