A California based start-up has developed a software program that notices  – and reads – facial micro-reactions. Does this spell the end of bluffing in poker? Or could scammers exploit their opponents with it?

The End of Bluffing in Poker?

Reading skills are maybe the most highly valued skills in poker. Top players like Phil Hellmuth can win more and lose less than others because they can pick up on the smallest changes in their opponents’ facial expression or posture.

Listen to what Hellmuth himself has to say about the significance of reading people well:

Watch him perform some of his “White Magic” here:

The essence is this: If you want to be a great poker player you better understand what the guy on the other side of the table is thinking.

Impossible to Trick

If it’s possible to automate these skills and have them performed by a computer this should give you an unbeatable edge over anyone, right?

Ken Denman.

Six former PhD students at the University of San Diego might have done just that. With their company Emotient, recognized as one of the 2015 TIE Top 50 Start-Up companies, they’ve developed a software program primarily to be used in analyzing response to TV advertising.

The software works at a rate of 30 frames per second so it won’t miss even the smallest of responses. According to the inventor the software works so well that it’s impossible to trick it.

Says CEO Ken Denman: “There are two pathways in the brain that drive facial movement. You can move your face deliberately … and then there’s another pathway … the emotional center.

“If you try to have a poker face, often what you get is micro-expressions.”

95% Accuracy

These can be so small a person wouldn’t notice but Emotient would still pick them up — just like Tim Roth did in Lie to Me.

The company promises 95% accuracy independent of age, gender and ethnicity of the test subject. Ethnicity is really of minor importance here, as facial expressions of primary feelings seem to be the same everywhere from highly industrialized countries to remote tribes.

There have been several experiments where the software has proven to work. For example it was used at the first TV debate for the Republican candidates running for President.

This was the first time face recognition software has been used in this context. You’ll probably not be surprised to find out Donald Trump was predominantly expressing “anger”.

The efficiency of the software has been generally accepted. Writes Charlene Weisler on mediavillage.com:

Emotient’s IP has the ability to capture hundreds of facial expressions at one time in a non-intrusive manner [and] offers the capability to amass big data by individuals’ gender and age. Emotient can track live responses at conferences, sporting events, concerts and business meetings.

So, obviously, it could also be utilized in poker tournaments.

Seven Basic, Universal Emotions

Just imagine what an edge it would give you if you knew how your opponent feels. You wouldn’t know what cards he has but that wouldn’t even be necessary.

Emotient distinguishes between seven basic, universal emotions (and about 20 additional face movements): Joy, disgust, anger, contempt, surprise, sadness and fear.

All of these give you a pretty clear indication of how strong your opponent’s hand is and how it evolves from street to street.

The technology could easily be used in televised events like the EPT where hole cards are not shown but the action is shown in real time. An accomplice could sit at home in front of his monitor, run the software and tell one player at the table what his opponent is feeling.

If you had a device similar to Google Glass, you wouldn’t even need that accomplice anymore. Or think of online poker with webcams. The possibilities are even more numerous because you could constantly survey the emotional states of each of your opponents at every table!

Maybe you personally think you would never cheat in such a perfidious way but you sure know that there will always be people with less scruples. Is it then time to say goodbye to the unreadable poker face? Is this software ultimately going to erode the game?

Not so fast.

The Limits of Software

Let’s say Emotient is as good as the experts say. But it suddenly becomes unusable if you don’t feed it the data it needs.

According to Emotient support the following things will cause the software to fail:

• Hands
• Large hats with wide brims
• Heavy facial hair
• Thick-rimmed glasses
• Dark glasses, especially sun glasses

Remember Chris Ferguson? With his hands folded in front of his face, fully bearded, with black sunglasses and his large black hat, he would be an enigma to Emotient just as much as to me and you.

If you’re following this year’s WSOP main event coverage you might also have noticed how many players are covering their faces with their hoodies or other parts of their clothes.

You might not be able to conceal micro-reactions of your face, as these can’t be willingly controlled, but you can conceal your face.

Televised poker events could be shown with a delay or a ban on all electronic devices. And you could just switch off your webcam.

For now facial recognition software doesn’t pose a threat to any form of poker. And as Emotient CEO Ken Denman told PokerListings:

“Poker is not a focus area for our business. Lots of things the technology can be used on but not a product for that area currently.”

Time to Be Suspicious

So while even the expert in the field says facial recognition software doesn’t have any significance for poker don’t overlook the last word in his statement – “currently.”

Google has already developed text recognition software that works in real time and converts images of text into actual text formats. Optical character recognition (OCR) has been available for mobile phones for several years.

It’s only a matter of time until they make the step to face recognition. But for now you don’t need to worry now about having your facial impressions electronically analyzed.

The night someone starts pointing his mobile phone at you at the poker table, though, it’s time to be suspicious.