Oilers analytic/advanced stat thread

[nexttothemoon]

You realize we're not talking about the same thing here, right? The issue isn't which measure is correlated with success, but which is most predictive of future success.

You posted a chart that showed the R squared value for pts per game vs Corsi%. (The R squared value on that chart isn't correct by the way... put the numbers into a spreadsheet and you'll see the figures I posted above are correct.)

Regardless, that chart you posted doesn't show anything about predictive value... it simply shows how well the data fits the regression line... ie not well compared to stats based on actual goals for/against.

 

[Stoneman89]

Bob Stauffer‏
The Edmonton Oilers are:
#1 in Fenwick (54.43)
#2 in Corsi (53.89)
#6 in shots for (34.5)
#7 in shots against (30.5)

Great metrics for us. How's that been working?

 

[KeithIsActuallyBad]

Funny enough the Jackets won the Corsi battle last night despite looking terrible for most of the game.

 

[nexttothemoon]

Bob Stauffer‏
The Edmonton Oilers are:
#1 in Fenwick (54.43)
#2 in Corsi (53.89)
#6 in shots for (34.5)
#7 in shots against (30.5)

Great metrics for us. How's that been working?

I'm an Anti-Corsi guy in general and think those stats are close to useless (reality based stats using GF/GA are logically and demonstrably far better and more reliable) but I will say that it is a good thing that they are outshooting their competition.

Once they get their shooting percentage rate up (it's slowly rising... now up to 12th worst in the NHL) and their team save pct up (a much, much slower process... still 2nd worst in the NHL)... this team will begin to rise back into playoff contention once again.

If the team had a low corsi/shots % AND a crappy save percentage and shooting percentage as well... they'd be truly doomed... (like Buffalo/Arizona).

 

[McGoMcD]

Bob Stauffer‏
The Edmonton Oilers are:
#1 in Fenwick (54.43)
#2 in Corsi (53.89)
#6 in shots for (34.5)
#7 in shots against (30.5)

Great metrics for us. How's that been working?

As much as advanced stats are not perfect the oilers have had a lot of bad luck to be where they are. They could easily be 16-13 instead of 13-16. They will turn in on the rest of the way.

 

[Connor McConnor]

The Oilers have the record they have because our special teams are a complete disaster.. They should not be 15th on the PP and last in the league on PK and expect to win many games. I've watched more than half of the games this season and we often dominate a lot of the 5 on 5 play minus brutal giveaways in the neutral zone that lead to odd man rushes and often times goals on the other end. However, anytime the other team gets a PP I am almost assured they are scoring whereas we seem to actually LOSE momentum on the PP.. If Pulj doesn't start getting more PP time McClellan should be fired just because of that.

 

[RebuildTheRebuild]

Those stats that Stauffer posted at that time just need to be looked at with a critical eye. Need to ask why is that our shot metrics are good but it is not translating to wins or goals for that matter. One thing is that I actually think the shot metrics are misleading due to the fact that Oilers have been on the wrong end of the scoreboard for most of the games they play and teams tend to take their foot off the pedal a bit when in the lead. So I am guessing score effects has skewed it to look better than what we actually see. I still think though Oilers would have at least decent shot metrics even without score effects.

The reason this wasn't translating to goals early on in the season is because a good proportion of our shots were coming from defenders from the point. These are the lowest % shots you can take. Now we are seeing more goals from the team recently a bit is luck and better shot selection. Then as well we have had some not so good goaltending as well. Also at that time of that tweet our even strength goal differential was +2.

With all this being said I think it does tell the story of the Oilers. They are a good even strength with the potential to really be one of the better teams there as well. The reason why we are so low is because of our horrible special teams. We cannot get it done on the special teams and it has really killed our season so far. This is the way I saw those stats.

 

[nexttothemoon]

Another quick point about those stats is that shots/corsi/fenwick are all very highly correlated so really they are all measuring the same thing... it's not as if there are really several different categories there that they are doing well in.

Also, as noted above... shots/corsi/fenwick are all inferior in their correlations to team success.

The best corsi team in the NHL missed the playoffs last season. That renders that stat useless imo.

Find me a season where the best PDO team or the best GF% team misses the playoffs... you obviously will not.. because those stats are obviously relevant and indicative of the best teams which put the puck in the net more often than their opponents do... which is the actual point of a game of hockey.

 

[Paralyzer008]

Yeah, the underlying 5v5 stats actually don't look too bad - the special teams are a disaster. At 5v5, Oilers have a 50.7% GF. That's interesting to me. That suggests a good team.

Also, some of the 5v5 stats are driven by a league high amount of D shots because Klefbom and Nurse shoot everything they see.

I personally think possession is just part of the puzzle - you need good goaltending and a good shooting percentage that's not solely driven by luck AKA shooting talent.

Oilers have a .894% save percentage for all situations, 30th in the NHL, due to horrid special teams and mediocre 5v5 goaltending play (and, terrible D mistakes). Their total shooting percentage is also a bit low at 8.67%, usually the best teams are higher than that. A 98 PDO suggests a bounce back is possible, especially in net cause Talbot can definitely play better.

 

[redgrant]

Who'd have figured firing a million shots into a goalies crest wont net a few goals?

 

[nexttothemoon]

Save percentage overall is now DFL in the entire NHL.

3.5% lower than LA. Every 100 shots against, the Oilers let in 3.5 more goals than the Kings do. Over the course of a full season that's going to be a differential of 85+ goals just based on that save percentage alone.

PK save percentage is 78.5% which is 13.3% worse than the Sharks who have the best save pct on the PK. The NHL median save pct on the PK is ~88% so the Oilers are 9.5% below even that level.

 

[oXo Cube]

Save percentage overall is now DFL in the entire NHL.

3.5% lower than LA. Every 100 shots against, the Oilers let in 3.5 more goals than the Kings do. Over the course of a full season that's going to be a differential of 85+ goals just based on that save percentage alone.

PK save percentage is 78.5% which is 13.3% worse than the Sharks who have the best save pct on the PK. The NHL median save pct on the PK is ~88% so the Oilers are 9.5% below even that level.

Yep. Goaltending is a massive problem and has been all season.

If you get scored on every single time your D makes a mistake you aren't being given a chance to win.

 

[CantHaveTkachev]

Save percentage overall is now DFL in the entire NHL.

3.5% lower than LA. Every 100 shots against, the Oilers let in 3.5 more goals than the Kings do. Over the course of a full season that's going to be a differential of 85+ goals just based on that save percentage alone.

PK save percentage is 78.5% which is 13.3% worse than the Sharks who have the best save pct on the PK. The NHL median save pct on the PK is ~88% so the Oilers are 9.5% below even that level.

this is our season right here

 

[KeithIsActuallyBad]

Who'd have figured firing a million shots into a goalies crest wont net a few goals?

For all that "dominance" the Oilers had in the first last night Natural Stat Trick had the Oilers and Preds with equal high danger scoring chances.

Working hard at accomplishing nothing.

 

[Stoneman89]

For all that "dominance" the Oilers had in the first last night Natural Stat Trick had the Oilers and Preds with equal high danger scoring chances.

Working hard at accomplishing nothing.

Won the metrics battle again last night. That should be all that matters, right?

 

[KeithIsActuallyBad]

Won the metrics battle again last night. That should be all that matters, right?

If nothing else this season is proving that advanced stats are basically worthless in determining anything.

 

[McDNicks17]

If nothing else this season is proving that advanced stats are basically worthless in determining anything.

I would say it's more proving that it doesn't matter how good you are at 5v5 if your special teams are historically bad.

 

[KeithIsActuallyBad]

I would say it's more proving that it doesn't matter how good you are at 5v5 if your special teams are historically bad.

Well they didn't score at 5v5 or on their special teams yesterday despite all that "dominance".

 

[nexttothemoon]

If nothing else this season is proving that advanced stats are basically worthless in determining anything.

I agree... "advanced stats" aren't even advanced... corsi/fenwick/shots/scoring chances all just measure attempts on the net (in ever so slightly different ways) which obviously don't mean anything when you can't put the puck in or keep it out.

 

[CantHaveTkachev]

I would say it's more proving that it doesn't matter how good you are at 5v5 if your special teams are historically bad.

and goaltending...the Great Equalizer

 

[Aceboogie]

Its been a bloodbath season but figure Id revive this thread to post on a cool new tool being developed. Basically it tracks shot contributions, zone entries and zone exits. Its still very early in development so no conclusions should be made. But defiantly something intriguing/neat to look at in the future.

There is data from last year too on this stuff out there. I don't have it posted here but its good to compare back to last year to see if things align

Ill post some random comparisons

-With all the data that was tracked on Nurse last year as well, hes really turning into an great puck manager. Both at zone exits and entries. Something special in him

- I am impressed with Russell here. Contributing offense and zone exits with possession are improving and maybe that's to do with being more comfortable with the systems. Also Sekera was not here to be the puck mover on that pair, as we can see Russell still a bit inefficient at getting the puck out with possession. Benning still has some solid underlying metrics

- Color me very impressed with Larsson. Hopefully this trend continues

Not much to say

And finally, for the laughs

 

[McDNicks17]

Interesting stuff.

You can see why Klefbom's offense has died.

 

[McDNicks17]

Is there an easy way to link those?

I was curious to see Davidson's and it's ugggggggggly.

 

[PinSeeker]

I like Sean Tierney on Twitter for charting: Tableau Public

 

[Aceboogie]

Interesting stuff.

You can see why Klefbom's offense has died.

Few thoughts:

1. His shooting % is shockingly low at 2.5%. Alot of the reason has to do with simply being unlucky. Its not the whole reason, but a low shooting % will sink offense for a season. Not really that considered as obviously hes still generating the chances and shots

2. To slightly contradict the first point. While he does have crap ton of shots, shots from D are just way lower danger chances. D that can move the puck around and get into closer slot areas get better average than D that just hammer the puck from the blueline. Klefbom moreso this yera has been firing a ton of pucks on net, but from the blueline area. Hes also alot more shoot first, whereas last year he was more of a puckmover. I think this has to do with coaching though, as in a preseason interview he said he talked with coaches and a goal was to get 250 shots this year. So Id have to think they wanted to set him up as a big time blueline shooter, vs as a nimble puck mover who also could shoot. We have some good forwards on this team, so maybe its better to move the puck to them for better chances

3. Other teams saw excel last year and gameplanned him big time. Aspects of his game could be figured out by the opposition. Itll be on him and the coaches to adjust for the future