Spikes, volume, open interest, spreads, calendars, multi-venue dossiers, weighted π, macro rows, and sentiment scores are sensors. Mispricing is the conclusion you earn when those sensors support a documented probability, an executable price, and positive net expected value after fees, slippage, and psychology—not when you merely disagree with the crowd.
Disagreement is cheap. Mispricing is scarce. The superforecasting module ahead asks whether your p′ is trustworthy; this chapter asks whether the market price is wrong for you, with math.
Mispricing versus disagreement, arb, and gamble
You and the market can differ without a trade. Mispricing requires edge large enough to clear costs plus a thesis you can state in three sentences. Arbitrage is locked payoff with matched rules on both legs. Gamble is clicking without a paper trail. Know which story you are telling.
A repeatable audit (ten beats)
Map the event: canonical question, cutoff, resolution source, one-line rules per venue—drop fatal mismatches from the comparison chapter.
Implied prices: mids and volume-weighted executable prices at your size on each surviving venue.
Synthesize weighted π with explicit weights logged before you look at social feeds.
Prior outside view: base rates, polls, economist medians, futures analogs—superforecasting chapters sharpen this step next.
Read tiered news and sentiment since the last close.
Compute posterior p′ with transparent Bayes logic, even qualitative.
Economics: gross edge is p′ minus ask (or NO-side equivalent); subtract fees, half-spread slippage budget, dispute haircut if rules are messy. If net edge is zero or negative, stop—no conviction override.
Decay: which calendar phase—discovery, catalyst, endgame—and does that raise your edge bar?
Execute with limit plan, hedge if portfolio correlation demands it.
Document journal row and calibration prep for after resolution.
Gate every trade on economics: beautiful stories with negative net edge are hobbies.
Worked audit: chamber control
Rules on Kalshi and a global venue match; capped retail is down-weighted. Kalshi ask 57¢, global ask 54¢, weighted π about 55.5¢. Historical base rate for similar open-seat cycles might be 52% class prior. Tier-one poll and ground-game evidence push p′ to 61¢. Gross edge seven cents minus two cents friction leaves roughly five cents net at the global ask—probe size under half-Kelly cap, not full bankroll. Kalshi at 57¢ is not an arb leg unless you can hedge both sides above costs.
If net edge were under two cents, watchlist only. Negative net edge means no trade regardless of narrative heat.
EV worksheet in prose
State p′, executable ask, gross edge, fee and slip budget, dispute haircut if applicable, net edge, Kelly cap, dollar stake. Net under two cents: watch. Two to four cents: small probe. Five cents and up: full plan if thesis survives red flags. Negative: walk.
Archetypes you will see again
Cap-stale extremes on retail-capped sites versus deep π. Rule-basis gaps that look like ten cents but are different events. Cascade overshoot after tier-three social on a spike. Decay drift without news mistaken for edge. Tree violations where sibling contracts sum wrong. Macro lag where PM moved but economists have not repriced yet—or the opposite. Liquidity mirage wide spread, lying mid. OI vacuum moves on hundreds of dollars.
Name the archetype in the journal; each suggests a different response (wait, arb, fade, fix dossier).
Multi-outcome and portfolio gates
On slates, test each outcome’s p′ᵢ against its ask and coherence of the whole distribution. Scalar bands need distribution thinking, not one mid. Conditional contracts require updating parents first. Before sizing, ask correlation with open bets, narrative duplication, sector caps, and whether another open p′ contradicts this one. Mispricing is per contract; risk is portfolio.
Red flags that end the audit
Cannot explain p′ briefly. Thesis is only “crowd wrong.” Net edge negative after costs. Fatal rule mismatch. Chasing a post-spike ask without refreshed p′. No liquidity for exit. Any one should veto the click.
Psychological bias chapters are the backstop when red flags whisper yes.
Weekly hunt without audit fatigue
Scan π versus outside benchmarks early week. Midweek, revisit high-dispersion dossiers. Daily tier-zero and tier-one narrative versus Δπ. Refresh macro rows pre-release. Friday rank candidates by net edge, not story quality. Sunday cull stale theses. Run the full ten-beat audit only on finalists—fatigue creates false positives.
How Module 08 fits together
Spikes classify moves. Volume and OI confirm conviction. Spread taxes urgency. Time decay sets edge bars. Comparison and consensus build π. Macro rows place π in context. Sentiment updates p′. This chapter decides if p′ − executable c is worth risk.
| Lesson | Role in the audit |
|---|---|
| Spikes | Timing and classification of moves |
| Volume | Confirms activity versus mirage |
| Open interest | Conviction versus churn |
| Spread | Executable cost |
| Time decay | Edge bar by calendar phase |
| Comparison | Map and rule diff |
| Consensus | Prior π |
| Macro | Outside benchmarks |
| Sentiment | Update to p′ |
Use the table as a checklist, not a substitute for the ten-beat narrative.
Core concepts to remember
Mispricing needs p′, executable price, and positive net EV. Archetypes repeat; name them. Portfolio gates matter. Fatigue creates false positives—audit finalists only.
Handoff to superforecasting
The Science of Superforecasting teaches outside view rigor, decomposition, calibration, and habits that make p′ more than gut plus Bayes adjectives. Signals find candidates; forecasting discipline decides whether you were right for the right reasons.
What comes next
You can read live prices as compressed arguments—continuously renegotiated, frictions included. The next module sharpens the forecaster behind the click.
Next: The Good Judgment Project and Superforecasters