Recency Bias in Politics and Poker

Was sitting here this evening with the teevee on hearing the talking heads reacting to early results of the New Hampshire primaries. Most polls had Donald Trump winning big on the Republicans’ side and Bernie Sanders similarly coming out on top for the Democrats, and that is just how things have played out.

Many of the commentators nonetheless seem to be expressing some amazement at the twin triumphs of Trump and Sanders. To be fair, they’re prefacing such comments with statements about how a few months ago it would seem unlikely that either of the two “outsider” candidates (a term that somewhat differently applies to each) would not just win in NH but by large margins. Even so, most seem to be influenced more by the fact that neither won in Iowa (although both nearly did), thus making the overwhelming wins in NH seem more dramatic by contrast.

Now I’m watching the speeches of the winners and other candidates. Man, they are going on and on. Couldn’t just say “nh” in NH, and move on, I guess.

(By the way, I’m assuming someone has already conceived of an op-ed comparing Donald Trump and Cam Newton and how each handles winning and losing, probably mapping those responses onto some conclusion about the relative maturity level of the culture as a whole.)

Thinking ahead, what happened tonight will surely inordinately affect the response on February 20 when the next round of voting occurs with the South Carolina primary and caucuses in Nevada and Washington state. The focus on what just happened is partly influenced by the 24-hour news “cycle” (which isn’t really even a “cycle” anymore but rather a kind of perpetuum mobile running without interruption). But it’s also, I think, just “part of the game.”

Many poker players are well aware of the idea of “recency bias” and that tendency to think things in the recent past are reliable indicators of what is about to happen next. And the fact is, those indicators often are fairly reliable, but not always -- another reason for the resulting bias.

Trying to pick winners each week during the NFL season highlights the problems that can arise from a too great reliance on recent events. Heck, the huge point spread in favor of the Panthers in Super Bowl 50 and all of the betting on Carolina -- more than 70% of the bets leading into Sunday, I read -- exemplified recency bias especially well, as everyone was influenced by their trouncing of Arizona in the NFC Championship.

The presidential race will continue to experience turbulence for the next few weeks, I imagine, but by the end of March more than half the delegates will have been determined for each party, thus making what happened most recently relatively less influential.

But what just happened will still affect responses to what happens next, as it always does.

Image: “New Hampshire Primary - Illustration,” DonkeyHotey. CC BY 2.0.