Why Poker Tells Are Unreliable
A mannerism is a tiny, easy-to-fake sample. Bet size and position are clearer information than a theory about faces.
A tell is a mannerism supposed to reveal a hand: a glance, a stillness, a joke, a hurried call. The idea is attractive because it promises a window past the cards. In practice the window is small, noisy, and often painted on. You can note behavior. You should not pay a large bet because you have a theory about someone's eyebrows.
The sample is tiny and selected
You see a few actions per hour from one person, and you remember the actions that fit a story. That is a small sample gathered by a biased collector. Why small samples mislead does not stop at sports. Three data points do not establish that a player always goes quiet with a strong hand. They establish that you noticed three moments.
Baseline matters. A quiet person is quiet with weak hands too. If you do not know how someone behaves when the pot is small and the hand is boring, you have no comparison for the interesting pot. Most home games never collect that baseline. They collect legends.
People know the legend exists
Anyone who has heard of tells can perform one. Stillness, a sigh, a show of doubt: these are cheap to fake and expensive to believe. A player who wants a fold may look strong or look weak, depending on which story they think you bought last time. Bluffing is already a message carried by the bet. A theatrical face is a second message that can be set to whatever helps the first.
The reliable information is harder to fake because it costs chips. A bet size, a position, a sequence of checks. Position at the table changes what a check means, because the player who acts first is not making the same statement as the player who acts last. Read the sequence before you read the sigh.
Culture and nerves look like patterns
People fidget when they are cold, when they are new, when they dislike silence, and when they are bluffing. Those causes produce similar motion. Assigning all of it to the cards is a story that flatters the observer. It also pressures newer players, who are nervous because the etiquette is unfamiliar, not because they hold the nuts. Punishing nervousness with a large call is a good way to be wrong and unkind at the same time.
If you are the nervous player, you do not need a false tell of calm. You need to act in turn and keep the stake small enough that a misread, yours or theirs, cannot wreck the night.
A limit on what you will pay for a hunch
You may notice something. You may not stake the rent on it. A hunch is not a count of outs, and it is not expected value. If the price already looks short on the chips alone, a theory about a face does not fill the gap. Fold, and let the theory go untested at a smaller cost.
This is adult table literacy, not a method for beating a game, and not a lesson aimed at anyone under the legal age. Faces are a poor calculator.
Frequently asked questions
Are tells completely useless?
They are weak evidence. A mannerism without a baseline and a count will mislead you more often than the movies suggest.
What should I read instead?
Bet sizes, the order of actions, and position. Those are visible and they cost chips to fake.
Can I fake a tell on purpose?
You can, and so can everyone else. That is one reason a face is a bad thing to pay for.