Stock traders swear by stop-losses; prediction market veterans argue they convert variance into realized loss. Both are partly right. This chapter defines when a pre-planned exit preserves bankroll and when it only sells cheap YES to a patient reader of news jumps.
Stops are the pressure valve on everything above: limits that filled, size that felt large, bias that spiked arousal. Prediction markets punish automatic price stops because probability jumps are discrete, not smooth.
Stops in zero-one land
Traditional stops trigger on price. Here, price is the crowd’s probability. A YES bought at 70¢ can trade 25¢ while your thesis is unchanged—the book panicked. A stop at 50¢ realizes a loss your model did not endorse. There is no corporate floor; only resolution or delisting ends the claim.
Mental stops, limit exits, occasional stop-limit if the UI offers it, calendar stops before dispute weeks, and thesis stops when facts break are the tools that actually exist. Percent drawdown stops on eighty-five-cent contracts confuse magnitude with probability.
Automated market maker exits can slip through your trigger; limits remain default on thin pools.
When exits help
Wrong resolution read discovered late—exit. Liquidity trap where you cannot size down later—exit. Broken arbitrage leg on one venue—flatten. Catalyst edge that was “before the debate” and is now gone—exit. Portfolio heat over cluster cap—trim newest leg. Dispute flag on oracle markets—reduce.
Remaining expected value below fees at the current mark is a thesis-agnostic reason to sell.
When price stops hurt
Noise after headlines often mean-reverts; thesis-only discipline may outperform a tight price band. Thin-book flickers are not trends. Contrarian trades include pain as part of the edge—pre-commit hold or do not enter. Long-dated macro lives through months of variance; size down at entry instead of panic-selling. A move from 93¢ to 88¢ is a huge percent loss but a small probability change—recompute, do not reflex.
Thesis stop versus price stop
You hold chamber-control YES at 54¢ with model 61%. A rumor without primary source drops price to 47¢; a price stop sells at 45¢ and feels smart if it keeps falling, but taxes you on noise. An official withdrawal drops price to 38¢; thesis stop sells because the race changed. Calendar stop before a regulatory approval window might exit at 68¢ to avoid a binary dispute week that could zero 71¢ exposure.
Design each trade with a one-sentence thesis, observable invalidation, max loss dollars from sizing, optional price level with minimum volume filter, time trigger, limit exit preference, and re-entry rule (“new thesis only”).
Three-layer policy without over-trading
Use at most one optional price trigger plus thesis and calendar—stacking all three guarantees churn. Primary: what fact tomorrow makes you wrong? Secondary: when does edge decay below fees? Optional price: only if liquidity is five times exit size and move exceeds eight cents on sub-eighty contracts to filter flicker.
Execute with limits through the touch; market only when thesis and gap align and you are slow, not wrong.
Stops, hedges, and platforms
Hedging can replace a stop if you monitor basis; otherwise they fight. Wide spreads on capped venues turn stops into charity. Practice thesis stops on play-money for habit; on-chain venues may make dispute window the real stop event.
Common mistakes
Moving stops wider after pain, stop on arb leg only while other leg runs, percent stops on high-priced YES, no liquidity check so limits never fill, and stopping out of contrarian trades on schedule you never committed to at entry.
Remaining EV test
Before any exit, ask: at today’s mark, after fees, is expected value of holding positive? If no, sell—even for passive labels. If yes, a price stop may be wrong unless liquidity is gone.
Contrarian and passive exceptions
Write at entry whether this trade allows mark-to-market pain. Contrarian and some passive holds explicitly include variance; a tight price stop contradicts the plan. Momentum and tactical trades should have clearer time and invalidation bounds.
Journal the exit
Record whether exit was thesis, calendar, price, or portfolio heat. Monthly review of exit tags shows whether you systematically sell winners early.
Worked calendar stop
You bought regulatory YES at 71¢ because approval was likely before month-end. By day twenty-three, no approval and no credible leak—only desk chatter. Your edge was always timing. Exiting at 68¢ gives up three cents per share but avoids a week where resolution risk dominates. A passive holder with a six-month thesis would not use that calendar; a tactical holder should.
Link to Module 08
Signal-reading chapters will show spikes and crashes on the tape. Stops interpret those moves through your invalidation, not through panic. Build stop policy now so spikes later are data, not emergencies.
Key ideas
Thesis and calendar beats blind price triggers. Liquidity filters on optional price stops. Size so exits are affordable without panic.
What comes next
Even good exit rules fail when bias overrides them at the keyboard.
Passive versus active exit rules
Passive holders should pre-write thesis and probability exits, not tight price bands. Momentum traders need time and invalidation exits more than percent stops. Contrarian traders pre-commit to pain budgets in the journal. Arb traders exit when basis breaks, not when a chart looks scary. Match stop vocabulary to the strategy sleeve you tagged at entry.
Misconceptions
“A stop protects my account” on thin books may mean you cannot exit at the stop. “I will hold to zero” on a dead thesis is choosing negative expected value to avoid realizing loss. “Stops are for stocks only” misses that thesis stops are universal. “Tight stop equals discipline” often equals paying spread to noise.
Write stops at entry when you are calm; execute them when rules trigger, not when mood spikes.
The journaling chapter will ask whether your exits matched your written rules—that is how stop discipline becomes measurable instead of heroic.
Next: Psychological Biases: Overconfidence, Recency, and Confirmation