I have been playing small stakes poker for a few years now, with the original intent to study the cues and tells that come from bluffing to see if I could translate generic observations from the poker table into real world deception detection.
Reading about a police interview technique developed by Professor R.E Geiselman of UCLA, an expert in detecting lies, gave me my first clue:
"When asked if they want to add anything, deceptive people tend to say NO quickly whereas truthful people either go ahead and add something new or they at least think about it before saying NO."
It occurred to me that a speeding up and slowing down behaviour translated directly into a specific bluffing scenario that went like this:
1 -- If you are strong, but bluffing to be weak, what do you do? You look at your cards and then your chips and pretend to think whether you are going to increase your bet. You stall.
2 -- If you are weak but bluffing to be strong, what do you do? You don't hesitate, you move your chips out quickly.
This poker table analogy holds directly in a real world scenario: People who move or act too quickly (quicker than their baseline!) are potentially deceptive or lacking confidence. It's a red flag moment at the table and it's a red flag moment in the real world too. Off course, this means you need a baseline, a situation where some small talk has taken place prior and where you have had an opportunity to gauge behaviour.
Something else I noted on the poker table was that people who I thought were bluffing seemed to be slightly more friendly, or polite at that point, as if not wanting to antagonise other players or draw attention to themselves.
Frank Enos from Columbia University in his thesis says:
"Preliminary findings suggested that pleasantness is the most promising factor in predicting deception..."
As expected, there is an overlap from poker table to real world scenarios making it a great laboratory to study human behaviour, it's just a matter of paying attention.