Watch a low-stakes cash game for an hour and you’ll see the same thirty seconds repeat itself. A player in the cutoff opens a hand they’d never play under the gun, gets called by the big blind, and fires a flop bet on autopilot. The big blind calls. The turn goes check-check. The river brings a bet from the blind, and the player with position folds, mutters something about “never getting there in position,” and racks up a slightly smaller stack than they sat down with.
Nothing in that sequence was a positional error in the way the word is usually taught. They had the button side of the table. They were last to act on every street. The information advantage was real and they collected it. They just never turned any of it into money.
That gap between having position and using it is where most of a beginner’s late-position losses live. The standard advice stops at “play more hands in position,” which is true, easy to apply, and about a third of the story. The other two-thirds is what you do after the flop comes out, and that part gets almost no attention because it’s harder to summarize on a chart.
The three “late” seats are not one seat
Most starting-hand charts group the hijack, the cutoff, and the button as late position, and most beginners round that grouping up into “these three are the same.” They aren’t close. From the button, nobody acts behind you before the flop. From the cutoff, one player does. From the hijack, two do, and one of them is on the button with the best seat at the table.
The practical difference shows up in the hands at the edge of your range. King-nine offsuit is a fine button open in a passive game. From the hijack, it’s a hand that gets called by the cutoff and the button, then plays out of position for three streets against a caller who has a positional edge on you. You opened a “late position” hand and ended up playing it from early position for the rest of the hand.
A useful check: before you open, count how many players can still act behind you and assume the most aggressive one will. If that thought makes you uncomfortable about the hand, it belongs in the muck from that seat even though the chart says otherwise. Hand requirements should move seat by seat, not in blocks of three, and the same logic drives how much short-handed play widens your opens compared to a full table.
Where the steal math actually breaks even
The arithmetic on a late-position steal is simple enough to do in your head, and it’s worth doing once so you stop guessing.
Say you open to 2.5 big blinds from the button in a game with no ante. The small and big blinds have 1.5 between them. If they both fold, you collect 1.5. You risked 2.5 to win 1.5. For the raise to profit on folds alone, you need everyone to fold about 62 percent of the time.
That’s a high bar, and you don’t actually have to clear it, because you win plenty of pots after the flop. But it explains why a 2.5x button open with seven-four offsuit against two blinds who defend properly is a losing proposition. The folds aren’t there, and the hand you’re left with can’t do much postflop.
Now add a big blind ante, where one player posts an extra blind that covers the table. Suddenly there’s 2.5 in the middle instead of 1.5. You’re risking 2.5 to win 2.5, and the break-even fold percentage drops to roughly 50 percent. That is a large change from one small structural detail, and it’s the mechanism behind why modern tournament players open so wide from the button. The ante effect rewrote late-position ranges more than any theory breakthrough did.
Beginners tend to import the tournament version of button aggression into a no-ante cash game, where the pot is smaller and the blinds are deeper. The seat is the same. The price isn’t.
Relative position beats seat position
Here’s the distinction that fixes more leaks than any chart. Absolute position is where you sit. Relative position is where you sit in relation to the player driving the betting.
Suppose the hijack raises, you call on the button, and the big blind calls too. Three players see the flop. You act last, which sounds ideal. But on every street the hijack is going to bet into the big blind, and the big blind will act before you. You get to see only one real decision before yours, the hijack’s bet; the big blind’s answer to it comes after you. Compare that to the same hand where you’re the big blind and the hijack raises. Now the button acts after you all the way down, and every street starts with you in the dark.
Where relative position genuinely bites is when you’re the preflop raiser and you get called by the big blind only. You’re last to act, but you’re also the one expected to bet. Your opponent checks, you bet, and the only information you collected is that they didn’t lead out. That’s thin. Being last to act is worth much more when you’re the one doing the calling, because then you get to see a real bet before deciding anything.
This is why calling a raise on the button with a suited connector can be more comfortable than opening one from the cutoff. In the first case you’re the closer. In the second, you’re the driver, and driving is expensive when you miss.
Why the turn is where most of the money goes
Take a $1/$2 game. You open the button with ace-jack offsuit to $7, the big blind calls, and the pot is $15. The flop comes jack-seven-three with two different suits. You bet $8, the big blind calls, and the pot is $31. The turn is an offsuit five.
The big blind checks. This is the moment where beginners check back, reasoning that they have one pair and don’t want to get raised.
Count what that costs. The hands still calling you here include worse jacks, pocket pairs below jacks, and every straight or flush draw that picked up the flop. A $20 turn bet gets called by a lot of those. Check the turn back instead and you’ve given free cards to the draws, and you’ve given the worse made hands a free pass to check-fold the river when a scare card lands. You collected information on the turn and spent none of it.
The exception is real, though. If the big blind is the kind of player who check-raises turns with air, or if the turn card genuinely favors their range more than yours, checking behind is fine. On a jack-high board that was checked to you, the strong hands mostly belong to you. Position is what lets you bet twice without fear of a big raise. Not betting at all is the one line that throws the advantage away.
Multiway pots quietly delete the advantage
Add a third player to a pot and the value of your seat drops sharply, because the thing position buys you is the ability to bluff and to make thin value bets. Both get worse against two opponents.
A continuation bet, the automatic follow-up bet from whoever raised before the flop, works because one player usually misses. Two players miss the same flop far less often. If you need folds from two people and each folds 60 percent of the time, you get through only 36 percent of the time. Your bluff frequency should drop hard, and the hands you keep playing should be the ones that make strong hands rather than the ones that make pairs.
Ace-nine suited is a reasonable button open. It’s a poor hand to take into a four-way pot where the top pair it makes is rarely good and the ace it holds is often dominated. Suited connectors and small pairs travel better with a crowd, because they either make something huge or fold.
When someone else’s aggression isn’t yours to copy
Late-position pressure scales with your stack, and that’s the part televised poker hides. At the 2026 World Series of Poker Main Event final table in August, Lucas Jumalon held the chip lead on every single hand on his way to the title, sitting down with 194 million chips, roughly 129 big blinds, against a short stack with 15. A player with that much room can attack from the button relentlessly because every call risks his opponent’s tournament and almost none of his own.
The short stack in that seat has a completely different job. Fold equity is the whole reason a light steal works, and a 15-big-blind stack opening the button has already committed a sixth of itself before the flop. The same hand from the same seat, and the correct play diverges entirely.
Near a money bubble the divergence gets sharper still, because chips you win are worth less in real money than the chips you risk losing. Wide button steals into a stack that can bust you get much worse. Wide button steals into medium stacks trying to survive get much better. Same seat, same cards, opposite answers, and the only variable that changed is what the chips are worth.