Pot Geometry: Planning Three Streets Backward
Introduces geometric sizing - choosing one bet fraction that, repeated on flop, turn, and river, lands exactly all-in - and shows the backward-planning habit of starting from the stacks and working toward the flop.
Assumptions: 100bb 6-max online cash, no rake unless stated. Baseline preflop ranges come from src/lib/poker/ranges.json. Quantitative claims are backed by commands recorded in content/mathlogs/postflop-turn-river/pot-geometry-and-sizing.md.
Pot Geometry: Planning Three Streets Backward
Pot Geometry: Planning Three Streets Backward sits inside the Pot Geometry & Sizing module, so the job is not to memorize one solver output. The job is to build a durable decision process for pot geometry, SPR, and sizing discipline. Introduces geometric sizing - choosing one bet fraction that, repeated on flop, turn, and river, lands exactly all-in - and shows the backward-planning habit of starting from the stacks and working toward the flop. Derives the roughly 75%-75%-75% geometry for a 100bb single-raised BTN-versus-BB pot and the much smaller fraction needed in a 3-bet pot, then applies it with AhAd on Td5c2s planning maximum extraction across three streets, and JsJh on the same board deliberately betting below geometry to keep the pot two-street sized. Explains when geometry is the goal (nutted, inelastic targets) and when deviating beats it (elastic calling ranges, protection needs). The quiz gives pot and stack pairs and asks students to compute the geometric fraction and the resulting all-in river pot. That scope is the anchor for this lesson. Every example assumes full stacks unless stated, one main villain, and a player pool where big turn and river actions are still more value-heavy than perfect theory would allow.
The first principle is that turn and river decisions are narrower than flop decisions. By the turn, ranges have been filtered by preflop action, flop texture, flop sizing, and the call or raise that followed. By the river, equity has collapsed into made hands, missed draws, and bluff-catchers. That means a vague sentence like "I have equity" is no longer enough. You need to name the value region, name the worse hands that continue, name the natural bluffs, and then compare the proposed bet size to the pot odds it gives villain.
The repo tools give several anchors used throughout this track. A nine-out draw on the turn has 19.57% to hit the river; an eight-out open-ender has 17.39%; a four-out gutshot has 8.70%; and a fifteen-out combo draw has 32.61%. A 75% pot bet needs 42.9% folds as a pure bluff and leaves the defender with 57.1% minimum defense frequency. A 150% overbet needs 60.0% folds and cuts MDF to 40.0%. A 200% shove needs 66.7% folds and cuts MDF to 33.3%. These are not decorative numbers. They decide whether pressure has enough fold equity, whether a draw can raise and tolerate a call, and whether a bluff-catcher is getting a realistic price.
Combo counting keeps the language honest. The checked-in combo tool counts AA, KK, QQ, JJ, TT, AKs, AQs, and AKo as 50 live combinations, 3.8% of starting hands before card removal. That slice is tiny, but it contains many hands that dominate high-card runouts. If a line claims a player has all of that region, the line can support bigger value bets and more bluff pressure. If a line removes that region because the player would have raised earlier, the river action must shrink or become more selective.
For Pot Geometry: Planning Three Streets Backward, start with the range that arrived. Ask which hands bet or checked earlier, which hands call the next size, and which hands raise. Do not start with your exact hand and search for a reason to like it. On the turn, medium made hands usually want either protection at a price worse hands can call or pot control against a stronger continuing range. Polar hands want a large bet that makes indifferent bluff-catchers suffer. Draws want pressure only when their equity and fold equity combine cleanly. On the river, there is no future equity, so every bet is value or bluff, and every call is a price against a range.
A useful practical split is value, bluff, showdown, and air. Value hands expect enough worse hands to call. Bluffs expect enough better hands to fold. Showdown hands beat enough checks and small bets that turning them into bluffs wastes equity. Air hands have no showdown value and need blockers or a specific population read before they get promoted into bluffs. Many expensive mistakes come from sliding one category into another because the hand feels too strong to check or too pretty to fold.
Sizing has to match the category. Small turn bets, roughly 25% to 40% pot, keep ranges wider and work well when the bettor has many medium hands that want value and protection. Standard 60% to 80% pot bets apply real pressure and start polarizing the range. Overbets are not a reward for confidence; they require nut advantage and a range that can stand behind the story. If villain can hold as many or more nutted hands than hero, an overbet becomes a donation to the top of villain's range.
Position changes how much risk a line carries. In position, checking back can realize equity and force the out-of-position player to reveal more on the river. Out of position, checking often hands villain the option to realize for free or choose a painful size. That is why block bets, probe bets, and turn leads exist. They are not signs of fear when built correctly. They are attempts to set price, deny equity, or attack a capped range before the in-position player gets to choose the cleanest response.
Against real pools, discipline matters more than theoretical elegance. Turn raises and river check-raises are under-bluffed in many games. Overbets from passive players are often too value-heavy. Small river leads from recreational players are often capped and bet-fold heavy. Those tendencies do not excuse lazy folds or automatic raises; they tell you where the burden of proof belongs. If a line is historically value-heavy, your bluff-catcher needs better blockers and a better price. If a line is capped and merged, your thin value raises get promoted.
Worked Hand 1
The first hand is the value-side version of the lesson. Hero begins by listing the worse hands that can continue, not by admiring absolute strength. If the board is static and villain keeps many second-best pairs, a medium or large value bet can be correct. If the turn or river card changes the nut advantage, the same hand may move from value to bluff-catcher. The key is whether the continuing range contains enough worse hands after this exact line and size.
Suppose hero bets 75% pot. The mathlog says a pure bluff would need 42.9% folds, but a value bet is judged differently: it needs calls from worse hands often enough to beat checking. That means the checklist is concrete. What worse top pairs call? What second pairs call? What draws call on the turn? Which worse hands fold and therefore stop contributing value? Which better hands raise? If the answer is mostly better hands and folds, the hand is not a value bet at that size.
The turn also threatens the river. A turn value bet that leaves a natural river shove creates leverage even before the river is dealt. That leverage is powerful with nutted hands and high-equity bluffs, but it can be poison with thin one-pair hands. A thin hand may be ahead now and still hate the pot it creates. Good turn strategy therefore asks whether the hand wants one street, two streets, or stacks. The correct answer often changes by a single card.
Worked Hand 2
The second hand is the pressure or defense version. Hero has blockers, a draw, or a bluff-catcher rather than a clean value hand. The default question changes from "what worse calls?" to "what better folds, and what happens when called?" A turn semi-bluff with fifteen outs can tolerate a call because the tool output gives that draw 32.61% to improve on the river. A naked gutshot with 8.70% river equity cannot pretend to be the same hand unless it has exceptional fold equity and blocker effects.
On rivers, blockers become more important because there is no card left to save a bad bluff. A blocker helps only if it removes hands from villain's continuing range or value range. Holding the ace of the completed flush suit can matter against a polar river bet because it removes the nut flush. Holding a random low card on a paired board may do nothing. The two-sided blocker test is simple: does my card block hands I want villain to have, and does it block hands I want villain not to have? A good bluff blocks calls and unblocks folds or missed draws. A good bluff-catcher blocks value and unblocks bluffs.
The same logic protects you from hero-call theater. Calling a river overbet is not brave because the pot is large. A 150% pot bet offers a price that requires about 37.5% equity by call/(pot+bet+call), and the MDF anchor is only 40.0%. You are allowed to fold a lot. The hands that continue should be high in the bluff-catching distribution and should not remove villain's likely bluffs. If your hand blocks busted draws and fails to block value, folding is not nitty; it is arithmetic.
Common Mistakes
The first mistake is carrying flop habits onto the turn. Flop c-bets can be frequent because ranges are wide and equity is diffuse. Turn barrels must be more selective because one card remains and the caller has already refused to fold once. The second mistake is saying "protection" when the bet actually folds worse and gets called by better. Protection has value only when worse hands with equity continue or better hands fold. The third mistake is bluffing rivers with the wrong blockers because the hand has no showdown value. No showdown value is a reason to consider bluffing, not a license to choose any combo.
The fourth mistake is failing to plan the next street. Before betting the turn, write the river map in your head: jam cards, thin-value cards, check-back cards, and give-up cards. Before calling the turn, know which rivers you will defend. Before raising, know whether a shove over the top is a disaster or part of the plan. Big-bet poker punishes players who create pots first and solve them later.
The fifth mistake is ignoring population. Against a balanced opponent, ratios and MDF matter. Against a player who never bluffs river raises, your near-nut hand can become a disciplined fold. Against a player who block-bets every medium pair, your value raises widen. The exploit is not to abandon theory. The exploit is to use theory as the baseline, then move in the direction the player pool actually gives you.
Table Routine
Use this routine for Pot Geometry: Planning Three Streets Backward. First, reconstruct the line street by street. Second, assign hero's hand to value, bluff, showdown, or air. Third, list villain's value combos and natural bluff or call combos. Fourth, apply blocker effects honestly. Fifth, choose the size or response that matches the range, not the emotion of the pot. Sixth, record the result after the session without rewriting the logic because the river happened to be kind or cruel.
The final rule is direct: big turn and river bets are range claims. Make claims your range can defend. Value bet when worse hands really call, bluff when better hands really fold, call when the price and blockers support it, and fold when the counted range says you are paying for a story that is not bluffing often enough.