EV of a Value Bet vs Checking
A river value bet is only as good as the worse hands that call it. Learn to compare EV(bet) against EV(check) with real combo counts, and spot the thin bets that quietly lose money.
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.
Betting a good hand feels automatic. It shouldn't be. A value bet is a decision with an alternative — checking — and the bet is only correct when its EV beats the EV of taking the showdown for free. Plenty of "obvious" value bets fail that comparison, and plenty of timid checks burn money the other way.
The test on the river is brutally simple to state: a value bet makes money only if, when it gets called, it's called by a worse hand more than half the time. This lesson shows where that rule comes from, then runs the full comparison on two hands — one clear, one thin enough to flip either way.
Why folds don't pay you on the river
Start with what each outcome of a river bet is worth, measured against checking instead:
- He folds a worse hand: you win the pot. But you'd have won the pot at showdown anyway. Net gain from betting: zero.
- He calls with a worse hand: you win the pot plus your bet. Checking only won the pot. Net gain: +1 bet.
- He calls (or raises) with a better hand: you lose your bet. Checking lost nothing extra. Net gain: −1 bet.
This is the asymmetry beginners miss. On the river, a fold from a hand you already beat is worth nothing — the bet's entire profit lives in the worse calls, and its entire cost lives in the better calls. So:
EV(bet) − EV(check) ≈ bet × (P(called by worse) − P(called by better))
The difference is positive exactly when worse calls outnumber better calls — that is, when more than half of the hands that call are hands you beat. (One footnote: if your bet folds out a better hand, that's pure extra profit, but it's a bluffing dynamic, not value betting — and against most river calling ranges your second-best hands aren't folding anyone better with a half-pot bet.)
"I have a strong hand" is not on the list of things that matter. The only question is what calls.
A clear value bet, counted out
You open A♦Q♦ in the cutoff, the big blind calls, and you bet top pair on the Q♠8♥3♦ flop. The turn 7♥ adds straight and flush draws, you check it back, and the river 2♠ changes nothing. The big blind checks with about 16bb in the pot. Bet 8bb or check back?
Put villain on a concrete range. He check-called a queen-high flop and checked twice more, so credit him with (combo counts computed with card removal for your A♦Q♦ and the board):
- Worse queens — QJ, QT: 16 combos. These call a half-pot bet; folding top pair to one modest river bet is not what this line does.
- Eights — T8s, 98s: 6 combos. Third pair on a four-card-straightish board. He folds these to the bet, but you beat them at showdown anyway.
- Busted draws — J9s: 4 combos. Jack-high gutshots that fold to anything and lose to your check.
- Better — slowplayed 33 plus Q8s: 4 combos. Bottom set and queens-up that check-called the whole way. (The bigger sets and two pairs mostly raise earlier, so they're discounted to these 4.)
That's 30 combos. First, the baseline:
EV(check) = you win at showdown against everything but the 4 better combos: 26/30 × 16bb = +13.9bb
Now the bet, branch by branch, measuring each leaf from the decision point (win the 16bb pot, plus 8 more when called by worse, minus 8 when called by better):
- Worse queens call: 16/30 × (16 + 8) = 16/30 × 24
- Eights and busted draws fold: 10/30 × 16
- Better hands call: 4/30 × (−8)
EV(bet) = (16 × 24 + 10 × 16 + 4 × (−8)) ÷ 30 = (384 + 160 − 32) ÷ 30 = +17.1bb
Betting beats checking by +3.2bb — and the shortcut formula agrees: 8bb × (16 − 4)/30 = 3.2bb. Check the headline rule too: 20 combos call (16 worse + 4 better), and 16 of 20 = 80% of calls come from worse hands. Way over the 50% bar. This is what a real value bet looks like: when the money goes in, you're usually winning it.
Notice the 10 combos that folded contributed nothing to the comparison — 16bb either way. All the action is in the call columns.
A thin one that can flip
You open K♥T♥ on the button, the big blind calls, and you c-bet the A♦K♠8♣ flop small. He calls. You check back the 5♥ turn, and the river is the 2♠. He checks. Pot: 14bb. You hold second pair — a hand that beats a lot and loses to every ace. Is a 5bb bet thin value or a leak?
Same drill. His river check range, with combos counted (card removal for K♥T♥ and the board):
- Worse kings — K9s, K9o, K7s, K6s, K4s, K3s: 16 combos. Standard big-blind defends that called one small bet on an ace-high board.
- Top pair — ATs, A9s, A8s: 7 combos. The Ax that didn't raise preflop and played it slow. All of these beat you (A8s is even two pair).
- Busted draws — QJ: 16 combos. The royal gutshot that called the flop and bricked out. Pure showdown losers.
39 combos total; you beat 32 of them. Baseline:
EV(check) = 32/39 × 14bb = +11.5bb
Scenario A — the sticky opponent. He hates folding any pair to a one-third-pot bet: all 16 worse kings call, all 7 Ax call, the 16 QJ combos fold.
EV(bet) = (16 × 14 + 16 × 19 + 7 × (−5)) ÷ 39 = (224 + 304 − 35) ÷ 39 = +12.6bb
Worse hands make up 16 of 23 calls — 70% — and the bet beats checking by about +1.2bb. Thin, real value.
Scenario B — the honest opponent. Same range, different policy: he folds every worse king ("he bet the river, my K9 is no good") and calls only with top pair. Now every single call has you beat.
EV(bet) = (32 × 14 + 7 × (−5)) ÷ 39 = (448 − 35) ÷ 39 = +10.6bb
The identical bet is now worth 0.9bb less than checking. Zero percent of calls are worse hands. You folded out 32 combos you were already beating and paid 5bb to the only 7 combos that continue. That is the textbook definition of lighting money on fire with a "value bet" — the hand is good, the bet is bad.
Same cards, same board, same sizing. The bet's sign flipped entirely on one read: does this player call with a worse king? That question — not your hand strength — is the value bet.
What about getting raised?
Both examples quietly assumed villain never raises, and on these rivers against these ranges that's close enough — but you should know what the raise branch does to the comparison, because it's the hidden tax on thin value.
When villain can check-raise, betting exposes you to a branch that checking never faces. Two costs arrive at once: his strongest hands extract more than one bet from you (you pay the bet, then face an awful decision for the rest), and — nastier — a capable opponent's raising range can include bluffs that blow your second-pair hand off the pot entirely. Folding K♥T♥ to a check-raise surrenders a 14bb pot you were 82% to win at showdown; calling it off against a range of slowplayed Ax-plus is throwing good money after bad. Either way, the raise branch contributes a meaningful negative term that checking simply doesn't have.
The practical effect: against opponents who check-raise rivers aggressively, the bar for thin value rises. The A♦Q♦ bet survives easily — top pair top kicker can profitably call most check-raise ranges, so the branch costs little. The K♥T♥ bet, already worth only +1.2bb in its best scenario, can be wiped out by a 6-7% check-raise frequency that forces folds. This is why "bet-fold thin against passive players, check back against maniacs" is standard practice and not nittiness: the line's EV genuinely changes with the opponent's raise frequency. The full tree treatment — explicit raise branches with probabilities and planned responses — is the final lesson of this module.
Sizing changes who calls
Everything above held the sizing fixed, but the calling range is a function of the price. Push the bet from 5bb to 12bb in the K♥T♥ hand and even the sticky opponent starts releasing K9 — the worse hands you're targeting are precisely the ones most sensitive to sizing. Top pair, meanwhile, calls 5bb and 12bb alike. Bigger sizing doesn't just thin the calls; it shifts their composition toward the hands that beat you.
That's the qualitative law to carry until the dedicated sizing material: the bigger you bet, the stronger the calling range — and the worse hands you want calls from are the first to leave. Strong value wants size; thin value usually wants the small bet that keeps second-best hands in. The 8bb half-pot bet with A♦Q♦ worked because QJ/QT stations off to that price. Overbet that river and watch your 80%-worse calls deteriorate.
You can put rough numbers on the trade-off even without solver outputs. Recall that the bet's edge over checking is approximately bet × (P(call worse) − P(call better)). Suppose in the K♥T♥ spot you're choosing between 8bb and 14bb, and your reads say: at 8bb, worse hands call 40% of the time and better hands 10%; at 14bb, the worse calls collapse to 20% while the better hands still call 12%. Compare the edges:
- 8bb: 8 × (0.40 − 0.10) = +2.4bb over checking
- 14bb: 14 × (0.20 − 0.12) = +1.1bb over checking
The bigger bet wins more per worse call and still earns less, because the calls it loses were exactly the profitable ones. Run the same structure with a nutted hand — where "better calls" is near zero at any size — and the big bet wins easily: 14 × 0.20 beats 8 × 0.40 only when there's no negative term eating the margin... it doesn't (2.8 vs 3.2), which is itself instructive: even monsters sometimes prefer the sizing that keeps the calling range wide, and the real answer depends on how steeply the calls decay with size. The frequencies above are stipulated for illustration — the framework, not the specific percentages, is the takeaway.
The reverse error matters too: betting 2bb "to get called by anything" with a monster throws away the bigger bets the worse queens would have paid. Both errors come from not asking the same question: at this price, what calls?
A river checklist
Before you fire what you think is a value bet:
- Name the worse hands that call. Actual combos — "worse Qx," "K9-type kings." If you can't list them, you're bluffing with showdown value, which is usually the worst of both worlds.
- Name the better hands that call. Slowplays, rivered two pairs, the Ax you can't beat.
- Compare the counts. More worse calls than better calls (at this sizing, versus this player) → bet. Otherwise check and take your 11.5bb baseline.
- Adjust for the player. The same bet was +1.2bb against the station and −0.9bb against the nit. Population default online: most low-stakes players call too much, which makes thin value better than it looks in a vacuum — but against the player who only continues with the nuts, your medium-strength hands should shut down.
One more habit: when your value bet gets called and you lose, don't reflexively conclude the bet was bad. In the A♦Q♦ hand you lose to 33 exactly 4 times in 20 calls — 20% of your called bets lose by design, and the bet is still worth +3.2bb over checking. Grade the bet by the range you faced, not the combo that called.