EV of a Bluff: Risk, Reward, and Breakeven Folds

Assumptions: All examples assume 100bb effective stacks in a 6-max online cash game with no rake, unless a different stack size is stated in the example.

A pure bluff has no equity when called. You're not betting your cards — you're betting that your opponent's hand can't stand the pressure. That makes the math cleaner than almost any other spot in poker: two outcomes, fold or call, and you can derive exactly how often the fold has to come for the bet to print money.

This lesson derives that breakeven number, tabulates it for every sizing you'll actually use, and then computes the full EV of two real bluffs.

Deriving the breakeven

When you bluff, you risk your bet to win the pot:

• He folds: you win the pot (the reward)
• He calls (or raises): you lose your bet (the risk)

Set up the EV with F as the fold probability:

EV(bluff) = F × pot − (1 − F) × bet

The bluff breaks even where this equals zero. Solve for F:

F × pot = (1 − F) × bet F × pot + F × bet = bet F = bet ÷ (bet + pot) = risk ÷ (risk + reward)

That's the whole derivation. The breakeven fold frequency of any pure bluff is your risk divided by risk plus reward. Solver literature calls this number alpha; you'll also see it called the breakeven percentage. Either way, it depends only on the sizing — not on your cards, not on the board, not on history. Those things determine whether your opponent actually folds that often; the formula just sets the bar.

The sizing table

Run the formula for standard sizings (all computed with the odds tool, bet relative to a pot of 100):

Memorize this table the way you memorized pot odds. Three things to notice:

• Small bluffs need very few folds. A one-third pot stab profits if your opponent folds just one time in four. This is why small "probe" bets at abandoned pots are so reliably profitable against straightforward opponents.
• The curve flattens. Doubling the bet from half pot to full pot only moves the requirement from 33% to 50%. Going from pot to a 2x overbet moves it from 50% to 67%. You buy folds at a worsening exchange rate.
• It never reaches 100%. Even a massive overbet needs only about two-thirds folds. No bluff requires certainty — it requires a fold rate above its alpha.

There's a mirror image worth knowing now because you'll meet it constantly later: the defender's side of this equation. If you must fold more than alpha, the bluffer auto-profits, so a balanced defender continues with at least pot ÷ (pot + bet) of their range — the minimum defense frequency (MDF). Against a half-pot bet, MDF is 67% (fold at most 33%); against a pot-size bet it's 50%. Alpha and MDF always sum to 100%. For this lesson we only need alpha; file MDF away.

A real river bluff, start to finish

BTN

Q♠♠Q♠

J♠♠J♠

BBunknown

Flop

K♦♦K♦

Turn

You open Q♠J♠ on the button, the big blind calls, and you barrel a K♦7♣3♥ flop and a 9♣ turn — a card that gives you a double-gutter (any ten makes your straight). The river 2♣ bricks everything. The big blind checks a third time with about 20bb in the pot. Your queen-high beats nothing that calls; checking back wins approximately never.

You bet 15bb into 20bb — three-quarters pot. First, the bar:

breakeven F = 15 ÷ (15 + 20) = 15/35 = 42.9%

If the big blind folds more than about 43% of the time, this bluff makes money even though your hand is dead when called.

Now suppose you have a real read: this opponent check-calls flop and turn with a wide mix of middle pairs, gutshots, and club draws, and by this river much of that range is unimproved third pair or busted clubs that can't call a big bet. You estimate he folds 55%. Compute the actual EV, step by step:

• Fold branch: 0.55 × 20bb = +11bb
• Call branch: 0.45 × (−15bb) = −6.75bb
• EV(bluff) = 11 − 6.75 = +4.25bb

Two checks on the arithmetic. First, the branches cover everything: 55% + 45% = 100% (we lump the rare check-raise in with "call" — either way you lose your 15bb, since you're never continuing). Second, sanity-check against the breakeven: at exactly 42.9% folds the EV computes to 0.429 × 20 − 0.571 × 15 ≈ 0bb, as it must. Your 55% estimate clears the bar by 12 points, and those 12 points are worth 4.25bb every time you pull the trigger.

Notice what the EV is not: it's not "+20bb when it works." Beginners mentally book the whole pot when a bluff gets through and the whole bet when it doesn't, and end up evaluating bluffs by their last result. The bluff is worth +4.25bb at the moment you bet, before you see what he does.

The cheap stab

Now the other end of the sizing table — where bluffing gets almost embarrassingly cheap.

BTN

6♦♦6♦

5♦♦5♦

HJunknown

BBunknown

Flop

A♠♠A♠

The hijack opens to 3bb, you call on the button with 6♦5♦, the big blind comes along: 9.5bb pot (the small blind's half blind is dead money). The flop is A♠K♦4♥. Big blind checks, and — the key event — the preflop raiser checks too, declining to c-bet the single best flop for his range. Your hand is six-high with a backdoor straight draw: nothing.

You stab 3bb into 9.5bb. The bar:

breakeven F = 3 ÷ (3 + 9.5) = 3/12.5 = 24%

Both players need to fold only about one time in four for this to break even. Think about what they'd need to continue with: the big blind check-called preflop with a range full of suited junk that whiffed this flop, and the hijack just told you he probably doesn't have an ace or king (most players c-bet those nearly always here). If the two of them combine to fold even a modest 40% of the time:

EV = 0.40 × 9.5 − 0.60 × 3 = 3.8 − 1.8 = +2bb

Two big blinds of profit on a 3bb investment with six-high. This is why good players' stats show high bet frequencies in checked-around pots: the formula says small bluffs at orphaned pots barely need to work, and ranges that check twice usually fold far more than the required quarter of the time.

The discipline that goes with it: when the stab gets called or raised, let go. The EV calculation already assumed you lose your 3bb on the call branch. Firing again "because he might fold the turn" is a brand-new decision needing its own numbers, not a continuation of this one.

Which hands should do the bluffing

The EV formula grades the bet, but it also quietly tells you which hands to bet. Compare what Q♠J♠ and a hand like 8♥8♣ (third pair on that K♦7♣3♥9♣2♣ river) each give up by bluffing:

• Q♠J♠ checks back: queen-high wins the showdown almost never. The check is worth roughly 0. Turning it into a +4.25bb bluff costs you nothing — the bluff's EV is pure gain.
• 8♥8♣ checks back: third pair beats every busted draw villain arrived with. If that's 40% of his range, checking is worth around 0.40 × 20 = 8bb of showdown value. Bluffing with it means forfeiting those 8bb to chase the same fold equity — and worse, the hands that fold to your bet are mostly the very busted draws you were already beating.

Same fold percentage, wildly different opportunity cost. That's why disciplined players bluff from the bottom of their range: hands with no showdown value have nothing to lose, so any positive bluff EV is found money. Hands with real showdown value should usually check and collect it. When you catch yourself "bluffing" with a hand that beats some of his bluffs, you've found the worst of both worlds — too weak to value bet, too strong to bluff.

One refinement you'll meet again in the combinatorics module: among your no-showdown-value hands, prefer the ones whose cards make villain's calls less likely. Q♠J♠ carries a small bonus of exactly this kind — on a K♦7♣3♥9♣2♣ board, holding a queen and a jack removes combinations of KQ and KJ, the top-pair hands most likely to call your 15bb bet. Card-removal effects like these don't change the breakeven formula at all; they nudge the real-world fold% a point or two in your favor, which on a 12-point edge is a rounding improvement and on a 2-point edge is the whole bluff.

What changes when your bluff has outs

Both examples so far were pure bluffs at the moment that mattered — Q♠J♠ had zero equity when called on the river, and we treated 6♦5♦'s backdoors as worthless. The formula F = risk ÷ (risk + reward) assumes exactly that: when called, you lose, full stop.

Bet a hand that can still improve — a flush draw on the flop, that same double-gutter on the turn — and the call branch is no longer a total loss. Sometimes you get called and spike the win anyway. Residual equity puts money back into the worst branch, which means semi-bluffs need fewer folds than alpha suggests — sometimes far fewer. A flop flush draw betting half pot doesn't need 33% folds to profit; the fold equity and the card equity share the load.

Doing that math properly means comparing whole lines — bet versus check, raise versus call — with multi-branch trees. That's the line-comparison lesson later in this module. For now, lock in the boundary: the table above is the requirement for pure bluffs, and it's an over-estimate of the requirement for any bluff that keeps outs when called.

Using alpha at the table

The breakeven number does its best work as a filter, in both directions:

• Before you bluff: state the requirement, then ask if the read clears it. "This 2/3-pot barrel needs 40% folds. Does a player who check-called two streets fold 40% of his range on this river?" If you can't articulate why the answer is yes — capped range, scare card, busted draws he must give up — you don't have a bluff, you have a donation.
• Watching opponents: flip it around. The player who bets pot needs 50% folds; if you and your reads say his target folds 70% of the time, his bluff is great — and your fold is exactly what funds it. Alpha is the start of exploiting over-folders and punishing over-bluffers alike.

And keep the two numbers in this lesson separate in your head: breakeven folds come only from the sizing; EV comes from the gap between your fold estimate and that breakeven. The sizing sets the bar, the opponent determines whether you clear it, and the gap — times the pot — is your paycheck.