Possession Stats

[Alan Jackson]

I just had a thought, and thought I might run it by the people who have a better grasp of "advanced stats" than I do.

If stats like Corsi and Fenwick, in very basic terms, are used in an attempt to measure possession, based on the idea that teams that shoot the puck more will generally win more often, would timing actual puck possession be more or less effective?

For example, would measuring the amount of time a player/team spends, with the puck, in the offensive zone as opposed to the opponent be a useful number?

Why or why not?

Click to read more...

 

[soireeculturelle]

it's another way to do it. a lot more labor-intensive since you'd need to either manually time things, which is open to error, or have a system which does it for you.

the other thing, is that a shot attempt has a probability of going into the net, whereas just having the puck (against your skates in the corners, let's say), doesn't. the way I understand it, is that a shot attempt means that you have meaningful control of the puck, and are trying to make a play to get it to the net for a goal. if all you do is chip the puck into the corner of the rink and get pushed up and down the sideboard for 40 seconds with zero shots attempted, it's not really conducive to scoring goals and doesn't really show that you're dominating the opposing defense.

 

[ot92s]

I just had a thought, and thought I might run it by the people who have a better grasp of "advanced stats" than I do.

If stats like Corsi and Fenwick, in very basic terms, are used in an attempt to measure possession, based on the idea that teams that shoot the puck more will generally win more often, would timing actual puck possession be more or less effective?

For example, would measuring the amount of time a player/team spends, with the puck, in the offensive zone as opposed to the opponent be a useful number?

Why or why not?

I mentioned that here before. It would be fantastic. light years ahead of any "advanced" stats when it comes to history reports.

 

[RaskY]

It'd be really difficult and laboring to collect those stats. A player usually keeps a puck on his stick for no more than 1 to 2 seconds it seems like. Not sure how you'd do it. Maybe if you had a panel of people with stop watches tracking each player's time with the puck.

 

[TheDevilMadeMe]

I would imagine that this is one of the stats that teams like the Blackhawks privately collect.

 

[RaskY]

So I decided to track Anze Kopitar's puck possession time tonight vs. the Ducks. So, here are the results.


Total Puck Possession Time: 1:29
Total Time on Ice: 20:09
Total Shifts: 29

PuckPoss.% of Total TOI: 0.06%

Avg. Shift Length: 0:41
Avg PuckPoss. Time Per Shift: 0:03


Here it is broken down period by period

1st Per. PuckPoss. Time: 0:16
1st Per. TOI: 5:39
PuckPoss.% of 1st Per. TOI: 0.04%
Shifts: 9
Avg PuckPoss. Time Per Shift: 0:02

2nd Per. PuckPoss. Time: 0:36
2nd Per. TOI: 7:03
PuckPoss.% of 2nd Per. TOI: 0.08%
Shifts: 10
Avg PuckPoss. Time Per Shift: 0:04

3rd Per. PuckPoss. Time: 0:37
3rd Per. TOI: 7:27
PuckPoss.% of 2nd Per. TOI: 0.08%
Shifts: 10
Avg PuckPoss. Time Per Shift: 0:04

The digits were rounded.

Overall, not sure if 1 minute and 29 seconds of puck possession time out of 20 minutes of ice time is good or bad since I have nothing to compare it to, but by watching him it seemed like Kopitar - along with the entire top line of LA had a pretty quiet game. He, along with Gaborik and Brown seemed to become more involved as the game went along as the stats indicate. Very quiet in the 1st period I thought.

Some notes:

* He drew a penalty at 8:54 of the 2nd period after completing his longest duration of play with the puck in the game. It was around 6 seconds with the puck as he worked along the boards before being hooked by Nick Bonino.

*The problem with this stat is if somebody else was tracking Kopitar's puck possession time tonight, I'm almost certain he would have different numbers than me, although I don't think they would be drastic differences. I thought I was able to start and stop the clock at the right moments throughout the game.

*This was overall pretty boring but at the same time mildly interesting.

 

[Ohashi_Jouzu*]

It'd be really difficult and laboring to collect those stats. A player usually keeps a puck on his stick for no more than 1 to 2 seconds it seems like. Not sure how you'd do it. Maybe if you had a panel of people with stop watches tracking each player's time with the puck.

Since players work as a unit on the ice, it might make more sense to track the time the unit possesses the puck. Or, alternatively, simply measure the length of time between puck retrieval and loss/relinquishing of possession. The aggregate would be like the field position stat like in soccer, and the average length could give you a measure of a team's puck control. Don't need the player level retrievals/losses, imo, although more is almost always better in cases like this.

 

[KaseMeOutside]

TSN has been tracking offensive zone time in this ducks/kings series, if the NHL adopts the sportsvu camera systems, then perhaps we could finally have a universal tracking system. to me the amount of total time spent in another teams zone is as important if not more so than shot attempts....the type of shots should matter

 

[mmbt]

TSN has been tracking offensive zone time in this ducks/kings series, if the NHL adopts the sportsvu camera systems, then perhaps we could finally have a universal tracking system. to me the amount of total time spent in another teams zone is as important if not more so than shot attempts....the type of shots should matter

I agree, because even true possession itself means little as far as scoring if you're mostly keeping possession in the defensive and neutral zones. Whereas even possession in your own zone still means you could easily give up a good scoring chance with a turnover; it's still in the "danger zone" until you clear it. To me, zone time would be more meaningful than shot totals.

OTOH, do any of these advanced stats weigh shot totals based on what percentage were taken by forwards vs defensemen? Because on average a forward shot has a lot higher chance of scoring.

 

[Ho Borvat]

I just had a thought, and thought I might run it by the people who have a better grasp of "advanced stats" than I do.

If stats like Corsi and Fenwick, in very basic terms, are used in an attempt to measure possession, based on the idea that teams that shoot the puck more will generally win more often, would timing actual puck possession be more or less effective?

For example, would measuring the amount of time a player/team spends, with the puck, in the offensive zone as opposed to the opponent be a useful number?

Why or why not?

I remember reading somewhere that Corsi % works out nearly identical to time on attack? I believe there has been studies where they did a "time on attack" stat and they found it pretty much mirrored corsi (much easier to track). I wish I had the source, but basically if you have a 60% corsi you would possess the puck for ~36/60 minutes of the game.

 

[MVP of West Hollywd]

Expect it to be tracked by the league within 5 years with tech similar to "SportVU" used in basketball right now. But right now it sounds like Corsi/Fenwick does a real good job of estimating possession

 

[Micklebot]

I remember reading somewhere that Corsi % works out nearly identical to time on attack? I believe there has been studies where they did a "time on attack" stat and they found it pretty much mirrored corsi (much easier to track). I wish I had the source, but basically if you have a 60% corsi you would possess the puck for ~36/60 minutes of the game.

This one is only for a half season, and only one team at that, but yeah, it has Corsi and Time of possession pretty much mirroring each other.

http://www.pensionplanpuppets.com/2013/9/16/4727746/leafs-attack-time-at-the-halfway-mark

 

[67 others]

Can someone explain Corsi stats and how to read them to me?

 

[Marc the Habs Fan]

Can someone explain Corsi stats and how to read them to me?

It's all about shot attempts. On net+blocked+shots that missed the net. You can break it down per team (easy) or per player which requires a lot of work and going on NHL.com's game play-by-play feature.

CBC was showing the viewers plenty of graphics with puck possession time during the 1st two rounds, not sure what method they were using beyond a simple stop watch.

 

[PepsiCenterMagic]

Possession --> More shots --> Shot Quality

Perhaps the next step in the research is figuring out how to incorporate the third dimensional analysis of the quality of shots.

Example.

Team A: +3.4 s/g differential
Team B: +2.5 s/g differential

Team A: On average, takes a shot that goes in 9.8% of the time
Team B: On average, takes a shot that goes in 11.3% of the time

And then figure a way how to balance the two dimensions to create a constant.

.......I might work on this.

 

[PoutineSp00nZ]

All you need to know about possession stats is that they are garbage and mean very little.