clock menu more-arrow no yes mobile

Filed under:

Groupthink, Confirmation Bias, Hockey Fans And Microstats

One of my first stops every morning is Economist Robin Hanson's blog Overcoming Bias. In Hanson's own words, the subject matter is "...honesty, signaling, disagreement, forecasting, and the far future." Hanson's depth of posting is impressive -- he discusses psychology, philosophy, economics, bias, fallacy, markets, science and the human condition. One of his posts from last year has stuck with me for awhile now as it applies to most things we discuss on a daily basis. The post, "Who Will Fight Group Think", does not discuss hockey, but it has an over-reaching message that applies to most of the subjects written about here and elsewhere throughout the Oilogosphere.

Hanson quotes Nickolas Wade from TierneyLab discussing Groupthink:

Conformity and group-think are attitudes of particular danger in science, an endeavor that is inherently revolutionary because progress often depends on overturning established wisdom

Vic Ferrari has talked about Roger Neilson's impact on coaching and the game in general. Neilson was an innovator that turned the game on it's ear by tracking all of that stuff that was going on between goals and penalties. While some, like the dashing Gabriel Desjardins of Behind The Net Hockey, argue that "hockey teams, by and large, do know what they are doing" (I have my doubts), hockey fans, on the other hand, are generally a clueless lot, especially when it comes to the inner-workings of the game.

Most fans don't have access to the video libraries that NHL teams do, and even the fans that have DVR technology don't use it. Fans, by and large, rely on what they see, and make snap decisions based on short bursts of data. From the time that decision is made, a fan will see what they expect to see, that is, they begin to notice the events and data that confirm the observations that led to their conclusion. They begin to seek new information to confirm their pre-existing bias, subconsciously ignoring the entire data set, especially the pieces that disagree with their conclusion. In psychology and cognitive science, this is known as confirmation bias.

In the world of sports fans, confirmation biases abound. It's impossible for individual fans to record, catalog, process, analyze and interpret the results of hundreds of independent events occurring constantly throughout a game, but it's much easier to pick out those events and sequences of events that support their conclusions. Any hockey fan that has sat silently shaking their head while the crowd piles on an undeserving player recognizes this immediately. It's a powerful psychological force, especially in a setting like sports. Fans can confirm their biases for themselves and immediately fall back on thousands, sometimes hundreds of thousands of fellow fans to confirm what they already know. This is the very foundation of groupthink:

Groupthink is a psychological phenomenon that can occur in groups of people. Rather than critically evaluating information, the group members begin to form quick opinions that match the group consensus.

Because it's impossible to critically evaluate those hundreds of independent events, fans end up rallying together in support of or against the lowest common denominator in observational terms. And this is why statistics and the analysis of those statistics is crucial. Statistics, at the lowest level, are simply a vast collection of events. Even though the fans' understanding and analysis of the game continues to grow (thanks to people like Vic, Gabe, JLikens and Tyler), there remains a vast majority of fans, people who know better and simple-minded folks that wallow in ignorance, sometimes willfully.

Roger Neilson realized the observational shortcomings of hockey and moved to correct it. Twenty years later we're beginning to advance the fan understanding of the game in the same way, and the biggest obstruction to the effort is groupthink. However, as Hanson notes, it's not enough to simply tell the group that they are wrong:

But after 37 years, the group think idea is pretty well known. The problem is that simply knowing that the group might be wrong is very different from knowing where in particular they are wrong.

Schremp_medium The "where in particular they are wrong" part is what leads to such concise breakdowns in the math and such detailed analysis of players, games and situations. Why was Rob Schremp a bust? Tyler found the answer in the math. Last year, even though Shawn Horcoff was being slagged on by the fans, he remained highly-effective. What was behind that? Vic found it in the math. There seems to be a common complaint among those that rail against statistics as a primary tool of analysis in hockey - that statistics can be used to reach any conclusion and that the stats are being manipulated. However, it's quite the contrary. With each of these posts, the author, be it JLikens, Tyler, Jonathan Willis, Vic, Scott or myself, explains the methodology he used to do his analysis and is open to debate and questions at any time. Contrast that with the opposing argument: "it is so because I saw it, and my observations are correct." In the world of logic, that's called Begging The Question. There is no way to prove it is so, there is no way to test the hypothesis, the argument is based solely on the confirmation bias of the person making the point. Which is more scientifically valuable? Which is more valuable to advancing the understanding of the game? Which is more predictive?

One other groupthink complaint towards the "statheads" is that they "hide behind the math" and "band together" in a form of their own groupthink. Someone that uses the math of hockey to set make a point or write an article isn't hiding behind anything, because from above, their methodologies are out in the open for public consumption. And if "banding together" were true, then how do you explain something like the public opinion of Dustin Penner? I didn't set out to prove the microstats guys wrong when I penned the Amicus Brief In Support Of Dustin Penner, I wrote it because I was hearing through the media and reading on the internet from even the "mathheads" like RiversQ, Tyler, Vic and Jonathan what a terrible player he was, which contradicted everything my eye was seeing. I argued the point at Lowetide's place for a long time before giving up. The groupthink was too strong. So I started digging through the underlying stats and found that he was a microstat star, and that he was, in fact, an extremely valuable player. Nine months later and among that group, the opinion on Penner is still split.

Hanson also notes another problem in fighting groupthink:

Far more people like to complain about groupthink when they think their contrarian ideas neglected,

I had this complaint about Penner and overcame the bias by doing the work behind the counterargument. David Staples at The Cult Of Hockey has been railing against Corsi's usefulness for quite some time, and while I disagree with him, and the math, in many different forms, has thus far agreed that he's barking up the wrong tree, Corsi is the strongest indicator of possession, zone time and winning. But David attempts to do the work behind his argument rather than railing against the groupthink argument, and I give him credit for that - it's more than most people are willing to do.

What we do here, what JLikens does at Objective NHL, what Tyler does at MC79hockey, what Vic and Rivers do at Irreverent Oil Fans, what the Contrarian Goaltender does at Brodeur Is A Fraud isn't new. Roger Neilson was doing the same thing twenty years ago and Kevin Constantine was doing it ten years ago. What these people do is attempt to bring a deeper understanding to the game we all love. Unlike a methodology that diminishes or outright ignores the value of statistics in establishing a point of view on a player, team or game, a methodology that values statistical analysis creates a point of view more open to being molded or challenged based on logical arguments from an established set of data. The "saw-him-good" crowd makes these judgments in the same way, but without tracking the information on which they base their opinions. Sometimes (even most times) these observations will be correct, but when a mistake is made, there is almost no way for an outside observer to effectively challenge what someone has seen with his own eyes, especially since, most times, he's not open to having that opinion truly challenged. When worse comes to worse, the fall-back becomes an individual data-set that no one can challenge instead of the group data-set that helps members in the discussion to become - at least a little bit - more objective. Statisticians make mistakes like anyone else, but when you have the numbers and an established methodology staring you in the face it's a lot easier to correct them.