Skip to content

Bet Sizing

How big to bet is mostly downstream of your range shape. Here's the theory, then the practical heuristics.

Geometric sizing (pot geometry)

Bet an equal fraction of the pot on each remaining street so your final river bet is exactly all-in.

For a perfectly polarized, static-equity range, geometric sizing is EV-optimal — it maximizes the total money the opponent contributes by forcing the widest cumulative defense.

Why geometric beats jamming

Pot $100, $1,300 behind. Three pot-sized bets → opponent defends 0.5³ = 12.5% of range, contributing ~$162.50. A single $1,300 shove → MDF 100/1400 ≈ 7%, contributing ~$91. The smooth geometric line extracts nearly double.

Caveat: this maximization is exact only for perfectly polarized, static ranges — which rarely occur. Most real solver spots are non-geometric.

Larger-than-geometric (overbets) on early streets

Solvers sometimes bet >geometric (even 150% pot on the flop). Why? Strong-but-vulnerable hands want to frontload their value — get money in now, while they're ahead, because future cards will erode their equity. Overbetting also leans on nut advantage: only a range full of nutted hands can credibly bet that big.

Practical c-bet sizing heuristics (single-raised pots, IP)

From solver aggregate data:

  • Dry/static boards → small & frequent (e.g., ~⅓ pot). Equities are locked; you deny equity cheaply and bet a wide range.
  • Wet/dynamic boards → bigger & more polarized, less often. Fold equity is valuable when draws are live; you bet your strong hands + good draws bigger and check more medium hands.
  • Paired boards → c-bet often, but small.
  • Boards that complete straights/flushes in the caller's range → size down and bet less (your nut advantage shrank).

The unifying principle

Big bets require a credible top of range. Size up when you're polarized and hold the nut advantage; size down (or check) when your range is capped or medium-heavy. Sizing is a consequence of range shape, not a free choice.

Sources: Pot geometry · Larger-than-geometric sizing · c-bet sizing mechanics · IP c-bet heuristics