Teams win more often when outshot?

[Alexei Yashvalev]

I was just looking at team stats on nhl.com.

The overall winning percentage for teams when outshot is higher than when they outshoot their opponent. .512% vs .494%.

I just thought that was kind of weird. Clearly when you look at a box score the team with less shots is actually the one with the advantage.

 

[How Ya Drouin]

Keep calm corsi's wrong.

 

[Era of Sanity]

In the the third period or late in the game the team protecting a lead has already scored their goals and often sits back a bit playing a defensive game and generates few shots while the trailing team is desperate and tries to play offense and throw pucks at the net.

 

[SnowblindNYR]

Keep calm corsi's wrong.

I believe Corsi only looks at the 3rd period if the score is tied. Many times teams pad their stats in the 3rd while playing from behind.

 

[Alexei Yashvalev]

In the the third period or late in the game the team protecting a lead has already scored their goals and often sits back a bit playing a defensive game and generates few shots while the trailing team is desperate and tries to play offense and throw pucks at the net.

I thought about that and also wondered if it has anything to do with goalies tending to play better the more shots they face.

 

[snizzbone*]

I believe Corsi only looks at the 3rd period if the score is tied. Many times teams pad their stats in the 3rd while playing from behind.

what? no..

Corsi = shots + missed shots + blocked shots aka shots directed at net

 

[Bear of Bad News]

I thought about that and also wondered if it has anything to do with goalies tending to play better the more shots they face.

There's no direct evidence that this happens (or the opposite, for that matter).

It's typically that the leading teams go into a defensive shell (allowing shots from the outside, and not taking risks themselves) in the third period.

 

[Rysto]

Yeah, it's called the score effect. Teams that are losing tend to shoot way more often. Some advanced stats only consider shots taken when the score is tied or a team is leading by only a single goal because of this.

 

[ClasslessGuy]

Team with Halak

 

[sbet1998]

Team with Halak

What? Halak had been playing for the Blues the last few years and they dont allow many shots. The only thing that changed was his GAA from his stint with the Habs.

 

[sbet1998]

I think its more the lack of high end possession teams. Mediocre teams make up the majority of the NHL and thus play each other a lot leading to a lot of close games where which ever team finds themselves ahead goes into a defensive mode and allows a lot of low % shots in the 3rd periods. Good teams also do this, but they usually still find a way to maintain enough possession to hold onto the shot lead in these situations. This is my theory, anyways.

 

[Hynh]

First of all, that stat would count games when the shots are 29-28, would it not?

Second of all, teams play different when they have the lead. For instance, when the Blackhawks trail by 2 or more they take 68.9% of the shots. But when they are leading by 2 or more they take 52.1% of the shots.

what? no..

Corsi = shots + missed shots + blocked shots aka shots directed at net

He's referring to score close corsi, which is generally held to be more predictive than just straight corsi.

 

[Avs]

I know one thing that happens with the Avs is that they play with the lead so often That the other starts throwing more, lower percentage shots on net

 

[sharks9]

When your team is leading, you usually sit back and the other team presses hard.

 

[Mayor Bee]

Yeah, it's called the score effect. Teams that are losing tend to shoot way more often. Some advanced stats only consider shots taken when the score is tied or a team is leading by only a single goal because of this.

Just like in football, where passing yards don't usually correlate to winning. A team with a sizable deficit will start passing the ball on every down, while the team with the big lead will drop into soft coverage to prevent the big play. The leading team is trading yardage for keeping the clock winding, but it's not uncommon to see the team with more passing yards end up losing big.

Generally, a QB with a ratio of over 150 passing yards per passing TD was in this situation, with the winning percentage getting worse the higher the ratio goes.

 

[MiniMoose]

That's due to the Leafs.

 

[Milan90]

It's pretty obviously because of how teams play while carrying a lead, but if this is such a big flaw with shot numbers then why is Corsi taken as gospel?

 

[Hynh]

It's pretty obviously because of how teams play while carrying a lead, but if this is such a big flaw with shot numbers then why is Corsi taken as gospel?

Score close 5x5 corsi (or fenwick) is the most commonly possession stat cited at a team level and attempts to avoid this by only looking at the events when the score is within 1 in the first two periods or tied in the third period and only including 5x5 play.

 

[Milan90]

Score close 5x5 corsi (or fenwick) is the most commonly possession stat cited at a team level and attempts to avoid this by only looking at the events when the score is within 1 in the first two periods or tied in the third period and only including 5x5 play.

But then would this not miss a good deal of games since there are a high number of games where the margin is greater than two goals within the first two periods?

 

[Hynh]

But then would this not miss a good deal of games since there are a high number of games where the margin is greater than two goals within the first two periods?

Yeah but in those situations, does it matter? When Calgary scored 4 in the first period against the Ducks the other night, was the rest of the game a true representation of the teams' abilities?

 

[Milan90]

You can't just make the assumption that the only time a team plays like themselves is when the game is within one.

 

[ukyo]

You can't just make the assumption that the only time a team plays like themselves is when the game is within one.

The thought is that using those times give you a better statistical correlation to winning.

 

[Arrch]

You can't just make the assumption that the only time a team plays like themselves is when the game is within one.

You're not, you're assuming that when the game isn't within 1, teams tend to play differently.

 

[Sonofamitch]

It's working for the 3rd place team in the east

 

[Arrch]

It's working for the 3rd place team in the east

Yeah...