Free poker tool
Heads-Up Preflop Charts
These charts are the starting point
Sharkling solves the spots the chart can't cover, and grades your play against the answer.
Heads-up baselines
100bb · no ante · cash
| Spot | Action | Frequency | Why |
|---|---|---|---|
| Button (small blind), first in | Raise | ~75-85% | Most solutions raise every pair, every ace, every suited hand. |
| Button (small blind), first in | Limp | ~5-20% | A real branch heads-up, because the button keeps position. |
| Button (small blind), first in | Fold | the remainder | Only the worst unconnected offsuit hands, and not always those. |
| Big blind vs a 2x raise | Continue | ~70-80% | 1bb to win 3bb is 3 to 1 — you need 25% equity to break even. |
| Big blind vs a 3x raise | Continue | ~55-65% | The price is worse, so the marginal offsuit hands drop out. |
| Big blind vs a limp | Raise | ~40-55% | Checking back the rest — a limp is not a hand you can punish blind. |
Frequencies are the range solver solutions land in across common raise sizes, not a single fixed chart. Heads-up strategy is unusually size-dependent: a min-raise strategy and a 3x strategy are different trees, and both are correct.
Static charts can't adapt to your opponents
Sharkling's neural solver does — it re-solves the spot for the table you're actually sitting at, then drills you on it.
Why heads-up ranges are so wide
At a 6-max table you open a hand knowing that five other players might wake up with something better. Heads-up there is exactly one other player, and the arithmetic changes completely. The worst starting hand in hold’em, 72o, still has roughly a third of the equity against a random hand; pocket aces, the best, has about 85%. Everything else sits between those two, and the spread is far narrower than the spread that matters at a full table.
Add the structure. The button posts the small blind, so half a big blind is already committed before the cards are dealt, and there is 1.5 big blinds sitting in the middle to win. A raise that folds out the big blind even a third of the time is immediately profitable with any two cards, and every hand that flops something is a bonus on top of that.
The last piece is position. The heads-up button acts first before the flop but last on the flop, turn and river — every street where the pot actually gets big. Position is worth several big blinds per 100 hands on its own, and heads-up you get it in every single hand you are on the button, which is half of them.
The button plays almost everything
A well-solved heads-up button strategy at 100 big blinds plays somewhere around 80-90% of hands. Every pair, every ace, every suited hand and every offsuit broadway raise. The folds are the genuinely disconnected offsuit holdings — 72o, 82o, 93o and their neighbours — and even some of those show a small profit against an opponent who over-folds.
It is worth being precise about what that means, because the number gets misread. It does not mean those hands are strong. J4o is a bad hand. It means that a bad hand played in position, against one opponent, for a pot that already contains 1.5 big blinds, still makes marginally more than nothing — and marginally more than nothing, repeated over half of all hands dealt, is where a heads-up win rate comes from.
Limping has a real role
In a 6-max game, limping is almost always a mistake, and the small blind limp is the one defensible exception. Heads-up the picture inverts. The button is the small blind, but unlike its 6-max counterpart it keeps position for the rest of the hand, so a cheap flop is cheap in the way that matters.
Solver solutions typically limp a supporting range behind a strong raising range: hands that play well in a small pot, flop enough to continue, and do not want to build a big pot against the top of the big blind’s range. The exact split depends on the raise size being used — a min-raise strategy limps less because the raise is already cheap; a 3x strategy limps more.
The big blind defends on price, not on hand strength
The big blind heads-up faces a raise every hand rather than once an orbit, so the defending frequency is the single highest-leverage number in the format. Facing a 2x min-raise you are putting in 1 big blind to win 3 — 3 to 1, or 25% equity needed. An enormous number of hands clear that bar against a range that contains most of the deck.
Against a 3x the price drops to 2 to 1 and the defending range tightens by roughly ten to fifteen percentage points, mostly at the offsuit end. The other lever is the 3-bet: heads-up the big blind 3-bets more often than at a full table, because the opener’s range is so wide that there is little risk of running into a monster and a great deal of fold equity to collect.
Stack depth changes everything
Heads-up ranges are more sensitive to stack depth than any other format, because the pot is already 1.5 big blinds and the blinds come round twice as fast.
| Effective stack | Button | Big blind |
|---|---|---|
| 100bb+ | Raise wide, limp a supporting range | Defend 70%+ vs a min-raise, 3-bet often |
| 40bb | Raise-heavy, fewer limps | 3-bets get larger relative to stacks |
| 20bb | Raise or fold; no limping | Mostly 3-bet jam or fold |
| 10bb or less | Open-jam roughly half of all hands | Call with roughly the top quarter to third |
The direction of travel is always the same: as stacks shorten, the post-flop tree that justified the marginal hands disappears, and the strategy collapses towards jam-or-fold. Deep, the button’s positional edge grows and the ranges widen again.
The four mistakes that cost the most
- Over-folding the big blind. Fold 60% to a min-raise and your opponent profits with any two cards, every hand, forever. This is the biggest leak in heads-up poker and it is not close.
- Playing the button like a 6-max button. Half of all hands is roughly right against four opponents. Against one, it leaves a large amount of money folded into the muck.
- Never limping, or only limping weak hands. Both are exploitable. A limping range that contains nothing strong can be attacked relentlessly.
- Ignoring the raise size. A 2x and a 3x demand genuinely different defending ranges, and most players use one range for both.
Where to go next
Heads-up ranges are the extreme end of a spectrum, and the fastest way to feel them is to see how the same seats behave with more players at the table. The full set of preflop charts covers all six 6-max seats, and the guide to reading preflop charts explains what the mixed squares actually mean before you try to apply them at 200 hands an hour.
Frequently asked questions
How many hands should the button open heads-up?
Most solver solutions play 80-90% of hands from the button at 100 big blinds, split between raising and limping, and fold only the worst unconnected offsuit holdings. The button posts the small blind, so it is already invested, faces only one opponent, and acts last on every street after the flop. Those three things together make almost any two cards playable.
Why does the big blind defend so wide heads-up?
Price and frequency. Facing a min-raise you are putting in 1 big blind to win 3, so you need only 25% equity to break even — and the worst starting hand still holds roughly a third of the equity against a random one. On top of that, the big blind faces a raise every single hand rather than once an orbit, so a folding leak compounds far faster than it does at a full table.
Is limping correct in heads-up poker?
Yes, and it is the one common format where limping is genuinely part of the equilibrium. The heads-up button is the small blind but still acts last after the flop, so a cheap flop in position has real value. That is the opposite of the 6-max small blind, which limps into a player who will have position on it for the rest of the hand.
How do heads-up ranges change when stacks get short?
The limping branch disappears first, then the flatting branch. Under about 20 big blinds the button plays a raise-or-fold strategy and the big blind answers mostly with 3-bet jams; under about 10 big blinds an open-jam-or-fold strategy is close to optimal, with the button jamming roughly half of all hands and the big blind calling with roughly the top quarter to third.
Can I use 6-max charts for heads-up?
Only as a rough floor. A 6-max button opens around half of all hands against four opponents plus the blinds; the heads-up button faces one opponent and opens far wider than that. Using a 6-max chart heads-up means folding hands that are clearly profitable, and the button seat is where almost all of a heads-up win rate lives.
Related tools
- Hand RankingsEvery hand, ranked — and what beats what.
- Equity CalculatorHand vs hand, hand vs range, range vs range.