Tuesday, April 16, 2013
When Bad Puck Possession Teams Turn Good
A week after going 24:33 minutes without a shot in a loss to the New York Rangers, the Leafs were back to their low-chance ways against New Jersey last night, managing only 13 shots.
Amazingly, the Leafs won. But what's truly amazing is that this wasn't abnormal for the team. The Leafs have been consistently outshot and outchanced all season, yet they continue to find ways to make their shots count and pick up wins.
The performance has come as a shock to many, but none more than those who ascribe to advanced statistics.
But despite its name, advanced stats in hockey aren't all that advanced.
The most important advanced stat in hockey is puck possession. And although no actual measure of how long teams possess the puck exists (at least publicly), measuring how many shots a team takes/allows is almost as good.
The logic behind the importance of puck possession is simple: Teams that posses the puck more shoot the puck more than their opponents. And if you outshoot your opponents you are more likely to out-chance your opponents; and if you out-chance your opponents you are more likely to outscore your opponents. That's it. Advanced stats in hockey are really that simple.
Sure, in any one game a goalie can stand on his head or some fortuitous bounces can lead to a goal-fest, but over time, like, say, an 82-game regular season, teams that routinely outshoot their opponents will find themselves with better records.
Ask yourself this: Do you feel more confident watching your favourite team when they are carrying the play and outshooting the other team, or are you more confident when your team is consistently hemmed in their own zone and outshot? The answer is clearly the former and you don't have to be a statistician to understand. There's a reason teams are generally described as "lucky to steal two points" when they win after being badly outshot.
The best measure of a team's puck possession is something called Fenwick close, a metric that measures shot attempts at even strength, minus blocked shots, in close situations (when the score is tied or within one goal through the first two periods and tied during the third period). Fenwick close only looks at these close situations because of score effects—which happen when the leading team goes into a defensive shell to protect their lead and end up allowing the trailing team to outshoot them.
Since first being collected in 2007-08, Fenwick close has a 0.61 correlation with the amount of points a team accumulates. That's a pretty strong correlation and shows that teams that own the shot clock generally own the standings. Since 2007-08, almost 75% of the top 16 teams in Fenwick close make the playoffs. This reality is wonderfully illustrated below, courtesy of Eyes on the Prize, the Montreal Canadiens SB Nation blog. An explanation from the author, Chris Boyle, below:
As
you can see, non-playoff teams generally have a Fenwick close of less
than 50%—in other words, they are giving up the majority of shots—and
playoff teams generally have a Fenwick close of more than 50%—they are
taking the majority of shots.
Having a strong Fenwick close mark is also predictive of how far a team will advance in the playoffs. A team's regular season Fenwick close score has a 0.47 correlation with how far a team goes in the playoffs (to calculate this I assigned teams a score of 1-5 based on which round of the playoffs they made, 5 indicating the team won the Stanley Cup). Although still moderately strong, puck possession isn't correlated as highly with playoff performance as it is regular season performance because a seven-game series is a much smaller sample than 82 games and anything can happen in a shortened series. Put another way, a good team isn't likely to play poorly over 82 games, but even good teams have slumps for a handful of games. And if you slump in the playoffs, or run into a hot goalie, there isn't a next week to make up for it. Plus, the playoffs is composed of mostly good teams and the difference between a four seed and a five seed, for instance, isn't usually very large.
Although good puck possession teams generally do better in the playoffs than bad puck possession teams, they don't always make the playoffs. A total of 16 teams with a Fenwick close score of over 50% have missed out on the playoffs since 2007-08.
Conversely, being a bad puck possession team doesn't necessarily disqualify you from the post-season. A total of 22 teams with a Fenwick close score below 50% have made the playoffs, almost a third of which have made it to the conference finals and beyond. Most famously, the Pittsburgh Penguins won the 2009 Stanley Cup as a sub-50% Fenwick close team, although they ended the season at 49.9% and were almost a 55% team from the day Dan Bylsma took over in February.
Cruelly, the Toronto Maple Leafs have twice iced teams that managed to break the hallowed 50% Fenwick close mark. In both years, 2007-08 and 2009-10, they missed the playoffs, the latter season being so bad that they forfeited the second overall pick, Tyler Seguin, to the Boston Bruins as part of the Phil Kessel trade. Yes, one of the worst seasons in Toronto Maple Leafs history occurred when they were actually a good puck possession team.
Now the Leafs are on the brink of making the playoffs, despite owning one of the league's worst Fenwick close marks (44.66%). The only teams worse are Edmonton and Buffalo, two teams essentially eliminated from post-season play.
So, even though puck possession is highly correlated with winning, it clearly isn't everything in hockey.
If we examine these outlier teams, those who buck the trend and succeed in spite of their crummy puck possession (the 2010 Montreal Canadiens going to the Eastern Conference Final) or fail despite their good puck possession (the 2010 Toronto Maple Leafs bequeathing Tyler Seguin), we may uncover some more information about what it takes to win, and maybe just as important, what it takes to lose.
Actually, it's not hard to uncover the hidden variable: it's goaltending.
Good puck possession teams that miss the playoffs almost always have a below-average goalie. Or, if you're the Toronto Maple Leafs, a miserable goalie sent from the future to torpedo your season (looking at you Vesa Toskala/Jonas Gustavsson). It doesn't matter how badly you outshoot the other team if you have a porous goalie waiting behind you to help the opposition convert their chances. Only three of the 16 above-50% puck possession teams that missed the playoffs had above-average goalies.
Conversely, bad puck possession teams that make the playoffs almost always have an above-average goalie, which helps explain why a team like Nashville could have such an abysmal Fenwick mark (46.08%) last season yet still succeed. It doesn't really matter if you're consistently shelled when you have a goalie like Pekka Rinne waiting to bail you out. In fact, only three of the 22 sub-50% puck possession teams that made the playoffs had below-average goalies.
The importance of good goaltending shouldn't surprise many, although it becomes particularly important for teams that can't limit the shots against. That's encouraging for the Toronto Maple Leafs, a terrible puck possession team, but one lucky enough to have God's favourite son, James Reimer, posting a .925 save percentage (considerably better than the league average rate of .912).
The Leafs better hope Reimer continues his stonewalling ways because if their puck possession mark remains under 45% they will be the worst puck possession team to ever make the post-season. The next worst team, the 2010-11 Anaheim Ducks (45.66%), were eliminated in the first round by the Nashville Predators.
Even if the Leafs are able to boost their Fenwick over the last few games their invite to Lord Stanley's dance will still be rare. Only six teams have made the playoffs with a Fenwick close score of less than 47%. Three of those teams were ousted in the first round, but the other three had considerably more success. Last year's Predators made the second round; the 2009-10 Montreal Canadiens went on a magical run to the Eastern Conference Final; and the 2007-08 Pittsburgh Penguins made the Stanley Cup Final.
Making the playoffs in Toronto—a city of sports misery for years—was already going to be a minor miracle for the Leafs. Doing it the way they are might qualify the 2013 Toronto Maple Leafs for sainthood.
Amazingly, the Leafs won. But what's truly amazing is that this wasn't abnormal for the team. The Leafs have been consistently outshot and outchanced all season, yet they continue to find ways to make their shots count and pick up wins.
The performance has come as a shock to many, but none more than those who ascribe to advanced statistics.
But despite its name, advanced stats in hockey aren't all that advanced.
The most important advanced stat in hockey is puck possession. And although no actual measure of how long teams possess the puck exists (at least publicly), measuring how many shots a team takes/allows is almost as good.
The logic behind the importance of puck possession is simple: Teams that posses the puck more shoot the puck more than their opponents. And if you outshoot your opponents you are more likely to out-chance your opponents; and if you out-chance your opponents you are more likely to outscore your opponents. That's it. Advanced stats in hockey are really that simple.
Sure, in any one game a goalie can stand on his head or some fortuitous bounces can lead to a goal-fest, but over time, like, say, an 82-game regular season, teams that routinely outshoot their opponents will find themselves with better records.
Ask yourself this: Do you feel more confident watching your favourite team when they are carrying the play and outshooting the other team, or are you more confident when your team is consistently hemmed in their own zone and outshot? The answer is clearly the former and you don't have to be a statistician to understand. There's a reason teams are generally described as "lucky to steal two points" when they win after being badly outshot.
The best measure of a team's puck possession is something called Fenwick close, a metric that measures shot attempts at even strength, minus blocked shots, in close situations (when the score is tied or within one goal through the first two periods and tied during the third period). Fenwick close only looks at these close situations because of score effects—which happen when the leading team goes into a defensive shell to protect their lead and end up allowing the trailing team to outshoot them.
Since first being collected in 2007-08, Fenwick close has a 0.61 correlation with the amount of points a team accumulates. That's a pretty strong correlation and shows that teams that own the shot clock generally own the standings. Since 2007-08, almost 75% of the top 16 teams in Fenwick close make the playoffs. This reality is wonderfully illustrated below, courtesy of Eyes on the Prize, the Montreal Canadiens SB Nation blog. An explanation from the author, Chris Boyle, below:
The rings of the graph represent each round of the Stanley Cup playoffs. The further the logo is away from the ring represents the distance from a playoff berth. I have also charter the percentage from .400 to .600. The further away from the .400 represents a stronger possession team. The ultimate on this index would be the 2008 Red Wings with a score of 59.39 located on the top portion of the Stanley Cup in the +.550 section. The 2008 Thrashers scrape the bottom of this index with a 41.23 and 28th position during the same season.
Click for a hi-resolution photo. |
Having a strong Fenwick close mark is also predictive of how far a team will advance in the playoffs. A team's regular season Fenwick close score has a 0.47 correlation with how far a team goes in the playoffs (to calculate this I assigned teams a score of 1-5 based on which round of the playoffs they made, 5 indicating the team won the Stanley Cup). Although still moderately strong, puck possession isn't correlated as highly with playoff performance as it is regular season performance because a seven-game series is a much smaller sample than 82 games and anything can happen in a shortened series. Put another way, a good team isn't likely to play poorly over 82 games, but even good teams have slumps for a handful of games. And if you slump in the playoffs, or run into a hot goalie, there isn't a next week to make up for it. Plus, the playoffs is composed of mostly good teams and the difference between a four seed and a five seed, for instance, isn't usually very large.
Although good puck possession teams generally do better in the playoffs than bad puck possession teams, they don't always make the playoffs. A total of 16 teams with a Fenwick close score of over 50% have missed out on the playoffs since 2007-08.
Conversely, being a bad puck possession team doesn't necessarily disqualify you from the post-season. A total of 22 teams with a Fenwick close score below 50% have made the playoffs, almost a third of which have made it to the conference finals and beyond. Most famously, the Pittsburgh Penguins won the 2009 Stanley Cup as a sub-50% Fenwick close team, although they ended the season at 49.9% and were almost a 55% team from the day Dan Bylsma took over in February.
Cruelly, the Toronto Maple Leafs have twice iced teams that managed to break the hallowed 50% Fenwick close mark. In both years, 2007-08 and 2009-10, they missed the playoffs, the latter season being so bad that they forfeited the second overall pick, Tyler Seguin, to the Boston Bruins as part of the Phil Kessel trade. Yes, one of the worst seasons in Toronto Maple Leafs history occurred when they were actually a good puck possession team.
Now the Leafs are on the brink of making the playoffs, despite owning one of the league's worst Fenwick close marks (44.66%). The only teams worse are Edmonton and Buffalo, two teams essentially eliminated from post-season play.
So, even though puck possession is highly correlated with winning, it clearly isn't everything in hockey.
If we examine these outlier teams, those who buck the trend and succeed in spite of their crummy puck possession (the 2010 Montreal Canadiens going to the Eastern Conference Final) or fail despite their good puck possession (the 2010 Toronto Maple Leafs bequeathing Tyler Seguin), we may uncover some more information about what it takes to win, and maybe just as important, what it takes to lose.
Actually, it's not hard to uncover the hidden variable: it's goaltending.
Good puck possession teams that missed the playoffs | |||
Year | Team | Fenwick Close | Team SV% (league avg) |
2011-12 | Winnipeg Jets | 51.05% | .905 (.914) |
2010-11 | New Jersey Devils | 52.16% | .906 (.913) |
2010-11 | Calgary Flames | 52.06% | .906 (.913) |
2010-11 | Columbus Blue Jackets | 51.92% | .900 (.913) |
2010-11 | St. Louis Blues | 51.92% | .902 (.913) |
2009-10 | Calgary Flames | 51.64% | .916 (.911) |
2009-10 | Toronto Maple Leafs | 51.55% | .896 (.911) |
2009-10 | St. Louis Blues | 51.40% | .916 (.911) |
2008-09 | LA Kings | 50.92% | .906 (.908) |
2008-09 | Dallas Stars | 50.61% | .895 (.908) |
2007-08 | Columbus Blue Jackets | 52.47% | .910 (.909) |
2007-08 | Carolina Hurricanes | 52.31% | .898 (.909) |
2007-08 | Toronto Maple Leafs | 51.33% | .896 (.909) |
2007-08 | Buffalo Sabres | 50.86% | .903 (.909) |
2007-08 | Tampa Bay Lightning | 50.73% | .890 (.909) |
2007-08 | Chicago Blackhawks | 50.18% | .905 (.909) |
Good puck possession teams that miss the playoffs almost always have a below-average goalie. Or, if you're the Toronto Maple Leafs, a miserable goalie sent from the future to torpedo your season (looking at you Vesa Toskala/Jonas Gustavsson). It doesn't matter how badly you outshoot the other team if you have a porous goalie waiting behind you to help the opposition convert their chances. Only three of the 16 above-50% puck possession teams that missed the playoffs had above-average goalies.
Bad puck possession teams that made the playoffs | |||
Year | Team | Fenwick Close | Team SV% (league avg) |
2011-12 | New York Rangers | 49.90% | .924 (.914) |
2011-12 | Washington Capitals | 49.71% | .912 (.914) |
2011-12 | Phoenix Coyotes | 49.17% | .927 (.914) |
2011-12 | Nashville Predators | 46.08% | .921 (.914) |
2010-11 | New York Rangers | 49.74% | .922 (.913) |
2010-11 | Anaheim Ducks | 45.66% | .915 (.913) |
2009-10 | Buffalo Sabres | 49.58% | .924 (.911) |
2009-10 | Vancouver Canucks | 49.30% | .913 (.911) |
2009-10 | Colorado Avalanche | 46.85% | .917 (.911) |
2009-10 | Montreal Canadiens | 46.07% | .919 (.911) |
2008-09 | Pittsburgh Penguins | 49.97% | .909 (.908) |
2008-09 | St. Louis Blues | 49.86% | .907 (.908) |
2008-09 | Vancouver Canucks | 49.28% | .914 (.908) |
2008-09 | Philadelphia Flyers | 49.06% | .914 (.908) |
2008-09 | Montreal Canadiens | 46.91% | .909 (.908) |
2007-08 | Calgary Flames | 49.95% | .906 (.909) |
2007-08 | Anaheim Ducks | 49.69% | .922 (.909) |
2007-08 | Dallas Stars | 49.35% | .908 (.909) |
2007-08 | Minnesota Wild | 48.28% | .916 (.909) |
2007-08 | Philadelphia Flyers | 47.12% | .915 (.909) |
2007-08 | Montreal Canadiens | 47.11% | .919 (.909) |
2007-08 | Pittsburgh Penguins | 46.51% | .918 (.909) |
Conversely, bad puck possession teams that make the playoffs almost always have an above-average goalie, which helps explain why a team like Nashville could have such an abysmal Fenwick mark (46.08%) last season yet still succeed. It doesn't really matter if you're consistently shelled when you have a goalie like Pekka Rinne waiting to bail you out. In fact, only three of the 22 sub-50% puck possession teams that made the playoffs had below-average goalies.
The importance of good goaltending shouldn't surprise many, although it becomes particularly important for teams that can't limit the shots against. That's encouraging for the Toronto Maple Leafs, a terrible puck possession team, but one lucky enough to have God's favourite son, James Reimer, posting a .925 save percentage (considerably better than the league average rate of .912).
The Leafs better hope Reimer continues his stonewalling ways because if their puck possession mark remains under 45% they will be the worst puck possession team to ever make the post-season. The next worst team, the 2010-11 Anaheim Ducks (45.66%), were eliminated in the first round by the Nashville Predators.
Even if the Leafs are able to boost their Fenwick over the last few games their invite to Lord Stanley's dance will still be rare. Only six teams have made the playoffs with a Fenwick close score of less than 47%. Three of those teams were ousted in the first round, but the other three had considerably more success. Last year's Predators made the second round; the 2009-10 Montreal Canadiens went on a magical run to the Eastern Conference Final; and the 2007-08 Pittsburgh Penguins made the Stanley Cup Final.
Making the playoffs in Toronto—a city of sports misery for years—was already going to be a minor miracle for the Leafs. Doing it the way they are might qualify the 2013 Toronto Maple Leafs for sainthood.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment