How to Call Bluffs in Poker... and Investing

To force opponents to crack at the poker table, try to look trustworthy.


In a recent study by researchers at Harvard, Cal Tech and M.I.T., versed but novice poker players were placed in front of computers and shown hands of cards from Texas Hold 'Em, a simple but strategy-filled variation of the game. Information was sparse. Players could see only their starting cards, the amounts wagered and their opponents' faces. The faces were selected from an online database, and each was presented in its original state or a digitally altered one designed to appear more trustworthy (relaxed smiles) or untrustworthy (tense frowns).

Players could only call or fold, which is to say, stake their chips on a win or not. The researchers found that players took significantly longer and made more mistakes (like folding when they shouldn't have) against smiling opponents. In other words, show up at your next poker night with plenty of cheer and you might leave with a bit of extra cash. Just try not to be intimidated by the grinning gambler next to you.

If only there were a way for stock market investors to call the bluffs of company managers who smile broadly when discussing quarterly financial results, only to retract the numbers months or years later and replace them with sour ones.

A new study suggests a way to give it a shot.

Researchers from Stanford University analyzed words and phrases used during the question-and-answer portion of earnings conference calls – the unrehearsed chat with analysts that typically follows managers' carefully prepared remarks.

Prior research had already linked some written accounting clues to negative financial outcomes. For example, when a company's paper earnings (which are closely watched by Wall Street) far exceed its cash earnings (which aren't as widely followed) over many quarters, it's often a sign that paper earnings are due for a plunge.

The Stanford researchers compared Q&A transcripts from FactSet, a data merchant, with a list of companies that gave dodgy numbers, provided by Glass Lewis, a corporate governance watchdog. They found some interesting correlations. Deceptive bosses were less likely to say that conditions were merely "nice" and more likely to say they were "fantastic;” they rarely mentioned anything "awful." They didn't like to talk about "I" and "we," preferring instead to point to "anybody," everybody" and "nobody." Liars steered clear of phrases like "value for investors" and often referred to general knowledge that "investors know well." They didn't express much certainty ("always"), but nor did they think things over with an "um" or "ah."

The results are perhaps best taken as a diversion rather than a hard-and-fast stock-picking strategy. Anybody can see that, as investors know well.