Scalar markets settle on a level—CPI prints at 3.2%, turnout at 61%. Index and spread products settle on a difference: popular-vote margin, point spread, index return of Nasdaq minus S&P, GDP surprise versus consensus. The random variable is relative, so you must model both sides of the comparison and the threshold K in the rulebook.
Vocabulary that actually matters
An index-style contract might ask whether a constructed statistic finishes above a line—weekly equity index up, composite approval index in a band. A spread contract asks whether one quantity beats another by at least K: Candidate J’s margin at least +2.0%, Chiefs covering −3.5 points, macro nowcast beating survey by 0.2%.
Venues label both as event contracts; your job is to write the formula: YES if (X − Y) ≥ K, with explicit definitions for X and Y.
Spread as a binary on adjusted score
“Chiefs −3.5” means YES wins if Chiefs score minus Bills score is greater than 3.5. Half-point lines avoid pushes; integer lines may refund or void on exact ties per the rules—read before you arb across platforms.
A +2.0% popular-vote margin market is not the same as “wins the presidency.” You can win the electoral college with a narrow popular loss, or lose both in a landslide. P(win) and P(margin ≥ 2%) are positively linked but not equal; portfolio sizing should treat them as correlated legs, not duplicates.
Building a belief for the difference
Start from forecasts of each component—vote shares, points scored, index returns—then derive the distribution of the gap. In a simple simulation, if 51% / 49% gives margin +2 and 53% / 47% gives +6, the fraction of draws with margin at least +2 is your P(YES). Compare that to a market YES ask of 44¢ only after you have aligned K, geography, and data source with the contract.
Illustrative election margin: YES ask 44¢, your P(margin ≥ 2%) = 52% gives roughly +8¢ raw edge per share before fees—meaningless if the contract uses a different national vote definition than your model.
When spreads look incoherent with win markets
If P(win) is 70% but P(margin ≥ 10%) is 45%, that can be coherent under a wide outcome distribution. If the tail margin price is impossibly high given the win price and identical rules, you may be seeing mispricing—or different jurisdictions (popular vote vs electoral college) smuggled into similar titles.
Cross-venue comparison fails when lines differ: margin ≥ 3% on one book and margin ≥ 2% on another are different instruments, not a clean arb pair.
Macro relative views
“Nasdaq outperforms S&P this week” embeds long tech beta and short broad beta. Both legs often rise on risk-on days; the spread may move less than either index alone, lowering variance but not eliminating macro correlation. Hedging a relative view may require a third listed contract if you need rates or volatility exposure.
On macro surprise indices, YES might mean “nowcast minus consensus ≥ 0.2”—both numbers come from distinct institutions; disputes follow the institution named in rules.
Sports and political margins on regulated books
Marquee NFL spreads can be deep; obscure political margin markets are thin with wide spreads. Using a niche margin contract for belief discovery while sizing elsewhere is common when caps and fees dominate.
Push rules and voids on postponed games are part of the payoff—props and spreads share cancellation risk.
Electoral college caution
Swing-state binaries do not sum to national win probability under winner-take-all rules. Selling “overpriced national” against states without a path model to 270 is a common retail error. Spreads on popular vote are informative about sentiment, not a substitute for electoral map math unless the contract says so.
Fréchet reminder for linked contracts
Given P(win) = 0.70, joint probability of win and a huge margin is bounded—often ≤ 0.70. If P(margin ≥ 10%) trades at 45% without rule differences, ask whether the tail is rich or your win definition differs. Bounds from joint probability material prevent impossible trees when rules align.
Line shopping across venues
Kalshi might list margin ≥ 3% at 40¢ while another venue lists ≥ 2% at 55%—not arb, different K. Harmonize lines before comparing. The same lesson applies to sports handicaps at −3.5 versus −4.
Variance on favorites and dogs
Spread favorites priced at 90¢ have low per-share variance but painful tail loss on NO. Underdog covers at 25¢ carry model risk on thin injury news. Size relative to the joint with the moneyline or win market you also hold.
Using spreads as relative macro views
When you only care that tech beats broad beta, a relative index contract may express the thesis with less nominal exposure than long both indices separately—if the listed spread matches your economic definition.
Ties to earlier modules in plain language
Foundations taught that price is probability under rules. Order-book chapters warned that mids lie on wide spreads. Arbitrage material stressed that cross-venue gaps often mean different contracts. Probability material gave joint bounds and correlation—essential for any spread paired with a win market. You are applying the same toolkit to relative payoffs.
Core concepts to remember
Write the formula YES iff (X − Y) ≥ K. Model both X and Y. Check push and void rules. Treat win and spread legs as correlated unless rules prove otherwise. Harmonize K across venues before calling arb.
A second worked margin example
Suppose your simulation gives 52% to popular margin at least +2% while YES asks 44¢. Raw edge is 8¢ per share. If fees and spread consume 4¢, net edge is 4¢—worth trading only at size your cluster cap allows alongside any win market you already hold. The spread is a relative bet; the win market is not a substitute hedge unless rules tie them tightly.
Common mistakes
Trading a spread without modeling both underlyings. Equating popular-vote margin with electoral win binaries. Assuming independence between spread and turnout scalars on the same election night. Arbitraging across books with different K or push rules.
What comes next
Index and spread markets compare two sides of a scoreboard. Proposition bets zoom into micro-outcomes inside the same evening.
Next: Proposition Bets (Props): Micro-Outcome Trading