The flop cbet is the first stat a reg checks after preflop, and one of the most useless in its overall form. 55% or 70% says nothing until you know in which pots, on which boards and against whom. Two players with the same 62% cbet can differ by a bb/100, and a player with the "wrong" 80% may play the spot better than both. The cbet stat is like the average temperature across a hospital: there is a number, but no patients in sight.
What determines the frequency
In the same spot the right cbet frequency changes several times over depending on three things.
Pot type and position. In a single-raised pot from the BTN against the BB the aggressor bets often and small on almost any board. In a single-raised pot from UTG against the BB on a low connected board they bet rarely. In a 3bet pot out of position the frequencies are different again. The overall cbet mixes all of this into one number, as if you averaged the weight of a cat and an elephant and then went looking for an animal of that weight.
Texture. On dry high boards like A72 rainbow the aggressor with a range advantage bets almost always. On medium connected boards like 876 two-tone the advantage passes to the caller and the frequency drops. One "flop cbet" stat averages boards where you need to bet 90% with boards where you need to bet 30%.
The pool. Against a pool that folds the flop more often than the break-even threshold, a small cbet pays off with almost any hand. Against a pool that calls the flop and folds the turn, the same bet changes its meaning: it does not take the pot, it builds it for the second bet. Same bet, different purpose, like one key for different doors.
What an expensive mistake looks like
In database reviews cbet mistakes rarely look like "too much" or "too little" overall. They look like this:
- A 65–80% pot bet on a board where the pool folds 35%. A large size against a pool that folds less often than the threshold for that size does not pay off with bluffs, and every such bet costs money. The pool has not read about MDF, it simply does not like folding.
- A check with hands that should bet for value against a pool that calls down. It is not a loss, but you lose against the best line.
- A near 100% cbet from the BTN against the BB on any texture. In flop frequency this may be almost free, but then the "bet / check / fold" line on the turn and river starts giving away pots, and the money leaves on the following streets rather than on the flop.
The last case is the most common among NL50–NL200 regs. The flop cbet stat looks normal or even aggressive, while the losses sit in the continuations. The flop is bet with confidence, the turn is checked thoughtfully, the river is surrendered in silence, and so a thousand times over.
What to look at instead of the overall percentage
- Cbet by pot type and positions. At least four cuts: single-raised pot in position, single-raised pot out of position, 3bet pot in position, 3bet pot out of position. Compare each with the top regs at the same stake separately.
- Bet size against the pool's fold to that size. Each size has a break-even threshold: a half-pot bet pays off when folds exceed 33%, a three-quarter bet when folds exceed 43%. If the pool folds below the threshold, the size does not work as a bluff.
- The whole line. Not "did I bet the flop" but "bet the flop and then what": bet / bet / bet, bet / check / fold, bet / check / call. Each line has its own frequency and its own result, and that is what has to be compared with top regs.
- Sample. The cbet frequency in a specific cut fills quickly, hundreds of hands. The price of a mistake in a spot needs thousands. Until then the number shows a direction, not money.
In short
There is no right cbet frequency separate from the pot, the board and the pool. The overall stat is needed for one thing only: to notice that it differs strongly from the top regs at your stake and go look at the cuts. The money is found in the cuts and in the turn and river continuations, not in the flop percentage. The flop is only the opening bid. You get paid on the turn and river.