The Traitors has become essential viewing for millions of fans. The TV game show, in which a group of “faithfuls” try to identify a smaller number of hidden traitors, is all about truth and lies.
And it looks very easy from the sofa. For the 2026 celebrity version, contestants including singer James Blunt, mathematician Hannah Fry, broadcaster Amol Rajan and actors Bella Ramsey and Richard E. Grant, will have been working on their game plan.
But inside the castle, things are more complicated than they might look. This is because the entire game is built in a way that makes any discussion among the contestants essentially useless.
Economic theory suggests that for all of the players, talk is cheap. Nothing they say is binding. Nothing they say can be checked. And if a contestant has any reason at all to deceive the person they’re speaking to, their words simply cannot be relied upon.
That does not mean the traitors say nothing, when they are holding forth at the round table discussion for example. It just means that everything they say is also something a faithful might say.
Traitors can ask a clever question, float a theory, look uneasy, defend themselves with feeling. A faithful member of the group can do exactly the same things. So listening to what anyone says is essentially futile.
Research has also shown that spotting liars is much more difficult than we might think. Across more than 200 studies, people successfully identify deception about 54% of the time.
Flipping a coin would get you 50%.
Nobody counts
Then there is the arithmetic. At the start of the game, when there might be three traitors among 21 players, any name picked has a one in seven chance of being right. The odds might improve as the numbers shrink, but for almost the whole series, any accusation made without real evidence is still more likely to be wrong than right.
And without real evidence available, we see how people reach for whatever is vivid and noticeable about a contestant’s behaviour. He was defensive. She went quiet. He was too quick to agree.
One study has indeed shown that when we are provided with a colourful description next to a boring statistic, the description wins and the statistic is ignored
For that research, participants were told that a group of 100 professionals contained either 70 engineers and 30 lawyers, or 30 engineers and 70 lawyers. They were then given a short description of one of the 100, who was said to be careful, uninterested in politics and fond of mathematical puzzles.
Most said that person was probably an engineer. And they did so whether engineers made up 70% of the group or only 30%. Had they taken the numbers into account, their answers would have reflected the fact that, in one version of the group, lawyers outnumbered engineers by more than two to one. But the words of description overpowered the numbers.
My traitorous students
To illustrate some of the economic theories demonstrated on The Traitors, I ran a version of the game with a class of 13 economics undergraduates.
In our version, six students, two of whom had been given the role of traitor, sat at the front of the room. Over four rounds of about five minutes each, the group argued about the traitors’ identities before the whole class voted on their phones to banish one player.
Before we began, each of the six players was handed a card with an observation about another player, such as: “Player F seemed unusually calm before the start of the game.” The observations were randomly generated before anyone was assigned a role. They were worth precisely nothing.
The first round went the way first rounds often go on television. With nothing else to go on, the players settled on the most outspoken student, and his confidence was read as guilt. Every student in the room voted to banish him, except the accused himself.
He was a faithful.
In the second round, suspicion turned on the two players who had led the first round accusation, with the reasonable theory that a traitor might steer the group towards an innocent victim. The theory was half right. One of those two really was a traitor.
The class banished the other one.
Only in the third round did the players start reasoning from evidence that could not be faked: who had voted for whom. That system was successful in catching the first traitor, and then the second, so the final two faithfuls were eventually victorious.
As for the students watching, when the six players agreed, the audience voted with them almost to a person. When the six were split, so was the audience.
Afterwards, the traitors explained their methods. One had deliberately gone loud, leading accusations and turning faithfuls against each other. The other stayed quiet, asked few questions and simply voted with her fellow traitor.
Both are exactly how a faithful might play, which is why neither strategy gave them away. Several players admitted that the meaningless observations on their cards had shaped their suspicions, even though none of them said so out loud during the game. All four faithfuls said they had felt confident, even in the rounds they got wrong.
With the TV version of course, the viewer usually knows the identity of the traitors from the start. And despite the whole event being harder than it looks, the errors, the misplaced confidence, the lies and the theories are precisely what make the programme so much fun.
Tony Syme does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.