An Early Look at the Rangers’ Penalty Killers

[Thirty One]

An Early Look at the Rangers’ Penalty Killers

I'm tracking scoring chance data for the Rangers on the penalty kill this season. Here are the early results.

Click to read more...

 

[silverfish]

Love the work you're doing, -31-, it's really great!

Are you using a tracking system to gather this data? Or are you just going back and DVRing games, watching them a little more closely? Either way, it's really cool analysis to see.

But honestly, did we really need another Glass thread

 

[Megustaelhockey]

I love this, too. Thanks for sharing your work with the community.

 

[Thirty One]

Love the work you're doing, -31-, it's really great!

Are you using a tracking system to gather this data? Or are you just going back and DVRing games, watching them a little more closely? Either way, it's really cool analysis to see.

But honestly, did we really need another Glass thread

With powerplays, I go back and re-watch to track what I'm looking at. For penalty kills, I can do it live as I'm only counting scoring chances and all the other information I need is tracked elsewhere.

 

[McRanger]

I am surprised you are using scoring chances over something involving shots like Corsi or Fenwick.

The usual advanced metrics seem to mostly dismiss shot quality. Does it matter more on 4-on-5 than 5-on-5?

 

[Raspewtin]

I am surprised you are using scoring chances over something involving shots like Corsi or Fenwick.

The usual advanced metrics seem to mostly dismiss shot quality. Does it matter more on 4-on-5 than 5-on-5?

Well they are a lot easier to track than 5v5 scoring chances, and Corsi isn't really an adequate stat when the team is shorthanded (corsi is pretty much always negative for the shorthanded team, as the team on the powerplay is very likely getting more shot attempts). Scoring chance quality matters more than shot attempts on the PP.

 

[Thirty One]

I am surprised you are using scoring chances over something involving shots like Corsi or Fenwick.

The usual advanced metrics seem to mostly dismiss shot quality. Does it matter more on 4-on-5 than 5-on-5?

They don't really dismiss shot quality, but over the long term ignoring it at even-strength tends to make little difference. The difference in percentage of shots that go in inside and outside the home plate area is large.

 

[McRanger]

Well they are a lot easier to track than 5v5 scoring chances, and Corsi isn't really an adequate stat when the team is shorthanded (corsi is pretty much always negative for the shorthanded team, as the team on the powerplay is very likely getting more shot attempts). Scoring chance quality matters more than shot attempts on the PP.

Obviously.

I am just surprised that the study does not focus on possession. Considering that a large part of killing a penalty is spending time outside your own zone.

Offensive/neutral zone pressure to stop the team from breaking into the zone and/or quickly clearing the puck is very important. Shot metrics would probably help measure how much time is actually spent in the zone which would probably lead to scoring chances. I would think the longer you are in your own zone the more likely you will give up a goal.

I'm not advocating using this metric, I was just surprised that shots weren't somehow factored in.

 

[McRanger]

They don't really dismiss shot quality, but over the long term ignoring it at even-strength tends to make little difference.

Is that solely based the variance of goalie save percentage? That is the only explanation I have ever read as to why shot quality could be ignored. And its never much made sense considering it seems to disregard the most important factor in whether or not a shot gets saved, regardless of its quality: the goalie.

The difference in percentage of shots that go in inside and outside the home plate area is large.

I would think so. I also just think that was the case no matter what the situation.

 

[Thirty One]

Obviously scoring chances are strongly correlated to zone time. Even in the short sample this season, the results show that clearly.

Shot quantity data is easily collected and widely distributed, so I didn't feel the need to include it here.

I forget the exact figures, but at even strength shot location accounts for a very small sliver of a team's success. I think that the highest shooting percentage team had less than 5% of their success explained by shot location. Lots of work is being done to capture those couple percentage points more easily, but until then we can focus on the quantity with little loss of information.

Does location account for more of the discrepancy on the powerplay than at even strength? I don't know. Intuitively it would seem so. Considering teams' 5-on-5 shooting percentages ranged from 5.84 to 9.83 last season, while 5-on-4 percentages ranged from 6.97 to 15.09, I think there's probably something there.

Comparing the shot charts from the top to the bottom in both situations seems to bear that out: