Poker Tells: What Actually Works (and What's Overrated)

Every poker movie is about the tell. Almost no real winning is. This guide covers what the classic tells literature actually says, which live tells are worth your attention, what "tells" look like online — and, most importantly, where all of it ranks against the boring, powerful alternative: knowing your opponents' frequencies.

Why tells are overrated: the read hierarchy

A tell, at best, adjusts your estimate of one opponent's hand in one pot. A frequency — how often someone enters pots, c-bets, folds to pressure — adjusts your strategy against them in everypot. That's the whole argument, and it's why the honest ranking of information looks like this:

• 1. Position and preflop ranges. What hands can this player even have, given where they entered and how? This is most of hand reading.
• 2. Frequencies and sizing patterns. Does this player limp every pot? Only raise with premiums? Bet half-pot with everything except the nuts? Stats like VPIP, aggression, and fold-to-c-bet — the things a HUD measures online and attention measures live.
• 3. Physical and timing tells. The tiebreaker. When the first two layers say a decision is genuinely close, a solid tell is allowed to cast the deciding vote — and no more than that.

If you play online, you can skip the guesswork about your own leaks entirely: the leak report computes your real frequencies from your hand histories. Live, the same discipline applies — count how often the target actually shows down what their betting claimed.

The classic live tells worth knowing

These come from the two standard references — Mike Caro's Book of Tellsand Zachary Elwood's Reading Poker Tells — filtered to the patterns experienced live players actually still talk about. None of them is a law; every one of them is a tendency that some players reverse:

• • Strong means weak, weak means strong (Caro's core theme). Acting disinterested, sighing, or announcing "I guess I'll bet" disproportionately accompanies strength; sudden aggression theater accompanies bluffs. Recreational players act; they rarely act their real hand.
• • Trembling hands on a big bet is classically a release of tension from a monster, not nerves from a bluff — bluffing players tend to consciously stiffen and go still.
• • Stillness and silence after betting (Elwood's relaxation framing): players comfortable with their hand can freeze happily; bluffers often avoid doing anything that might trigger a call — including talking, moving, or making eye contact.
• • Glancing at chips after the flop hits — an involuntary "I want to bet this" signal that often means the board connected.
• • Reaching for chips out of turn is usually a defensive act meant to discourage your bet — disproportionately a weak or marginal holding.
• • Snap decisions: an instant call usually rules out a monster (raises take thought) and rules out air — it's the sound of a draw or a medium pair that never considered anything else.

Online tells: timing and bet sizing

Online, nobody sees your face — but everyone sees your clock and your sizings. These patterns are commonly observed population tendencies at micro-stakes; individual players vary, so confirm against showdowns before exploiting any of them:

• • Instant check — frequently the pre-selected auto-check/fold box. The range is capped and weak; a stab wins immediately far more often than it should.
• • Snap call — the decision was easy but not raise-worthy: draws and medium-strength hands. Monsters and close bluff-catches both take longer.
• • Long tank, then big raise — at low stakes this is value-heavy. Recreational players tank-bluff far less often than they tank with "how much can I get paid?"
• • Min-bets and tiny probes — commonly weak-to-medium hands "finding out where they're at." Raising them prints against players who fold the bottom of that range.
• • Sudden size jumps — a player whose value bets have been half-pot all session suddenly potting the river is telling you something changed. At micro-stakes, unusual big sizing is value far more often than bluff.

Notice that every one of these is really a frequency claim— "this pattern is weighted toward X in this population" — which is why the online version of tell-reading converges on the same skill as stat-reading. If you want the honest version of your own patterns, run your hand histories through the hand history analyzer and see what your sizings have been announcing.

Using tells without inventing reads

The failure mode of tell-hunting isn't missing tells — it's seeingones that aren't there and using them to justify calls you wanted to make anyway. Three rules keep you honest:

• ✓ A tell may only break ties. If ranges and sizing say fold, a hunch about a nervous glance doesn't override them. If the spot is genuinely close, the tell gets a vote.
• ✓ Patterns, not one-offs. One shaky bet proves nothing. The same behavior paired with the same hand type at two or three showdowns is a read.
• ✓ Weight by player type. Tells are most reliable on relaxed recreational players in big pots (real emotion, no concealment training) and least reliable on quiet regs with a routine — the players you most want to read give you the least.

And keep the meta-lesson from the etiquette side of live poker: the information you give away for free — acting out of turn, commenting on folded hands, visible frustration — costs more than most tells earn. Our etiquette guide covers the habits that keep your own signal quiet.

Common questions