Steer: On cognitive bias, and avoiding harmful narratives in hockey

Follow me on Twitter

An unlucky redirection on the first. A virtually unstoppable Evgeni Malkin howitzer on the second. A costly misjudgement by Reto Berra on the third, and suddenly the Colorado Avalanche found themselves chasing the game in Pittsburgh. You certainly don’t often see a minor penalty influence the outcome of a game to such an extent, but on occasion that’s just how the cookie crumbles.

The ensuing narrative came easily to most, embraced for its familiarity, perpetuated widely as though it were a form of Colorado sports gospel handed down by Joe Sakic himself. “The Avalanche are mentally fragile” was again a popular refrain, discounting the fact that the ebbs and flows of psychological momentum impact all teams mostly equally. Ignoring the reality that a hockey game isn’t a single variable event in which team x’s play is the sole factor that determines the game’s outcome.

Your team of choice isn’t more or less focused, or determined, or driven than another, generally speaking. You’re just more aware of the peaks and valleys in the quality of their play, the variance in their output. This cognitive bias often results in inferences about people and situations being drawn in an illogical fashion, leading to the construction and propagation of false narratives.

False narratives are all too common in sports, and are particularly common in hockey, a game where many of the factors that contribute to the success of a team have long been considered indefinable, or rather, intangible. Often, this has resulted in the undue attribution of importance to some player traits, and a blatant disregard for others. Similarly, humans are inherently driven by their emotions and experiences, and often unknowingly distort reality to better suit their own preconceived notions.

In this piece, I will examine six common cognitive biases and an Avalanche-related narrative that they inform, in an attempt to show the manner in which narratives are often constructed and disseminated. What you see isn’t always what you get.

Fundamental attribution error

What is it?

Fundamental attribution error is the tendency for people to place an undue emphasis on internal characteristics when explaining someone’s behaviour in a given situation, rather than considering the situation’s external factors.

The harmful narrative

“Gabriel Landeskog needs to step it up. Given that he’s captain, I expect him to be producing at rate x, and contributing y specifically.”

What’s the real story?

Leadership is typically a latent quality in hockey. Certainly you have famous instances of strong, public leadership, such as Messier’s Game 6 victory prediction in 1994, but these are exceptions, rather than the norm. For the most part, a captain’s impact on a team is something that’s only truly understood by the team itself. Media and fans aren’t privy to the inner workings of the captain-team relationship, and drawing conclusions based on incomplete information is a questionable practice at best. As Kevin Durant said so eloquently;

“You guys really don’t know [expletive]… To be honest, man, I’m only here talking to y’all because I have to. So I really don’t care. Y’all not my friends. You’re going to write what you want to write. You’re going to love us one day and hate us the next. That’s a part of it. So I just learn how to deal with y’all.”

Truly, the media doesn’t really know “expletive” about the true dynamic of the dressing room, and the concept that your internal characteristics contribute more heavily to your play if you’re the captain is extremely faulty. An athlete’s performance is largely governed by how talented they are, full stop. If Landeskog isn’t producing, criticize him because he’s playing at a level that isn’t acceptable for someone with his abilities, not because he’s got a damn letter on his jersey. Sometimes, I feel as though people forget that wearing the “C” doesn’t bestow super powers upon the recipient. It’s the man who makes the sweater, not the sweater who makes the man.

Confirmation bias

What is it?

Confirmation bias is the the tendency to search for, interpret, favor, and recall information in a way that confirms one’s beliefs or hypotheses, while giving disproportionately less consideration to alternative possibilities.

The harmful narrative

“Colorado is a perpetual joke. Roy is too ignorant and unintelligent to alter his tactics and will never learn from his mistakes, meaning that the Avalanche will be forever mired in mediocrity. I would also like to lump Joe Sakic in with Roy in this regard, because I can specifically recall several bad deals he made in July 2014, and am unwilling to acknowledge his successes”

What’s the real story?

Let’s be crystal clear here. This is not me absolving Patrick Roy and Joe Sakic of any wrongdoing, not in the slightest. Both have made mistakes, and Roy’s tactics and rhetoric are often extremely flawed.

https://twitter.com/lukeasteer/status/668876911536291840

Colorado is not the worst team in the NHL, despite their abysmal start. By virtue of expected goals they rank 23rd, and trail divisional rivals Chicago and Minnesota by only the slightest of margins. They’re 25th in SCF% and last in CF%, yes, but they’re rapidly climbing. So, what gives? Why are the Avalanche most often subject to more criticism than any other bottom-tier team?

https://twitter.com/GarretHohl/status/668921135841415168

Is this perspective partially a product of the vindication the analytics community felt after the unsustainable Avs fell from their tenuous perch atop the Central? Do people simply dislike Patrick Roy, and hope to see him fail? At times, it seems as though some members of the media are personally offended by the Avs’ front office, their criticism bordering on vitriol. Poor performances are held up as an indictment of the franchise, while strong performances are largely ignored. Roy is lambasted for his opinions on shot attempts, while the impact of Chris MacFarland is rarely discussed. Sakic is lumped in with Roy, the distinction between their roles blurred, his successes forgotten and failures put on display.

It’s not that I feel Colorado should go uncriticized. This is quite obviously a franchise whose front office has made a number of missteps, especially during the early portion of their tenure. The Brad Stuart trade will probably keep me up at night for years. However, I feel that analysis and judgement should be unemotional and objective, and it often hasn’t been.

Perhaps Colorado will continue to be mediocre, it’s certainly possible. Some would even say probable. Regardless, that doesn’t mean that every step is a misstep, every move the wrong one. At this point, writing a gleeful account of the Avs’ woes is akin the plucking the juiciest plum from the tree of low-hanging fruit. It’s lazy journalism.

Self-serving bias

What is it?

A self-serving bias is any cognitive or perceptual process that is distorted by the tendency to perceive oneself in an overly favorable manner. It may also manifest itself as a tendency for people to evaluate ambiguous information in a way beneficial to their interests.

The harmful narrative

“Radulov has yet to sign a new deal with CSKA Moscow. As such, I’m now fully expecting Radulov to join the Avalanche next summer and will be disgruntled (with management) if he does not”

What’s the real story?

Radulov might sign with Colorado this summer. He’s got a positive history with Patrick Roy, and these murmurings obviously have some basis in reality. But he likely won’t. If you’ve been slotting him into your prospective 2016-2017 Avalanche line-up you should probably give it a rest.

Belief bias

What is it?

Belief bias is the tendency to judge the strength of arguments based on the plausibility of their conclusion, rather than how strongly they support that conclusion. Or, similarly, it refers to when one’s evaluation of the logical strength of an argument is biased by their belief in the truth or falsity of the conclusion.

The harmful narrative

“Tyson Barrie is a good point producer, but is poor defensively. Travis Hamonic is elite defensively, and I tend to value two-way guys more heavily. Honestly, I would do Barrie + 1st to get the deal done.”

Similarly:

Dude i got first hand proof. Ive played with Barrie before and I know Hamonic is a more valuable asset. @lukeasteer @TGS42084

— DJ Mika (@eflarrow) November 20, 2015

What’s the real story?

First of all, Tyson Barrie’s production isn’t merely good, it’s borderline elite. In 2014-2015, Barrie averaged the most P/60 at 5v5 of any defenseman who played at least 200 minutes, eclipsing both noted superstars like Erik Karlsson and P.K. Subban, and promising up-and-comers such as John Carlson and John Klingberg.

Secondly, what does “poor defensively” truly mean? Does that mean Barrie allows a significant number of shots against? A significant number of scoring chances? Moreover, what is information is truly relevant when assessing a player’s defensive impact? Let’s take a look at what Jack Han had to say about evaluating defensemen over at Hockey Graphs.

Coaches and GMs want guys who block shots, don’t turn the puck over and can play good D-zone coverage. That makes sense because their focus is on having people who can make sacrifices in the name of the team, manage risk, and prevent the opponents from creating high-quality scoring chances.

But those things do not matter to me nearly as much as two other things, which I believe to be integral to a defenseman’s job requirement:

• The ability to support the puck across the two blue lines
• The ability to prevent the opposition from supporting the puck across the two blue lines

As Han notes, many people are still very hung up on the concept of the defensive defensemen, the value of a player who can limit high-danger chances, block attempts, and clear opposing players from the front of the net. However, having the opportunity to excel at in-zone defensive play means that you’ve already ceded the blue-line to the opposition, which is a failure in and of itself. What’s truly valuable is an aptitude for controlling transitional play, and ability to create controlled breakouts, prevent controlled entries and maintain possession in the offensive zone. To paraphrase Han, being able to do these things eventually leads to a positive shot differential, a positive goal differential, and, further down the road, a positive win differential.

Tyson Barrie is better at both generating and suppressing shots than Travis Hamonic. In 2014-2015, Barrie’s CF% RelTM was a sterling 4.5, which was the 17th best mark in the NHL. Conversely, Hamonic’s CF% RelTM was -2.5, good for 181st in the league. Yes, Barrie does make the occasional costly mistake in the defensive zone that will leave you tearing your hair out. However, a successful outcome in hockey is one in which you’ve scored more goals than the opposition, and Barrie’s net impact on this outcome is greater than Hamonic’s. To simply state that “Barrie isn’t good defensively” because you’ve seen him turn the puck over once or twice is grade-A baloney.

Framing

What is it?

In the context of mass-media communication, a frame defines the packaging of an element of rhetoric in such a way as to encourage certain interpretations, and to discourage others. This is often done by utilizing a narrow description of the situation or issue.

The narrative

The Colorado Avalanche are really missing Gabriel Landeskog, the captain who abandoned them to make a cowardly and selfish cheap shot

— J.J. (@JJfromKansas) November 18, 2015

What’s the real story?

Gabriel Landeskog plays a physical game. He always has, likely always will. This involves him playing with a bit of an edge, which doesn’t always endear him to other fans and players. On the hit in question, Landeskog set Brad Marchand in his sights as he moved laterally across the top of the defensive zone, then visibly slowed when he realized that Marchand couldn’t see him. He made shoulder-to-shoulder contact, but hit Marchand’s head as he followed through.

Colorado's Gabriel Landeskog has been suspended 2 games for an illegal check to the head of Boston's Brad Marchand. https://t.co/e0PHYMclPM

— NHL Player Safety (@NHLPlayerSafety) November 13, 2015

Personally, I was fine with Landeskog getting two games for this hit. Whether that punishment is consistent with what other players have gotten for similar hits is another story. All that said, framing it as “cowardly and selfish” is irresponsible.

Hindsight bias

What is it?

Hindsight bias is the inclination, after an event has occurred, to see the event as having been predictable, despite there having been little or no objective basis for predicting it.

The harmful narrative

“Oh man, we should’ve really traded Paul Stastny at the deadline in 2014. Losing him for nothing was an enormous blow. Way to go, Sakic”

What’s the real story?

This is revisionist history at its finest. Name one instance in which a GM has traded their first line centre at the deadline when they were sitting in a playoff spot, with the opportunity to make a push for the division title. That’s an extremely questionable front office move; you’re essentially giving your roster the middle finger, and letting them know that you don’t think they have what it takes to capitalize on their current success. With a young roster like the Avalanche had at the time, that’s a dangerous precedent to be setting.

I’m strongly in favour of good asset management, believe me. But there’s also a human aspect to sport that shouldn’t be discounted. Furthermore, Stastny did indicate to Sakic that he’d be willing to accept a hometown discount to stay in Denver. He eventually decided not to, and left for St. Louis. These things happen, let’s move on.

Like it or not, both the media and online communities play a large role in influencing our perception of the sports that we watch, and as such are obligated to disseminate content (and associated narratives) responsibly. Always take the time to critically assess the things that you read. Don’t share lazy attempts at journalism. We’ll all be better off for it.