Overconfidence tightened your probability and size. Loss aversion hits on the exit: losses feel roughly twice as painful as equivalent gains feel good in classic prospect-theory framing—your mileage varies, but the asymmetry is common. Prediction markets turn that into disposition effects: sell winners early, hold losers, double down, and break thesis stops you already wrote.
This chapter gives expected-value exit discipline and journaling that separates P&L color from forecast quality.
Definition and trading translation
The psychological weight of losing one dollar exceeds the pleasure of gaining one dollar. In markets that shows up as holding underwater YES, doubling down losers, selling eighty-cent winners too soon, and framing “down fourteen cents” as shame instead of information.
Expected value is linear in money; your brain is not. Systems must encode linear exits. The market does not know your entry price; it offers you today’s bid. Treat every hold as a fresh buy at that bid minus fees—that mental move is how professionals avoid nursing losers.
Disposition effect in one paragraph
The disposition effect is the empirical cousin: investors sell winners too soon and ride losers. In binaries the asymmetry is stark because marks move in cents that feel like moral verdicts. Naming loss aversion helps you see why disposition appears; naming disposition helps you score patterns in logs.
Loss aversion versus endowment and sunk cost
Loss aversion fears realizing red. Endowment (next chapter) overvalues what you already hold. Sunk cost treats past research hours as reason to continue. They cluster: you own it, you hate closing red, you cite effort spent.
Prospect theory on a binary contract
Settlement is zero or one dollar, but mark-to-market swings trigger mid-trade pain.
Entry 40¢ falling to 25¢ tempts hold to avoid locking loss; expected value says exit if revised probability is below 25¢. Entry 40¢ rising to 75¢ tempts sell to feel safe; hold if revised probability still above 75¢. Flat at 40¢ brings relief freeze—re-anchor belief instead of hugging comfort.
Thesis stops invalidate on belief, not on red ink alone.
Disposition pair (worked example)
Position A: YES from 38¢ now 60¢; revised probability 62%—still small positive edge. Position B: YES from 52¢ now 38¢; revised probability 28%—negative edge.
Loss-averse behavior sells A to “bank win” and holds B for “comeback.” Correct behavior holds A if edge remains and exits B regardless of pain. Comfort sold the winner; hope kept the loser—a negative-EV bundle.
Doubling down (worked example)
You hold NO from 55¢; YES rips to 70¢ after a surprise whip count. Revised NO probability near 20%—passage likely.
Loss-averse path averages down with narrative “they can’t pass twice.” Expected-value path closes NO, accepts loss, respects bankroll freeze after a bad week, re-reads resolution text. Doubling turns a fifteen-cent mistake into a ruin candidate.
Insurance at bad prices (worked example)
Macro tail YES on “default event” at 8¢; fear spikes; you buy more at 14¢ without raising structural probability.
Outside tail rate might be 3%; calm probability 5% against 14¢ ask—negative edge on YES. Negative-EV “hedge” is a tax, not therapy.
Thesis stops and journal fields
Pre-write invalidation as an observable event, time stop to re-evaluate by date, max loss in dollars not negotiable in pain, and emotion rule—if five of five, no adds.
Belief triggers exit when whip count flips regardless of mark. Time triggers reduce size forty-eight hours before vote with no new info. Portfolio heat triggers trim worst expected value, not necessarily winners first.
Loss aversion whispers wait one more poll; thesis stop shouts the pre-registered rule.
Realizing versus paper loss
Unrealized red feels “not real yet”—but mark is a tradable exit. Realized red feels like failure—Brier cares about probability, not shame. Win-rate ego hides calibration.
Post-mortems should separate forecast error from exit discipline error. A good forecast with a bad exit is still a process bug worth fixing; a bad forecast with lucky hold is still a scoring problem.
Breakeven stops are not thesis stops
Waiting to “get back to even” is a price target, not a belief update. Thesis stops fire when the world changes—whip count flips, court rules, data series breaks trend. If your stop is only a cent level, you are trading P&L psychology, not information.
“Would I enter fresh?” (worked example)
You hold YES at 71¢, two hundred shares, entry 44¢, revised probability 68%, market 71¢.
Enter fresh at 71¢? No—edge near zero. Holding because you are up is loss aversion on giving back hypothetical upside. Sell and redeploy. You stopped a negative-EV hold; that is discipline win.
Integration without a spreadsheet
Mark-to-market pain appears in order books. Variance is not the same as loss feeling. Hedging chapters distinguish panic hedge from real hedge. Bankroll freezes after drawdown exist for a reason. Time decay on long-dated positions bleeds while you hug losers from pain. Bad bins plus loss streaks should trigger shrink, not revenge size. Fear from affect chapter buys tail YES at bad prices—cousin trap.
Weekly loss-aversion audit (fifteen minutes)
List open reds; compute revised probability versus mark; exit one hope position with edge at or below zero before week ends. Note premature green sales; pair tags with disposition notes. Enforce heat caps and no martingale after stops.
Red flags
“Can’t sell until breakeven” mixes anchor and loss aversion. Adding to losers is martingale. Selling only winners on Friday is classic disposition. Ignoring written invalidation breaks thesis stops. P&L mood setting next probability is affect returning. After a red week, ban increasing size until scorer day reviews tags—revenge trades dress as conviction.
Cooling after stops
When a thesis stop fires, a twenty-four-hour no-add rule prevents martingale storytelling. The next headline will feel like redemption; usually it is another fee. Log the stop reason in one sentence so hindsight cannot rewrite it Sunday.
Tax framing
Some traders describe exits as “paying tax” to free capital. The frame helps if it reduces shame; it hurts if it excuses random churn. Pair the metaphor with expected value: tax is acceptable when redeployment beats hold; it is not therapy.
What comes next
Beliefs and exits now meet ownership. The endowment effect makes the contract you hold feel worth more than the same contract you do not own—mark-to-thesis beats mark-to-heart.
Next: The Endowment Effect: Overvaluing What You Already Hold