Binaries, categoricals, scalars, spreads, and props each describe what might happen. Conditional markets describe what happens on a branch: presidency given nomination, CPI path given a July cut, policy outcome assuming Senate control. They are where Bayes meets the order book—and where traders confuse timing with logic.
What “conditional” means on a ticket
Wording varies: “given,” “if X then Y,” “after X,” “assuming.” The resolution appendix decides whether X must be true at measurement, resolved first, or merely hypothetical for a what-if listing. A conditional pays on Y in the world where X is satisfied according to those rules—not in the world of your social feed.
An unconditional contract prices P(Y). A listed conditional prices P(Y | X) under the contract’s definition of X. A true joint event X and Y has probability P(X) × P(Y | X) when rules describe the same universe.
A coherence check with siblings
Illustrative nomination-to-general tree:
Candidate A wins the nomination at 42%. Conditional “A wins the presidency given nomination” at 58%. Unconditional “A wins the presidency” at 24%.
Multiply: 0.42 × 0.58 ≈ 0.24—internally consistent. If the conditional were 70% while the unconditional stayed 24%, the product 29.4% would exceed the standalone price—either mispricing, different measurement dates, or non-comparable rules. Before trading size, build a two-column rule diff; arb across venues without parity is betting on clerical errors.
Timing: the silent killer
When X is not yet resolved, the conditional quote is the market’s P(Y | X) under current beliefs—not P(Y). When X resolves YES, the conditional often collapses to an ordinary binary path on Y. When X resolves NO, the contract may void, pay NO, or freeze—read void language before you treat a dead branch as free lottery.
Some platforms list conditionals only after a primary result; others price them throughout. The listing date matters as much as the headline.
Counterfactuals and survey-style listings
“What-if” markets—“if the incumbent drops out, who wins?”—may price narrative curiosity before X is even possible. Without a clean unconditional sibling tied to the same oracle, treat them as standalone stories, not nodes in a strict tree.
Survey-style what-if tags can attract retail flow without a tight link to formal conditionals—edge is selection, not multiplication.
Hedging and expression
You might be bullish on P(Y | X) but bearish on P(X). Buying the conditional and related legs on X can be cheaper than buying unconditional Y if joint prices embed a discount after fees. Custom hedges—long unconditional win, short conditional win given nomination—are basket thinking previewed in the bundled-contracts chapter.
Macro paths use the same algebra: P(CPI below 3% given July cut) × P(July cut) should relate to unconditional CPI markets when definitions align.
Fed path illustration
Suppose P(July cut) = 55%, P(CPI below 3% by December given July cut) = 40%, and unconditional P(CPI below 3%) = 28%. The product 0.55 × 0.40 = 0.22 sits below 28%—either cuts are not sufficient in traders’ minds for the CPI story, or CPI definitions differ. A two-column rule diff beats a hot take.
Liquidity on branches
Conditional books are often thinner than headline binaries. Wider spreads mean mids lie; simulate fills on CLOB or curve moves on AMM. When X resolves, branch liquidity can vanish in minutes—plan exits before the parent collapses.
Platform habits
Regulated U.S. books often expose election series with PDF rule hierarchies. Crypto-native venues run long-tail “if Fed cuts, then…” props with fast repricing when X moves. Decentralized designs can mark outcomes invalid when the oracle cannot map “given” to reality.
When two venues list the “same” conditional, compare resolution strings before comparing cents.
Cross-venue comparison
Kalshi may list a conditional in a series hierarchy while a crypto venue lists a standalone if/then prop. Compare resolution strings line by line before comparing P(Y | X) in cents. Cross-market arbitrage lessons apply unchanged.
When the conditional is the cheaper expression
If P(X) × P(Y | X) implied by listed prices is below unconditional P(Y), buying the conditional path plus exposure to X can beat buying unconditional Y—after fees and liquidity on both legs.
Invalid and void on branches
Decentralized markets may mark a conditional invalid if X cannot be observed cleanly. Centralized books may void the ticket. Model invalid as a third outcome, not as remote noise.
News repricing X and Y|X together
When nomination odds jump, conditional presidency legs reprice in the same minute. Do not double leverage on the same headline through X and Y|X without a joint plan.
Core concepts to remember
Conditionals price paths. Check P(X) × P(Y | X) against unconditional Y. Timing and void rules matter as much as multiplication. What-if listings may lack strict siblings.
Hypothetical listings versus true branches
Survey-style what-if markets can trade for weeks without a clean sibling. They teach narrative prices, not tight tree math. Separate them mentally from exchange-listed branches with linked resolution.
Common mistakes
Multiplying prices from unrelated contracts. Ignoring void rules when X fails. Confusing P(Y | X) with calendar “after X” on props. Sizing the conditional leg without risk on X itself.
What comes next
Conditionals are edges in a graph. Combination markets wire parents, children, and sibling sets into explicit event trees.
Next: Combination Markets: Parent-Child Event Trees