Dispute resolution on crypto-native venues is not an edge case—it is a priced risk on ambiguous props and a calendar risk on every market with a challenge window. Earlier chapters introduced oracles and oracle families. Arbitration is the economic and governance process that picks which reading of the rules wins when asserters, disputers, reporters, or operators disagree.
A trained forecaster can be right about the world and still lose because the contract disagrees with reality. Arbitration is where contract law wins over headline law.
The stack in one narrative
Rules text defines what winning means. Someone asserts an outcome on-chain or operations applies the rulebook. Challengers post bonds or appeals. Escalation enlarges stakes or invokes votes. Finality locks state. Remedy pays winners, zeros losers, or refunds on invalid. Traders feel this as mark gaps, frozen redemption, and weeks of uncertainty.
Centralized finality
On regulated-style paths, operations maps the appendix to an official feed. Appeals are rare. Void or cancel follows contract language. CPI restated next month does not reopen a March contract that already settled—settlement is a single shot at the oracle moment you bought.
Bet macros when the appendix maps one-to-one to a data series. Avoid judgment language unless you accept operations risk.
Optimistic oracle fights
You hold eight hundred YES at seventy-one cents on an ambiguous prop. The asserter claims NO. Mid collapses to thirty-eight cents; a disputer posts YES; a vote looms; final NO pays zero. Selling at seventy-one cents the day before assertion would have been flat versus a large loss. Pre-assert exit often dominates conviction when dispute priors exceed five percent.
Dispute gamma is bimodal. Quarter-Kelly caps on props make sense. Hedges fail if redeem is locked—plan capital as zero liquidity through the window.
Reporting rounds and invalid
A product ships beta only; rules demanded commercial launch. Reporters split. INVALID might refund near sixty cents per NO share while a strict NO would have paid one dollar. You were right commercially and wrong contractually. Track world calibration and settlement calibration separately.
Invalid as an arbitration endpoint
Invalid triggers include vague wording, missing scenarios in trees, unverifiable sources, and contradictory bundle legs. Expected value for YES is not simply probability YES wins times one dollar minus price. Add probability of invalid times refund per share. Never assume invalid is zero on permissionless community wordings.
Dispute-aware sizing in prose
You forecast fifty-five percent; executable price is fifty cents; bankroll twenty-five thousand dollars; dispute prior eight percent; invalid two percent; fourteen-day freeze. Frictionless edge is five cents. A full Kelly fraction might be ten percent of bankroll—but tail risk deserves a haircut. A few hundred dollars at stake reflects dispute tax, not the full frictionless story.
Cross-venue arbitrage into an assert window is dangerous: frozen capital kills loops that need both legs free.
How venues differ in arbitration
Kalshi optimizes fast ops finality with narrow retail bond play. Polymarket-style UMA paths show assert, challenge, and vote on explorers with medium freeze. Augur-style reporting is slow, transparent, and invalid-heavy on DIY props. The comparison is not which logo is decentralized—it is which tail distribution you are buying.
Pools may quote ghost mids while redeem is disabled—do not trade the print. Same headline on two platforms is a legal statement, not a tweet summary.
Habits that help retail survive
Score ambiguity before entry. Build dispute priors from category and creator history. Calendar assert windows and avoid new arb legs inside them. Use thesis stops on broken rules, not noise mids. Exit before assertion when dispute prior and mark disagree with your world view. Do not post bonds unless professional expected value is positive. Correlate exposure—many politics props on the same oracle family are one beta. Screenshot UI, rule URI, and block at entry for later taxes and journals. Update base rates after finals.
Retail edge is selection, not litigation
Professional disputers treat bonds as a business with legal research and capital. Retail edge usually comes from choosing clean macros, exiting before assert on props, and sizing for tails—not from posting a counter-bond because the mid feels wrong. If your dispute expected value is negative at posted bond sizes, the market is telling you the fight is priced.
Correlated dispute risk
Holding five politics props on the same optimistic stack is one dispute beta event. Resolution may differ wording by wording, but challenge windows and governance drama overlap. Portfolio thinking from bankroll chapters applies: correlated tails deserve smaller total exposure.
Worked narrative: holder through challenge
Imagine four hundred YES at twelve cents on an obscure regulatory approval market. True outcome is NO, but the asserter posts YES with a minimum bond and no disputer appears within the window. YES pays one dollar; you lose forty-eight dollars plus time. The failure mode is economic security, not your research. Defense is market selection and pre-assert marks, not outrage after the fact.
When arbitration helps transparency
Public bonds and votes are ugly but legible. You can see the fight that a regulated desk handled privately. That helps researchers and journalists; it does not guarantee you profit. Use transparency to avoid bad markets, not to assume fair mids.
Common mistakes
Holding for justice. Buying cheap YES during fights as if disputes were free options. Ignoring invalid on Augur-style markets. Arbitraging across oracle families. Assuming media calls equal finalize.
Dispute windows and portfolio liquidity
Treat open dispute windows as zero liquidity for sizing redeploy, even if the UI still quotes a mid. Capital stuck in a challenge is capital not available for a better market elsewhere. Opportunity cost is real on high-rate environments and crowded election calendars. Arbitration is not only P&L variance—it is balance-sheet lockup.
Synthesis
Disputes are markets about rules, fought with bonds instead of bids. Retail wins by avoiding bad rules and exiting early, not by posting the heroic counter-bond. Size and correlate dispute exposure like any other tail risk.
What comes next in this module
Arbitration turns ambiguous rules into bonded outcomes. The next chapter follows collateral release—redeem mechanics, complete sets, and when trustless settlement still has last-mile friction.
Media calls versus finalize
Television calls races; contracts wait for assert and challenge clocks. Your P&L path follows finalize, not the chyron. When everyone celebrates on TV, check whether redeem is even open yet.
Practice note
Estimate dispute prior before entry on the next prop you trade—use category history, not vibes. If prior times freeze days exceeds your edge in dollars, skip.
Reader takeaway
Arbitration is priced risk, not drama. Exit before assert when dispute priors are high; do not post bonds for sport. Correlate dispute exposure across props on the same stack.
Next: Settlement and Payouts Without Intermediaries