Continuation betting is the most over-taught, under-specified action in Hold'em. People remember "c-bet a lot" from a decade ago, then fire three streets into a calling station because the first bet was supposed to work. A c-bet is a flop bet from the preflop aggressor. A delayed c-bet is the same idea on the turn after a flop check. A double barrel is a turn bet after a flop bet. A triple is the river. An overbet is a size larger than the pot. A block bet is a small river bet that wants a call from worse and a fold from some better. A give-up is a check that admits the story ended. All of these are tools. None of them is a personality.
Texture, range advantage, and SPR are the floor; they live on postflop fundamentals. This page is the betting tree: sizes, success frequencies, and when to stop. If you need the call-price formula, jump to pot odds, equity, and SPR or the pot odds tool.
Flop c-bet sizing is a sentence, not a habit
Small (25–33% pot) is the default on dry boards where you have range advantage and you want to bet often. You can bet small with a huge slice of your range because the defender misses a lot, and because a small bet still denies equity from ace-high and from gutshots that would peel a bigger size. Large (66–75% pot, sometimes a pot-sized bet) is for polarized spots: wet boards, nut-advantage flips, turns and rivers where medium hands check. If you use one size for every flop, the size stops meaning anything. If you use small on 987 two-tone and large on T62 rainbow, you have the mapping backwards.
On ten-six-deuce rainbow, bet small, bet often. Your range has the overpairs, the ace-high, the missed Broadway that still has six outs. Their range has a lot of air. A third-pot bet makes them fold the air, continue the pairs and the rare set, and pay with the gutshots. You do not need a 70% fold to show a profit. You need them to fold often enough, plus you keep equity when they call.
Does a small flop c-bet need to work?
Pot $6. You bet $2 (33%). You risk $2 to win $6.
Break-even fold frequency if you had 0% equity when called: 2 / (2+6) = 25%.
MDF for the defender: 1 − 2/(6+2) = 1 − 0.25 = 75%. They are "supposed" to continue 75% if they never want to be exploited by this size.
Live big blinds do not continue 75% on T62. They continue 40–55% and then ask you why you always have it. Bet small, bet often, until they start check-raising. Then mix in checks with your air and some of your medium pairs.
On nine-eight-seven two-tone, that 25% break-even number is a trap. They continue too often, they raise too often, and your air has to fight a range that flopped. Check more. When you do bet, go larger and polar: sets, two pair, strong draws, and some air that blocks the straight. A third-pot stab here is how you donate to a check-raise.
Delayed c-bets: the turn as a second first bet
You checked the flop as the raiser. The defender checked back, or you were out of position and they checked to you on the turn. A delayed c-bet is you betting the turn after that flop check. It works when the flop was a bad c-bet texture (connected, monotone) and the turn is a brick that missed their draws and their check-raise range. It also works as a protection line: you had top pair on a wet flop, you checked to keep the check-raise in the box, the turn bricked, now you bet.
Delayed c-bets fail when the turn completes the obvious draw, when the defender is the type to float the flop with air and stab any check, and when your checking range on the flop was so strong that betting now is only the nuts. Mix. If you delay-c-bet every brick, they start raising you. If you never delay-c-bet, they start stabbing every checked flop. The leak is the 100% policy, not the play.
Double barrels: the turn has to change something
A second barrel is not "I bet the flop so I bet the turn." The turn should increase your fold equity, improve your hand, or both. Good barrel cards: overcards that pick off their pair range (an ace on T62 when you have King-Queen and they have a ten they hate now), blanks that miss the draws they peeled (a deuce on Q93 two-tone), pairing cards that scare the one-pair calling range. Bad barrel cards: the third heart when you do not have a heart, the eight on 972 when they have all the 86s and T8s in the defend.
PREFLOP Hero (BTN) opens KQo to 2.5bb. BB calls. Pot 5.5bb.
FLOP Ts 6c 2d. BB checks. Hero bets 1.8bb. BB calls. Pot 9.1bb.
TURN Ah. BB checks. Hero bets 6.5bb.
The ace is a good barrel card for King-Queen. It is in your opening range a lot (every AX you opened), it is in their defend as some AX they would have check-raised or peeled, and it makes their ten feel sick. You still lose to AT, A6, 22, 66, TT, and some stubborn AX. You are not betting because you "hit." You are betting because the story got worse for their calling range and better for your raising range.
Turn barrel, break-even folds
Pot 9.1bb. You bet 6.5bb (~70%).
Break-even fold% if zero equity when called: 6.5 / (6.5 + 9.1) ≈ 41.7%.
You are not at zero equity. King-Queen has two overs to a six, a gutshot if a jack rolls off, and showdown versus some missed junk if they peel and check river. Real required fold frequency is lower than 42%. If this opponent folds the turn more than half the time after calling T62, the barrel is mandatory. If they never fold a ten, pick a different card to fire, or give up and use KQ as showdown.
A triple-barrel tree, with the give-up marked in ink
Same hand, same flop. We will walk three rivers. You are not supposed to hero-call these. You are supposed to know which one is a bet and which one is a check-back.
PREFLOP Hero (CO) opens A5s to 2.5bb. BB calls. Pot 5.5bb. SPR ~18.
FLOP Kc 7d 2h. Dry, king-high. BB checks. Hero bets 1.8bb (range advantage, small size). BB calls. Pot 9.1bb.
TURN 3c. Backdoor clubs, a wheel card. BB checks. Hero bets 6.5bb. Two pair and wheels just got real in your range; their sevens got worse; some king-x still calls. BB calls. Pot 22.1bb. Stack behind ~89bb.
You have ace-high, a gutshot to the wheel, and the ace of spades, which is not a club. You barreled because the three of clubs is a good card for a cutoff opening range (A3s, 45s, club draws, AK still betting) and a bad card for a lot of their 7x and 88–TT. Now the river branches.
River A: the four of hearts — fire
RIVER 4h. You made a wheel. Pot 22.1bb. BB checks. Hero bets 16bb (value, not a tiny block; they still have kings that cannot fold). If BB raises, you continue. This is the nuts. Do not get cute and check.
You arrived here as a bluff and you left as value. That happens. The triple was going in even without the four, as we will see, but the four turns the hand into a stack-off. Size up. Stations will call a king. Nits will call a king. Give them a chance to pay the straight.
River B: the nine of clubs — fire as a bluff, with a number
RIVER 9c. Flush completes. Straight does not. You missed. Pot 22.1bb. BB checks. Hero bets 22bb (pot, polar). Represent the club. Ace-five of spades does not block their clubs, which is not ideal, but it blocks some AX they might call with, and the nine of clubs is a scare card for one-pair kings.
River bluff, pot-sized
Pot 22.1bb. Bet 22.1bb.
Break-even fold frequency: 22.1 / (22.1 + 22.1) = 50%.
You need this to work half the time if they never call worse and you never have showdown value. Ace-high actually has showdown versus missed 86s, missed 54s, and some stubborn jack-x. So you need less than 50% folds. Against a pool that overfolds completed flushes on the river — which is most live pools and a lot of low-stakes Zoom — this is a print. Against a station who "always calls a king," it is a donation. Name the opponent before you name the size.
River C: the seven of spades — give up
RIVER 7s. Pair the middle card. Pot 22.1bb. BB checks. Hero checks. Give up.
The seven is a terrible bluff card. Their 7x just made trips. Their 87s, 76s, T7s, J7s — all the junk that peeled the flop — just got there. Kings still call. Your ace-high still beats a missed club draw and still loses to every seven and every king. Betting here is continuing a habit. Checking back is how you keep the 22bb you already put in from turning into 44. If they are a maniac who check-raises this river as a bluff more than they show up with a seven, you can deviate. That is a person. The card itself is a stop sign.
Three rivers. One holding. One triple-barrel tree. The give-up is the part people skip in study, because give-ups do not make clips. Give-ups make win rates.
Overbets and block bets are opposite tools
An overbet is larger than the pot: 1.3x, 1.7x, 2.2x. You use it when your range is polar and theirs is capped. Classic: you are the in-position raiser, flop and turn went bet-call, river is a brick, they check, you have the nuts or air. Medium hands check back. The overbet makes the bluff-catcher miserable because the price is ugly and your range is "nuts or nothing." If your river betting range is still full of top pair, do not overbet. You will get called by the exact hands that beat top pair and you will fold out the exact hands you beat. Polarize first. Then size.
1.7x river overbet, required folds
Pot $80. You bet $136.
Break-even fold%: 136 / (136 + 80) ≈ 62.96%.
They need to fold almost two times in three if you are pure air. That is a high bar. It is also why overbet bluffs need a blocker (you hold a card that makes their calling hands less likely) and a population that overfolds. If they call 50% of the time, this bluff is negative EV even before rake. Value overbets, on the other hand, want calls. Same size, opposite job. Do not mix those jobs in your head.
A block bet is a small river bet, often 20–40% pot, from a medium-strong hand that wants two things: worse hands to call, and better hands to not jam. You block with a missed combo draw that made a pair, with top pair on a wet river you cannot bet large, with a third pair that might be good. Stations call block bets with worse. Aggressive regs raise them as bluffs. If the player on your left raises every block bet, stop blocking and start checking or sizing up polar. The block bet is not "pot control." Pot control is a check. The block bet is a small value bet with a defensive side effect.
Give-ups, and the ego problem
Folding out of a barrel sequence is not the same as giving up a betting lead. Giving up means you check, you let them have a cheap showdown or a cheap stab, and you do not turn ace-high into a 1.7x bluff on a pairing river. The ego problem is that the first two streets already went in, so the third "has to" go in or the first two were a waste. That is the sunk-cost fallacy in a hoodie. The first two streets had their own EV. The river has its own EV. They do not owe each other.
Minimum defense frequency is the other ego trap. MDF says a defender should continue X percent to not be exploited by a certain size. It does not say you, as the bettor, are entitled to that fold rate. If you fire three barrels into a station, MDF is a bedtime story. They will continue too much. Your bluffs die. Your value prints. So you stop bluffing and you start thinning the value. That is the whole exploitative loop, and it is why GTO versus exploitative is a sibling, not a contradiction.
MDF reminder, so nobody weaponizes it
Facing a bet: MDF = 1 − bet / (pot + bet)
Pot $100, bet $50: MDF = 1 − 50/150 = 66.7%. They should continue about two times in three if they want to be unexploitable versus this size.
That is their problem. Your problem is whether they actually do. If they fold 60% to a half-pot river, your bluffs print and you should fire more air. If they fold 20%, your bluffs die and you should fire more value, thinner. MDF is the floor of a solved game. It is not a moral claim on a Friday night $2/$5 seat.
Study this as a tree, not as a vibe
Take one holding and one flop. Write the turn cards that are barrels and the turn cards that are give-ups. For each barrel, write the river cards that are value, bluff, or check-back. That is a tree. It will feel slow. It is still faster than firing $136 into $80 because the clip you watched had a blocker.
When the sizes have numbers attached, you are done with this page. Go back to texture and SPR if the board still looks like a Rorschach test. Go to the math pillar if the required fold frequency still feels like a rumor. And when a river actually goes in, record whether it was a barrel or a habit. Sample size will sort the rest.