College Basketball Bubble Watch

Updated: February 25, 2014, 11:10 AM ET
By Eamonn Brennan | ESPN.com


When it comes to the 'eye test,' trust, but verify

Editor's note: This file has been updated to include all games through Monday, Feb. 24.

In 2011, Rockstar Games published "L.A. Noire," an action-adventure game set in postwar Los Angeles. Surprisingly removed from the open-world satire of the "Grand Theft Auto" franchise, "Noire" instead guided players from one linear crime scene to the next, where, as complex gumshoe archetype Cole Phelps, they searched for clues and tracked down suspects for one-on-one interviews.

The game's main selling point was the sophistication of its characters' facial expressions. Thanks to some fancy newfangled camera technology not worth detailing here, professional actors were recorded, and their hyper-detailed facial expressions were faithfully recreated as animations.

This was the core game mechanic: Phelps sat across from suspects and asked them questions. Based on how the suspects responded -- whether their eyes moved left or right, or if they looked nervous or sad -- the player then chose the next conversation option: "lie," "doubt," or "truth." At the end of each case, based on their responses, the game assigned players a score.

The tech of "L.A. Noire" was stunning, audacious and fundamentally broken. Too often, characters' expressions were too hard to read, or unintentionally misleading, or just plain uncanny. You'd think a character was lying, and then Phelps would shout an accusation -- and then lower his voice jarringly when the thread went nowhere. It was a noble failure, but a failure all the same.

It was also, if the Watch may say so, a fitting metaphor for the reliability of the "eye test."

Every year around this time, we start hearing about the "eye test," or what the members of the NCAA men's basketball selection committee see when they watch teams play.

Late last week, Sports Illustrated's Seth Davis briefly brought it back to the fore by tweeting: "People underestimate how much the 'eye test' matters to people on the committee. Those folks watch a LOT of games. They trust what they see."

There are a handful of ways to unpack that statement. The first is as a helpful reminder, same as when committee chair Ron Wellman introduced the concept of "schedule intent" in a teleconference two weeks ago: The bracket is predictable, but only up to a point, because the selection committee is nothing more than a group made up of human beings, each with their own set of biases and opinions.

These opinions differ not just about teams but about what criteria should be used to judge the merits of those teams in the first place. Some will hew to the RPI. Some will mix in advanced metrics. And some -- maybe most? -- will trust their eyes. How can we be sure?

The larger problem, of course, is the reliability of an eye test in the first place. The data our eyes pass to our brain is full of noise -- emotional reactions, misplaced memories, tunneled viewpoints.

Let's just admit it: It is impossible for a human being to watch so much basketball (and believe the Watch when it says this, for it has tried) that he or she can properly contextualize so many teams accurately. That's why advanced metrics systems exist. It's why assistant coaches spend their lives hunched over film. It's why the RPI, for all its flaws, was invented in the first place: to minimize the unfairness of subjective vision.

The goal of "L.A. Noire" was to turn the vague overconfidence of a gut feeling into a scoreable game experience.

This person is lying. I can see it in their eyes.

For 30 years, the selection committee has gotten better and better at limiting this impulse, with ever-more-inclusive information at its disposal.

Maybe we're underrating the modern belief in the eye test, but the Watch hopes not.

Our eyes are notoriously difficult to score.

Note: All RPI data via ESPN RPI is updated through Feb. 24.