In 1960, the English psychologist Peter Wason showed people three numbers, 2, 4 and 6, and told them the sequence followed a secret rule. The rule was almost embarrassingly simple, yet participants had enormous trouble finding it. The reason why ended up naming one of the most stubborn habits of the human mind.
A guessing game with a hidden trap
The setup was disarming. Participants could propose their own sequences of three numbers to test their hunches, then try to name the rule behind 2, 4, 6.
Most of them did something that feels perfectly reasonable. They came up with a rule, then tested examples that fit it. Imagine you suspect the rule is “add two each time.” You try 8, 10, 12, and it works. You try 20, 22, 24, and it works again. Feeling confident, you announce your answer, and you are wrong.
The real rule was simply any increasing sequence. Participants kept proposing rules that were far too specific, because they only tested examples that agreed with what they already believed. A single test such as 1, 7, 50 (which still climbs, but breaks the “add two” idea) would have exposed the mistake in seconds.
The experiment that gave a bias its name
Wason’s work sharpened the idea and gave it the name it still carries today: confirmation bias. It describes our general tendency to search for, interpret and remember information in a way that confirms what we already think.
Psychologists break it into three mechanisms: biased search, biased interpretation and biased memory. The 2, 4, 6 game captures the first one perfectly, since nobody looked for the sequence that could prove them wrong.
And nobody was cheating. The bias comes from automatic, involuntary mental processes, not from any deliberate attempt to deceive. It also grows stronger when we want a particular outcome, when the topic is emotionally charged, or when the belief is deeply rooted.
Thinkers spotted it long before the lab did
Wason gave the phenomenon a name, but others had noticed it long before. In the 5th century BC, the Greek historian Thucydides already observed it in his History of the Peloponnesian War.
Centuries later, the English philosopher and scientist Francis Bacon (1561-1626) described it in his Novum Organum, noting that the human mind drags everything else toward whatever confirms and agrees with it. The German philosopher Arthur Schopenhauer (1788-1860) put it even more vividly: once we adopt a hypothesis, it gives us the eyes of a lynx for anything that supports it.
When a whole city found damage that was always there
The bias does not only live in laboratories. During the famous Seattle windshield pitting epidemic, residents heard reports of mysterious damage to car windshields and started inspecting their own.
Sure enough, many found tiny pits and marks. The damage had been there all along, they had simply never looked. Each new discovery seemed to confirm an epidemic that never actually existed.
The skeptic investigator James Randi found the same pattern in psychic readings. He compared the real transcripts of sessions with what clients later remembered. Clients recalled the rare correct predictions and forgot the large majority of wrong ones. It is the same trap behind survivorship bias: count only the hits, and almost anything looks impressive.
Being smart offers no protection
It is tempting to think this only happens to other people, ideally less educated ones. The research says otherwise. High intelligence does not shield anyone, and the so-called myside bias affects recognized experts just as much as everyone else.
In a 1979 study at Stanford University, people with firm opinions on the death penalty read fictional studies with deliberately mixed results. Even after a full debriefing, they drifted back to their initial beliefs, picking apart the studies that contradicted them.
There is one piece of good news. In 2010, researchers Brendan Nyhan and Jason Reifler coined the term “backfire effect,” the idea that correcting someone makes them cling harder to a false belief. Later, much larger research by Ohio State University and George Washington University, involving more than 10,100 participants and 52 topics, found no confirmed case of it at all.
The real lesson hidden in 2, 4, 6
Today, social media filter bubbles and algorithmic curation serve up more of what we already agree with, which makes the 2, 4, 6 trap easier than ever to fall into.
Yet Wason’s puzzle carries a lesson that is mathematical at heart. A rule is never proven by the examples that agree with it, only by surviving the ones that could break it. The next time you feel completely certain, try your own version of 1, 7, 50.

