A 2nd Look at the Detroit Pre-Season Game
For years the Buffalo Sabres Jim Corsi tracked even strength shots attempted as a way to get a handle on whether they were outplaying the opposition or being outplayed. Attempted Shots For and Attempted Shots Against are an imperfect measure of scoring chances and time of possession -- but at a gut level they do in fact record where the puck is located. If your line or defense pairing is giving up far more shots than the rest of the team on a consistent basis--well that's a problem.
The Oiler and Flames bloggers have been tracking the "Corsi numbers" or attempted shots since the NHL started making this data available in the middle of last season. So let's take a look at how all four lines and three defense pairings fared last night. Attempted shots are calculated as follows = Shots on Goal + blocked shots + missed shots. The Corsi number is simply Attempted Shots For minus Attempted Shots Against. This only counts the SAF/SAA at Even Strength (excluding all Empty Net situations).
Corsi +/-
Goaltending/Team
-12 Lehtonen
In the Detroit pre-season game each club scored 1 goal at even strength (Detroit added 2 PP goals and a EN goal). A tie at even strength is good, but the Thrashers were outchanced at even strength by a margin of 12.
Defensemen
+5 Enstrom
+1 Havelid
When the top pairing was out the Thrashers outplayed Detroit.
-4 Hainsey
-2 Bogosian
The second pairing was mildly outplayed.
-12 Exelby
-12 Kulda
The third pairing was just hanging and was lucky to get off the ice without allowing a goal. Did I mention that having Exelby on the roster is holding this team back?
Forwards
-2 Kovalchuk
0 Christensen
-2 Williams
The top line was even or slightly worse. Even is not good enough. You need your top line to outplay the opposition to have a chance to win most nights.
-3 Sterling
-4 Little
-6 Armstrong
Detroit had a mild advantage against the 2nd line.
-8 Reasoner
-9 Stuart
-6 Thorburn
The third line was in their own end a ton. When I saw these numbers I assumed that the Reasoner line was matched up against Zetterberg's line and that is why they were so badly out chanced. Then I looked at the shift chart I was very surprised to see that Reasoner was almost never matched up against Detroit's top scoring line. Those are some ugly numbers considering they didn't even face the toughest line Detroit could put on the ice.
+2 Slater
+0 Stevenson
+2 Boulton
The forth line was the only positive forward line. They did exactly what you expect on the 4th line--play at the enemy's end of the ice.
Conclusion
The 4th line and the top defense pairing outplayed the opposition using Net Shots Attempted as a measure. If the readers enjoyed this I'll do some posts like this in the future.
EDIT/UPDATE: Here's some more cool information. Vic Ferrari has also made this cool program to show you which players where on the ice when a shift ends with a stoppage. (You want stoppage in the opposition end because that means you were applying pressure down there.) This is another plus/minus sort of stat, how many of your shifts ended in the opposition end (good) minus the number of your shifts that ended in your defensive end (bad).
Goaltending/Team
+0 Lehtonen
For a team that was lost the shots attempted battle, the Thrashers did much better in this other measure of territorial pressure having an equal number of shifts that ended in the offensive and defensive ends of the ice.
Defensemen
+2 Enstrom
+2 Havelid
+5 Hainsey
+5 Bogosian
-7 Exelby
-7 Kulda
The top two lines had more shifts that ended in the opposition's end. This is a second measure confirms that the 3rd pairing was struggling to hang on. I find it VERY interesting that there is almost no difference when you compare forward lines, but a wide difference when you compare defense pairings. It is just one game but if this data pattern holds over multiple games it suggests to me that improving the 3rd pairing could have a real benefit to the overall team performance.
Forwards
+1 Kovalchuk
+1 Christensen
+1 Williams
-1 Sterling
+0 Little
-1 Armstrong
-1 Reasoner
+0 Stuart
+0 Thorburn
+0 Slater
+0 Stevenson
+0 Boulton
I find it interesting that the Reasoner line was out shot so badly but didn't have a lopsided pattern of where shifts ended. Did Lehtonen face long shots that he immediately played to the defense when the Reasoner line was out there? Because if they were getting out shot and he was covering the puck or deflecting it into the netting we should see an end-of-shift-face off gap.
The Oiler and Flames bloggers have been tracking the "Corsi numbers" or attempted shots since the NHL started making this data available in the middle of last season. So let's take a look at how all four lines and three defense pairings fared last night. Attempted shots are calculated as follows = Shots on Goal + blocked shots + missed shots. The Corsi number is simply Attempted Shots For minus Attempted Shots Against. This only counts the SAF/SAA at Even Strength (excluding all Empty Net situations).
Corsi +/-
Goaltending/Team
-12 Lehtonen
In the Detroit pre-season game each club scored 1 goal at even strength (Detroit added 2 PP goals and a EN goal). A tie at even strength is good, but the Thrashers were outchanced at even strength by a margin of 12.
Defensemen
+5 Enstrom
+1 Havelid
When the top pairing was out the Thrashers outplayed Detroit.
-4 Hainsey
-2 Bogosian
The second pairing was mildly outplayed.
-12 Exelby
-12 Kulda
The third pairing was just hanging and was lucky to get off the ice without allowing a goal. Did I mention that having Exelby on the roster is holding this team back?
Forwards
-2 Kovalchuk
0 Christensen
-2 Williams
The top line was even or slightly worse. Even is not good enough. You need your top line to outplay the opposition to have a chance to win most nights.
-3 Sterling
-4 Little
-6 Armstrong
Detroit had a mild advantage against the 2nd line.
-8 Reasoner
-9 Stuart
-6 Thorburn
The third line was in their own end a ton. When I saw these numbers I assumed that the Reasoner line was matched up against Zetterberg's line and that is why they were so badly out chanced. Then I looked at the shift chart I was very surprised to see that Reasoner was almost never matched up against Detroit's top scoring line. Those are some ugly numbers considering they didn't even face the toughest line Detroit could put on the ice.
+2 Slater
+0 Stevenson
+2 Boulton
The forth line was the only positive forward line. They did exactly what you expect on the 4th line--play at the enemy's end of the ice.
Conclusion
The 4th line and the top defense pairing outplayed the opposition using Net Shots Attempted as a measure. If the readers enjoyed this I'll do some posts like this in the future.
EDIT/UPDATE: Here's some more cool information. Vic Ferrari has also made this cool program to show you which players where on the ice when a shift ends with a stoppage. (You want stoppage in the opposition end because that means you were applying pressure down there.) This is another plus/minus sort of stat, how many of your shifts ended in the opposition end (good) minus the number of your shifts that ended in your defensive end (bad).
Goaltending/Team
+0 Lehtonen
For a team that was lost the shots attempted battle, the Thrashers did much better in this other measure of territorial pressure having an equal number of shifts that ended in the offensive and defensive ends of the ice.
Defensemen
+2 Enstrom
+2 Havelid
+5 Hainsey
+5 Bogosian
-7 Exelby
-7 Kulda
The top two lines had more shifts that ended in the opposition's end. This is a second measure confirms that the 3rd pairing was struggling to hang on. I find it VERY interesting that there is almost no difference when you compare forward lines, but a wide difference when you compare defense pairings. It is just one game but if this data pattern holds over multiple games it suggests to me that improving the 3rd pairing could have a real benefit to the overall team performance.
Forwards
+1 Kovalchuk
+1 Christensen
+1 Williams
-1 Sterling
+0 Little
-1 Armstrong
-1 Reasoner
+0 Stuart
+0 Thorburn
+0 Slater
+0 Stevenson
+0 Boulton
I find it interesting that the Reasoner line was out shot so badly but didn't have a lopsided pattern of where shifts ended. Did Lehtonen face long shots that he immediately played to the defense when the Reasoner line was out there? Because if they were getting out shot and he was covering the puck or deflecting it into the netting we should see an end-of-shift-face off gap.
posted by The Falconer at
1:06 AM
9 Comments:
Very interesting stats Falconer, thanks for sharing ! X really looked all over the place as usual Wednesday, and I don't mean that in a good way. These figures only confirms his lack of sound positioning...
By
Anonymous, at 3:01 AM
Great analysis.....not sure how much it meams, but interesting nonetheless. Would be curious how that analysis plays out over a few games during the season. It does reflect my thoughts that X and Kulda did not look good at all in that game and the Slater line seemed to play well in the short time they were out there. keep up the good work.
By
h, at 8:36 AM
Great analysis. Poor Exelby, I love the guy. Maybe some of the more experienced D we get can teach him positioning and passing skills. I think that those things can be learned, while hitting and aggression on the other hand are innate attributes of a player. Since he already has the aggression, he could become a good D player if he acquired the other skills.
By
Anonymous, at 9:45 AM
Facloner,
Here's a bizarre question - how does the NHL track shifts so precisely? I score youth hockey games and have enough trouble keeping track of penalties, SOG, etc.., I'm guessing the NHL attaches some type of RFD device to the players uniform so that they can electronically capture when a player is stepping on and off the ice. I'm particulary curious as to how the NHL is monitoring the locations of the players when a shift ends with a stoppage.
By
Anonymous, at 12:59 PM
Awesome!! I love numbers. :D
Kovy's line played against Detroit's top line most of the game. I remember Slater's line being out against McCarty's group. And I remember seeing Reasoner and Co. spend a lot of time battling around the net - extra chances for the opposition.
You spoke of improving the third d pairing but that's about an automatic given. With Schneider's insertion into the line-up, Hainsey likely moves down to that third pairing with X/Valabik.
Also re: the D stats, they all seem to have pretty much the same numbers (which would make sense as they go out in pairs) so in fact a piss-poor performance by one player on D could skew the numbers for his partner - or vice versa. So is XLB still that bad or did his numbers get swayed by a young rookie? Perhaps vice versa, the young rookie performed decent but Mr. -21 dragged him down.
By
Anonymous, at 1:48 PM
d: every thing is time stamped to the game clock. There are about 5-6 guys who have responsibilities to cover certain players. So guy #1 might have the Thrashers defensemen all night. When they change on the fly he simply presses the new player's number and it goes into the time file.
When a players attempts a shot someone else has screen and they press the touch screen where the shot was taken and if it is blocked or misses the net (otherwise it is counted as an official SOG).
Using the game file you can see who was on the ice for every Shot attempted, who was on the ice at each stoppage and for each faceoff.
There is some error in the data, but the ice time is highly reliable. The errors that have are discussion are in when they mark the location of where shots occur. The data are not perfect, but even with a small error rate the bulk of it can tell us things about our players that we would not otherwise know.
By
The Falconer, at 2:08 PM
sara: you're absolutely right that D pairs. It is impossible to distinguish who is doing what out there since they nearly always play together. Over the course of the season you get more mixing and matching so you MIGHT be able to get some leverage on which guy is better, but in the short run we can't tell if it was Kulda or Exelby that was struggling just looking at these stats. That's one reason I grouped them by lines and d pairs, since it only really informs our understanding of GROUP performance.
By
The Falconer, at 2:11 PM
Great stuff as always, Falconer.
Any idea how someone gets into the "NHL Stat button pusher" profession? I'm a button pusher already but I'd rather get paid to push buttons while watching hockey.
Regarding the stoppage of play stat and what end it is in when play is stopped, I find it interesting that Lehtonen was +/- 0 in that stat given the shot differential. I've always read that as a goalie, you want to stop play when you can unless it is 100% safe to move the puck to one of your skaters. Logic being that a goalie wants to control the pace of the game as much as possible. I wonder if there is a stat for where on goal a shot is placed- i.e. shot at blocker side high or 5-hole, etc. or if there is a stat for rebounds given up. If a team is purposfully trying to go blocker side or low it should lead to less stoppages and more rebounds for them. I'd be interested in seeing how stats like that influenced other stats like the stoppage one.
By
Anonymous, at 7:11 PM
PF:
1) I too find it surprising we lost the "shots attempted" battle but were even in the "end of shift face off location" battle. I will say that when I was watching the game I thought the Thrashers were being very good about moving the puck out of their own end. If I was there in person tonight I would pay more attention to that.
2) I know that for a while Fox Sports (I think) had a chart showing what part of the net a shot was directed at. A while back I posted a graphic made by someone else that showed Lehtonen's SV% by area of the net: high glove, high blocker, 5 hole, low stick, low glove. I'll see if I did up that picture for you.
By
The Falconer, at 8:20 PM
Post a Comment
<< Home