G5 2016-10-20 Oilers have some Cordon Blues for dinner

Game Notes EDM vs STL

St. Louis Blues (Head Coach: Ken Hitchcock) at Edmonton Oilers (Head Coach: Todd McLellan)

NHL Game #60, Rogers Place, 2016-10-20 07:00:00PM (GMT -0600)

Oilers 3-1

Referees: Justin StPierre, Graham Skilliter
Linesmen: Shane Heyer, Mark Shewchyk

Three Stars: Cam Talbot; Milan Lucic; Jake Allen

  • Even *I* don’t care what the charts say, I’ll take a regulation win over the Blues any day of the week!
  • It was a pretty dull game for the most part. One thing to remember, though, is that STL is a very very good team, especially defensively, so ‘hanging with STL’ will usually mean a lot of time playing a somewhat dull heavy game. I don’t think the Oil are anywhere near good enough yet to force them into a track meet consistently, but that third period though!
  • Go Yak City! He gets a goal, the Oilers get a win, the Flames lose, life is good.
  • Tarasenko is quite a terrifying player.
  • Some things to look for in the fancies: the Blues had the better of all them, except for two: Dangerous Fenwick … and the score!
  • That’s not a fluke, other sites show the Oilers won the chance and high danger chance battle. And the name of the two warriors who made it happen are McDavid and Lucic.
  • You *have* to check out the rink map of shots against Klefbom and Larsson. Unbelievable. And they were matched against the Tarasenko line. The other two pairs?  Hmmm, perhaps not so good.  Seksy was good apart from a few  minutes when he was feeling a little generous with the puck.
  • Larsson and Klefbom were negative on the raw numbers, but positive on the danger numbers. This is a consistent thing. We saw this in Larsson’s numbers with Jersey, so I don’t think it’s a small sample fluke. (though in this game it was moreso Klefbom)
  • McDavid was also negative on the raw numbers (rare), but positive on the dangers (common). It’s a thing. You can’t stop the McDavid, you can only hope to preserve at least one of your limbs when facing him.

Obligatory Russell Rant

OK, OK, no one wants to hear more about this Russell vs Analytics nonsense … but please hear me out anyway.

Good Lord, it’s bizarre watching the hardon the MSM has for Kris Russell. I assume that is because they think he is somehow “disproving analytics”.

Except it’s not at all true. If Russell were playing this well and his stats were similar to last year’s poor numbers, sure, I guess you could argue that. But that’s not the case. At all.

The reality is that Russell’s fancystats this year are right in line with the visual. They’re way better so far this year. If he was doing this well statistically while in Calgary, there would have been less trepidation about his signing amongst fancystatters.

He is in fact running at +1% or so on the CorsiRels, which is about 5% better than he was doing in Calgary.

He is running at +18% (!!) or so on the xGFRels, which is about 20% better than he was doing in Calgary.

That’s a massive change. In other words, there is no ‘disproving’ of analytics going on. The analytics are basically saying exactly what the eye is telling you. Kris Russell is playing pretty well. In fact, this is one of the best four or five games statistically he’s had in the last three years. (Not unheard of for him, but very rare).

The question now is whether you believe that in one offseason he’s really improved that remarkably – or whether he’s just on a hot streak and going to regress back to the player he was the last 200+ games.

I expect the latter, though I am cheering for the former.

Honestly, I’ve quite liked Russell’s game. Yeah, he’s still a bit of a tire fire when not with McDavid. But he seems to know his job with McDavid and Sekera is to make the simple defensive play and get it to one of those guys, and he’s been doing that well.

In theory, anyone should be able to look good playing with McDavid. But he’s doing better in that role than some others, and that has value IMO.

Russell being the player he has been this year instead of the player he has been the last three years would be a genuine help to a still not fully-baked defense, and I am cheering for him – but we’ve seen shows like this before, and they didn’t last, so you’ll forgive me if I remain skeptical that this sudden and dramatic a jump is sustainable.

Game at a Glance


  • Tyler Pitlick, Milan Lucic, Connor McDavid each had 1 goal
  • Patrick Maroon, Milan Lucic, Connor McDeity each put 4 shots on goal
  • Milan Lucic was most active on the team with 9 shot attempts
  • Leon Draisaitl was the faceoff champ at 77.8% (7-2)
  • Anton Slepyshev led the team in on-ice 5v5 shot attempts at 57.9% (11-8) – hey Anton!
  • Anton Lander, Mark Letestu trailed the team in on-ice shot attempts with 5v5 CF% of 18.2 – hey Anton!
  • Zack Kassian, Benoit Pouliot each thumped the other team 3 times

Goal Overview

Team Period Time Strength ShotType ShotDist Danger
STL 2 3:37 EV Wrist 6.0 2.91 YAKS!
EDM 2 6:22 EV Tip-In 12.0 1.84
EDM 3 0:36 EV Wrist 14.0 2.17
EDM 3 19:22 EN Backhand 24.0 0.43

Who Won the 5v5 Shot Battle?

Which Battle Who Won By How Much
Shots STL 26 to 21 (55.3%)
Average Shot Distance Against (ft) EDM 37 to 28
Corsi STL 63 to 45 (58.3%)
Score & Venue Adjusted Corsi STL 64 to 44 (59.0%)
Fenwick STL 36 to 30 (54.5%)
Dangerous Fenwick Tied 35 to 35 (actually, the Oil are just ahead)

Detailed Metrics

Shot Metrics
EV 45 63 41.7 44.5 64 41 30 36 45.5 35.5 34.7 50.6
All 51 77 39.8 50.7 77.8 39.4 33 47 41.2 40.7 46.5 46.7
Other Metrics
Team PP PPG PIM FO Hits Giveaways Takeaways
Oilers 2 0 6 57.6 22 12 12
Blues 3 0 4 42.4 22 12 7

How the Players Did (On Ice Shot Attempts)

Danger Tables

Forwards are sorted by decreasing CF%. Defensemen and pairs sorted by increasing DFA60. Forward lines by decreasing DFF%. Positions are as listed by the NHL roster page, not necessarily where they played.

Eberle 16:38 62.5 22 23 48.9 47.8 50 60.8
Drai 13:07 37.5 14 15 48.3 47.8 57.9 54.3
McDavid 19:47 52.6 22 29 43.1 42.2 45.2 54
Nuge 14:10 22.2 9 18 33.3 33.3 35.7 31.5
Pitlick 06:47 25.0 5 11 31.2 30 37.5 35.4
Letestube 04:57 20.0 2 9 18.2 17.5 28.6 45.9
Lander 06:11 8.3 2 9 18.2 17.3 28.6 33.6
Slappy 10:06 50.0 11 8 57.9 57.5 61.5 76.2
Looch 16:43 71.4 22 19 53.7 53 53.8 64.3
Maroon 13:15 37.5 15 15 50 49.2 55.6 54.2
Kassian 13:30 33.3 6 14 30 29.9 30.8 21
Poo 14:12 20.0 6 17 26.1 25.9 28.6 17.7
Klefbom 16:23 26.7 14 15 48.3 48 61.5 67.2 15.75
Larsson 18:42 31.2 13 23 36.1 35.8 44.4 46.3 28.56
Russell 19:54 61.1 19 29 39.6 38.8 41.9 51.2 41.31
Seksy 18:56 58.8 17 30 36.2 35.4 41.4 42.3 50.39
Nurse 13:32 16.7 13 15 46.4 45.4 45 50.4 58.97
Gryba 13:10 16.7 13 16 44.8 44.1 42.9 49.4 62.89
Defense Pairings
Larsson Klefbom 15:40 16.47 36 11 15 42.3 41.9 54.5 58.2
Seksy Russell 17:17 43.39 42.6 16 25 39 38.2 42.3 47.9
Nurse Gryba 12:12 65.41 27 13 14 48.1 47.3 45 50.4
Forward Lines
Maroon Drai Slappy 09:20 11 8 57.9 57.5 61.5 76.2
Eberle Looch McDavid 14:01 21 17 55.3 54.3 52 63.7
Pitlick Lander Letestube 03:07 2 5 28.6 27.4 40 51.8
Kassian Poo Nuge 11:21 5 12 29.4 29.2 30 22.3

Game Flows

Rink Maps

Defense Pairings – Shots Given Up

Forward Lines – Shots Taken

Head to Head

NHL Media Highlights

Click the link to play the associated highlight video

Allen’s high blocker save
Yakupov opens the scoring
Lucic finishes off McDavid’s feed
Pitlick’s bizarre deflection
Allen stones Eberle
Allen’s two nice saves
Talbot’s strong pad save
Allen’s fantastic pad save

14 thoughts on “G5 2016-10-20 Oilers have some Cordon Blues for dinner

  1. Awesome analysis after every game! And I liked the comments about Russel. The people who dislike analytics are the people who don’t understand what you are trying to say with them. Cheers!


      1. To be fair, you didn’t miss it. I looked at your Yak column on the head to head and was compelled to look it up. I don’t know how effortless your kung-fu is but… since you have the oil cf% and dff% totals head to head, the opposition totals would be insightful as well. Another one for the suggestion box.


  2. “The question now is whether you believe that in one offseason he’s really improved that remarkably – or whether he’s just on a hot streak and going to regress back to the player he was the last 200+ games.”

    I disagree with the above.

    I didn’t hear anything last year that Russell was having a poor season. All I heard is Russell’s corsi is bad but he was in demand by Hartley and later the Stars. It was later insinuated that he was being used too much by Hartley and the Stars and that these people were behind the times for analyzing players, which I completely bought into.

    What I think is that he was playing fine last year and he was a good player and is a good player. I think the system that the Flames play wasn’t good for his possession numbers but he still does a lot of things well that teams covet; it’s that these things aren’t captured in corsi, fenwick, etc. Hockey analytics has a long way to go still and it isn’t capturing everything.

    In the end, Russell 1, Analytics 0.


    1. It’s certainly possible it was the system. I figured if it was, he’d do well in Dallas and he was terrible there. I live in Calgary, I watch Flames games more often than I like, and I saw him bad there too, lots. Those bad numbers came from repeated tire fires, and Russell was in the middle of a lot of those tire fires. He’s been as good so far this year as any stretch of games I’ve seen out of him the last few years. He’s had streaks like this before. He always regressed. I assume that’s why the Flames didn’t re-sign him. I assume that’s why the Stars didn’t re-sign him, even though they paid a pretty penny for him. I assume that’s why the Oilers only gave him a one year show-me contract on the eve of the regular season. Hey, it would be great if he didn’t regress this time, he’ll help the Oilers win and that’s what we all want. But like I say, you’ll forgive me if I remain skeptical.


  3. Great write up G. Amazing how much info you pull from a game in such a short period of time. Truly remarkable. Klef/Larsson’s performance against Tarasenko is ridiculous. Wow.

    Regarding the Russell/analytics conversation, is it possible that the fancy stats are fantastic for analyzing what has happened (previously) and evaluating components of an individual’s game (historically in very specific situations), but not-so-much for predicting what will happen going forward or how “good” a player actually is?

    I guess what I’m asking is, you suggest that the only way Russel keeps up the improved stats is if he has – in fact – improved as a player (and you’re hopeful but skeptical). But…. could it be that he is the exact same player with the exact same “value” but on a different team with a different approach and different intangibles which leads to improved stats? If so, then to me this would suggest the predictive value of fancy stats is limited by the degree to which that player’s previous environment/intangibles remain the exact same. In that case, Russel’s fancy stats with Calgary and Dallas would tell us very little about how well he’ll do in Edmonton, and his performance thus far shouldn’t be so shocking.

    Thanks, and again, really enjoy reading your recaps…. very impressive stuff.


    1. Thanks Sean!

      Yes, you raise a couple of really good points that are worth delving into.

      1 – fancystats (all of them, even points!) are heavily dependent on context. Change the context, and you should *expect* a change in fancystats. Radical idea, hey! New system, new teammates, new role, perhaps the eye opening humility of not getting a contract until October! We should be able to say before it even happens that it’s not even that Russell’s fancystats *might* change, but it’s almost certain they *will* change now that he’s in Edmonton.

      2 – I think this is doubly true for defensemen. What they do seems to be a bit more complicated than with forwards, so they are inherently harder to analyze (I think that’s true by eye, not just by stats). My observation is that fancystats, and scouting, ‘break’ more often with defensemen. I wrote about this a little over a year ago: https://oilersnerdalert.wordpress.com/2015/10/16/fancystats-and-defensemen-an-uneasy-combination/

      Put 1+2 together and you get a situation where the outcome is a lot more uncertain than people seem to act it will be. (One of the ways I always cheat in my player analysis is I conclude “I think this guy will help” or “won’t help” … that way on those occasions where I’m wrong, as I inevitably will be, I have an out!)

      The other thing we don’t talk about enough is value and role. A $5M signing to play top pair is a way different situation than a $2M signing to play bottom pair – and that’s with the same player.

      Now that said, when I say I “remain skeptical”, it is that I would be very surprised if such a radical change in the fancystats, even with all the uncertainty I noted above, sticks.

      Poor players can have hot streaks and look good, just as good players can have cold streaks and look bad.

      At about the 15 game mark, the TOI for Russell should have increased to the point that his fancystats will start to stabilize, and I will begin to believe the predictive value of his numbers. We’ll have a better sense of what EDM/Russell is going to be, in contrast to CGY/Russell. And I hope that’s a good number, or at least a better one than he posted last year.

      Until then, most of the arguments (on either side) are still mostly bickering over small samples!


      1. Thanks for the response G! I appreciate your balanced perspective.

        Truth be told, I’m very intrigued by the emergence of fancy stats in the NHL as a tool for evaluation/prediction, I just can’t quite settle on how much is cause/effect, and how much is correlation. Regardless, if the Oilers keep winning I don’t care!

        Liked by 1 person

  4. Very good points Sean. I am in agreement however What G is saying is similar with everything in statistics or design or some other way we try to “predict the future”. If an engineer designs a retaining wall for one site. He can just blindly use it for another. There may be a total set of different circumstances that need to be considered. That’s where I think understanding of statistics and how they are used and why they are used is so important. Any body can just spout of numbers saying this Corsi for example doesn’t reflect the players “eye test”. They don’t necessarily understand what the Corsi stat is trying to tell you.

    Liked by 1 person

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s