Which Oilers played against the opposition’s best defenders? Taylor Hall and Jordan Eberle

Quality of Competition is one of the few advanced statistical metrics that virtually nobody argues lacks value. Success is success, but there’s a difference between succeeding against the other team’s first line and their third line.

Until now, however, most Quality of Competition rankings have been based on plus/minus or one of the shot metrics, and have failed to differentiate between playing against the other team’s best forwards and their best defencemen. What happens if we tweak those numbers for the Edmonton Oilers?

Eric T. has been killing it at NHLNumbers, and he has the answer for us. On Thursday, he suggested a quality of competition metric based on total ice-time – with the idea being that rather than try to use some sort of statistical metric to rate players, we just trust that the coaches have a good idea of which guys are the most talented and play them accordingly. He proposed another change as well – rather than a single number for Quality of Competition, a chart which showed each player’s position against forwards and defencemen. In other words, if Ryan Nugent-Hopkins was matched up against the Malhotra line in a Vancouver game, but saw Bieksa/Hamhuis as the defence pairing, he’d get a low number for the forwards he played against and a high number for the defencemen – better reflecting what situations he was playing in than a single number.

On Friday, Eric published charts for every team in the league. The following is the one he published for the Oilers.

One note on the chart before we dive into this data: the defenseman axis (on the left) has a difference of 2.4 minutes between top and bottom, while the difference on the forward axis is 4.5 minutes. In other words, the same size gap in both directions means almost twice as much on the forward chart as it does on the defenceman chart.

The guys high up the chart played against the opposition’s best defenders, and it’s not hard to see that NHL coaches saw things sensibly. Jordan Eberle and Taylor Hall faced the toughest defenders, followed by Nugent-Hopkins, with Gagner, Hemsky, Smyth and Horcoff all seeing impressive opponents. Jones, Paajarvi, Belanger, are in the third tier, and the fourth-line guys could be comfortable in knowing that they’d be playing against the other team’s versions of Colten Teubert, Cam Barker and Theo Peckham virtually every game.

The toughest opposing forwards is a different story. Horcoff and Smyth lead the way in that department, followed by Hemsky and then with the kids, Gagner and Jones tightly grouped behind them. Paajarvi and Belanger are once again on the third tier, the fourth line guys rank lower, and Darcy Hordichuk sits in his own little world playing against the weakest sisters of the NHL.

How Does This Compare To Regular Quality of Competition?

The link to regular Quality of Competition for this same group is here, and we can see that basically the same information is conveyed. Six guys on the team have a number above zero – Horcoff, Hemsky and Smyth lead the way, with Eberle, Hall and Nugent-Hopkins following tightly behind. Jones and Gagner are next, followed by Paajarvi and Belanger, followed by the fourth line, followed by Hordichuk, alone at the bottom by a mile.

Regardless of what metric is used to rank players – time on ice, Corsi, plus/minus, even total points – over a large number of games it all boils down to basically the same thing. The guys who play the most typically have the most points, the best plus/minus, and the best Corsi ratings – all four metrics, over a large population and a large number of games rate player successes. There are minor differences using each system, but they’re all in the ballpark.

The thing I like about Eric’s proposed set-up is the differentiation between top defenders and top forwards. Most conversations about matchups typically focus on matching lines – but the defenders out there matter too.

What Does This Mean For The Oilers?

It means that despite the fact that Tom Renney worked to give Eberle, Hall and Nugent-Hopkins some advantages this past season – lots of starts in the offensive zone, ducking the opposition’s top forward line when possible – they still ended up playing the best defencemen on the other team. Not only that, but they produced.

It also shows why guys like Ryan Smyth, Shawn Horcoff and Ales Hemsky have value, despite being routinely run through the mud by fans. When Renney faced the other team, more often than not those were the guys he put out against the Sedins, Getzlafs and Iginlas of the world.

On the other hand, there’s no relief on this chart for Eric Belanger – he struggled despite seeing the opposition’s depth on most nights. The fourth line, similarly, got destroyed despite playing against terrible players. Also, for those wondering why Darcy Hordichuk played only four and a half minutes per game – this chart shows exactly how much trust Renney had in him.

  • No trust at all. Why even bother with him was Renny’s attitude. As much as he went to the wall with/for Belanger he did exactly the opposite for Hordichuck. The metrics for Belanger proved out that Renny’s assessment and use of him was misguided and detrimental to the teams play.He was death to MP and whoever else he was playing with. Pro scouting either missed the boat on him or he just didn’t fit into the system that Renny had in play. Either way Belanger’s play as contrary to his history. I don’t think he plays anything other than 4th line minutes with PK and defensive zone draws added in. If he produces at last seasons pace it won’t adversely affect the teams play as much as it did last year because the expectations will be lower and his role reduces.

    My main concern is for Jones and Eager. I think that we can’t have both of them in the lineup on opening day. For me Eager either gets traded,sent to OKC or finds himself on waivers. Jones produces but his inability to be physical on a consistent basis makes him expendable IMO. I would rather see Tyler Pitlick or Petrell play on the 4th line night in and night out. I want and need for physicality out of the fourth line.

  • paul wodehouse

    JW – I note that Jones appears twice and RNH does not appear at all on your summation of the Quality of Competition stats. Is it safe to assume that the 6th name should be Nuge?

  • John Chambers

    Jon, this is a fascinating way to evaluate players and I thank you for the post.

    This totally throws water on the argument that RNH and Eberle were eting marshmallow minutes as they were scoring their EV points against Keith, Weber, Bouwmeester, and the like. I personally think QComp is a far more important constraint or advantage than Zonestarts, and Ebs and the Nuge have been unfairly typified as ‘sheltered’.

    The reason I don’t like Zonestarts is that players begin shifts on the fly or with NZ face offs most shifts. But overly weighting Zonestarts, offensive players who are preferred by the coach to start in the offensive zone are penalized using stats like ZS adj relCorsi.

    • On the other hands, the kids were not playing against the best forward lines. I think when you combine middling forward competition plus good zone starts, the word “sheltered” can start being used.

      Personally, I don’t think Eberle was being protected – I think Renney had a plan to get Nugent-Hopkins’ feet wet at the NHL level, and because Eberle was the designated wingman he also got the benefit of the somewhat easier minutes.

      With all of that said, there’s no question about their ability to post offence when we look at who the defenders they were playing against were. The shelter we’re talking about is defensive – very few starts in their own end, not playing against the other top offensive threats.

  • Wanyes bastard child

    Why is it always the coaches fault? When are the fans of Edmonton gonna stop blaming the guys behind the bench for the performance of the players on the ice. Theres only so much you can do with an inexperienced group of players surrounded by marginal veteran players. The same thing with Mactavish, Quinn, and Renney. All three are credible coaches with credible reputations and a history of success.

    When the young group of forwards learn how to win in this league with experience and time are we gonna suddenly praise the coach behind the bench for being around at the right time and place? I sure hope people erect statues of the coach at the time based on his survival.

  • As a stand alone tool this offers great insight into how the players were used and coached against. In order to measure success it needs to be cross referenced with points? I’m just trying to make sure I’m understanding the data accurately

  • John Chambers

    Great work by Eric T to slog through the data and produce this really valuable chart. Great work by you JW to find it, present it to us and add first rate analysis. This puts the evaluation of the performances of individual forwards on a much solider footing.

    This kind of statistical analysis does not answer every question but any organization that is not relying on this kind of information is really missing the boat.

    • Thank you for the hard work and for pointing out the absolute value to every NHL team of statistical evaluations—how these stats are used is system dependant but their value is core to opposition tactical system evaluation in my NHS.

  • John Chambers

    I think it was a tie with all the guys they played with because it takes 6 man units to execute a system properly and this is micro-managed and assesed on a shift by shift basis,I dont see the value in useing statistics to define either the opposition or the Oilers,because every system in the world is executed by 6 men on a shift by shift basis,teams can play a different system for each line if they can manage it.

    Because of this ,any statistics that do not include exactly the same players as a 6 man unit cannot be used as a baseline for anything at all,never mind assesing individual players or their system suitability or value.

    You did find someone who sees no value in statistics regarding players assesments of any type.

    Statistics are however very valuable for predicateing opposition systems ,by identifying trending individual opposition players numbers,or a lack of numbers,or a disporportionate number of a specific type of numbers.We CANT tell how good they are as individuals though—and it doesnt matter because in a well executed system players are never in one-on-one dynamics long enough to taste the venom.

    And the sweetest part of all is that my NHS is invulnerable to this “system radar”via statistical analysis and trend identification.Because of how the offense is integrated into the base of the system.

    I saw 6 man units last year that were executeing something pretty powerful when they were clicking,some units WERE executeing an effective system but almost in a renegade fashion.It wasnt really a 4 line buy-in every night.Except for our early streak,that was all system induced through our focus on three transitions and puck possesion for several reason,none of which were !00% NHS,that committment came after we re-grouped from the free-fall and began to regain our balance very late in the season.The results from the early possesion/transition style and the later NHS focus were the same essentially,sucess and wins.It is an excellent comparative example from last season,and it is a PERFECT illustration of a hybrid adjusted with NHS data{includeing offensive integration data to some individual players}purer NHS core values–the first win streak,and a NHS adjusted hybrid like LA used to win the cup—our last streak where we were righting the ship useing EXACTLY the same NHS adjusted hybrid that won the cup{a lees NHS based system,more hybrid }—-and we even began to implement it at the same time as LA with around 25 games to go when LA also got serious.

    So no really the stats helped identify opposition systems and ,system strengths projected by individual players but in my considerations I only included tactical system engagement attention,and I use the stats a lot to plan tactics sometimes just before games to stay current on trends caused by injury replacments and other system adjustments,it can be that accurate as to identify very recent trends.So the stats really served me no purpose in identifying how good the opposition players or any individual player really was,just helped ID tactical system trends the opposition was experienceing pror to gametime.

      • John Chambers

        Of course,its relevant to the thread,do you have any questions?there are a lot of areas for analysis and challenge technically when i refer the the NHS,why not take a chance and present a challengeing tactical question about my data as it applies to the thread.This is a great thread.I aready explained a tactical application of conventional stats that exposes opposition systems more than individual when executed within the NHS,and questions about that?

  • master of my domain

    off topic- the Nuge and his family just had dinner at the restaurant where I work in New Westminster. he looks like he’s packed on some weight this summer, in a good way too.

  • Reg Dunlop

    First, I would like to thank NewAge for the abridged version. I still fell asleep though.

    Second, is it not possible that sometimes playing against a forward line of defensive specialists,a line that doesn’t get a lot of ice time otherwise,could skew results meaning Hall and Ebs actually faced forward lines that were harder to accumulate points against? It could even indicate that other teams were trying to shut down Hordichuk. I think back to Gretz always facing Kasper in Boston or Carroll in Long Island. Facing top defence pairs is hard to avoid when some teams have 30 minute men like Weber and JBow. They seem to never leave the ice.

    Third, I want to share with you my own newagesystem, guaranteed to bring success. SCORE MORE THAN THE OTHER TEAM.

  • paul wodehouse

    4Tiger…I’m not confused or at least i don’t think i am…imo Renney got kinda railroaded and i think Krueger is being brought in here as some sort of ‘savior’ because some say Renney didn’t do a good job with the kids or for that matter with the whole bunch that our GM gave him to work with…i’m not an advanced stats guy AT ALL but without taking my shoes and socks off i can still add two and two … Renney did a great job with this sample of raw talent…sure there were times when he made you scratch your melon but if he ‘protected’ players it was a logicical thing to do as a coach and it wasn’t his fault that PRV went into an historic sophomore slump like no other as he did…apart from other factors that led to 91s’ demotion but i’m thinkinmg it was Renney who said the AHL was going to be the best place for 91 without getting the ‘he gave up on him’ retort in return…Renneys’ strength IS working with kids, didn’t he take a couple of his junior teams to win a Memorial Cup or two in his coaching career?…at the NHL level he has shown via the charts and graphs of the world that he did a credible job but got run in the end…i think Renney was brilliant ‘protecting’ 4 & 14 from opposing teams best forwards instead of’sheltering’them from opposing teams best Dmen…add 93 at times and y’have a beauty 3on2 don’tcha…!? Like I said i’m not a stats guy but didn’t Renney do what the numbers said would be the best way to utilize all the young talent at his disposal without threatening early career destruction? OR
    am i confused and i don’t even know it?

  • OB1 Team Yakopov - F.S.T.N.F

    I think some are getting a little confused about this data, the fact that Eberle/RNH played top defenders doesn’t mean they weren’t shelterd… sure it made scoring goals harder, but they didn’t have the added burden of playing against top forwards, who would put the defensive pressure on against them.