Multi-Outcome Markets: Finding Mispriced Contracts
Multi-outcome markets misprice 5-15% of contracts due to overround and longshot bias. Learn 3 de-vig methods to identify and exploit these inefficiencies.
Why Multi-Outcome Markets Misprice More Than Binary Markets
A binary market has one degree of freedom. Yes or No. The crowd gets this roughly right most of the time because the constraint is simple: the two prices should sum to $1.00. A multi-outcome market has dozens of contracts that must sum to 100%. They almost never do.
Multi-outcome markets include presidential primaries, award shows (who wins Best Picture), World Cup futures, and any event with more than two possible outcomes. Each contract in the market is priced independently by buyers and sellers. The market has no built-in mechanism forcing the total to 100%. That structural gap is where edge hides.
Take a hypothetical presidential primary with 8 candidates. Add up every contract's Yes price: $0.35 + $0.25 + $0.15 + $0.10 + $0.08 + $0.05 + $0.04 + $0.03 = $1.05. That 5-cent excess is the overround. It means the market is collectively overpricing the field by 5%. But the overround is not distributed evenly across candidates. Some carry more excess than others. That uneven distribution is your opportunity.
The Overround Problem and How to Strip It Out
An overround of $1.10 on 8 outcomes means an average of 1.25 cents of excess per contract. In practice the distribution is heavily skewed.
Favorites are priced efficiently. They attract the most volume and attention. Market makers compete tightly on the top 2-3 candidates. The spread is narrow and the price reflects genuine information.
Long shots carry disproportionate overround. Candidates trading at $0.03 to $0.08 are mispriced for three reasons:
- Low-probability events are inherently harder to estimate. The difference between 2% and 5% is enormous in percentage terms but tiny in dollar terms.
- Thin liquidity means prices are stale. A contract last traded at $0.05 might reflect information from days ago.
- The favorite-longshot bias is real. Recreational traders overpay for long shots because the potential payout is exciting. A candidate with a true probability of 2% might trade at $0.05. That looks like a 3-cent difference, but the market is pricing them at 2.5x their fair value.
To find fair probabilities, you need to strip out the overround. The simplest method is proportional normalization: divide each contract's price by the total sum.
If the sum is $1.10 and a candidate trades at $0.45:
- Raw implied probability: 45%
- Normalized probability: 0.45 / 1.10 = 40.9%
That 4.1 percentage point difference matters. If your estimate of the true probability is 43%, you are not comparing against 45%. You are comparing against 40.9%. Your edge is 2.1%, not a negative 2%.
More sophisticated de-vig methods exist for multi-outcome markets. The power method and Shin method handle the overround distribution differently and can produce meaningfully different fair probabilities, especially for long shots. Run the numbers through the de-vig calculator to compare methods side by side.
Worked Example: 5-Candidate Market
Here is a complete walk-through. Five candidates in a primary. You have price data from the order book and your own probability estimates from a model or research.
| Candidate | Market Price | Normalized Fair | Your Estimate | Edge |
|---|---|---|---|---|
| A | $0.45 | 40.9% | 43% | +2.1% |
| B | $0.30 | 27.3% | 25% | -2.3% |
| C | $0.15 | 13.6% | 14% | +0.4% |
| D | $0.12 | 10.9% | 12% | +1.1% |
| E | $0.08 | 7.3% | 6% | -1.3% |
| Total | $1.10 | 100% | 100% |
Step 1: Identify the overround. Total = $1.10. Overround = 10%.
Step 2: Normalize. Divide each price by 1.10. Candidate A's fair probability is 40.9%, not 45%.
Step 3: Compare to your estimates. Candidate A shows +2.1% edge (43% - 40.9%). Candidate D shows +1.1%. Candidates B and E are overpriced relative to your model.
Step 4: Calculate EV on the underpriced contracts.
For Candidate A at $0.45 with a 43% true probability:
- Win: 43% x ($1.00 - $0.45) = 43% x $0.55 = $0.2365
- Lose: 57% x (-$0.45) = -$0.2565
- EV: -$0.020 per contract
Wait. That is negative. The raw price of $0.45 is above your 43% estimate. The edge only exists relative to the normalized fair price. This is the critical insight: you are not buying at $0.45 because you think the probability is above 45%. You are buying because $0.45 represents 40.9% in this overround-inflated market, and your estimate is 43%.
The correct EV calculation uses the normalized comparison. The contract at $0.45 in a $1.10 market is equivalent to buying at $0.409 in a fair market. Your 43% probability applied to a $0.409 entry:
- Win: 43% x ($1.00 - $0.409) = 43% x $0.591 = $0.254
- Lose: 57% x (-$0.409) = -$0.233
- Adjusted EV: +$0.021 per contract (5.1% return)
This is why normalization matters. Without it, the raw price comparison gives you a misleading signal. Use the multi-outcome calculator to run these calculations across all candidates simultaneously.
Real-World Multi-Outcome Mispricings: What the Data Shows
Multi-outcome markets on Kalshi and Polymarket consistently exhibit predictable mispricing patterns. Here are examples from actual market structures on these platforms.
Award and entertainment markets. Best Picture, Grammy, and similar multi-outcome markets routinely have 8-12 nominees. Overrounds of 15-25% are common because each contract is thinly traded. The top 2-3 favorites attract 80%+ of the volume. The remaining 5-9 nominees sit at stale prices for days. When a credible review or nomination announcement shifts sentiment, the favorites adjust within hours. The long shots take days. That lag is exploitable.
Political primary fields. Presidential primaries with 6+ candidates have been among the most mispriced multi-outcome markets on both platforms. The total implied probability frequently exceeds 115%. After a candidate drops out, the remaining contracts take 12-24 hours to fully absorb the redistribution. Traders who model the redistribution faster than the market captures the gap.
Sports futures. Conference winner and MVP markets are classic multi-outcome structures. An NBA MVP market with 15 candidates on Kalshi might sum to $1.20. The same market on Polymarket sums to $1.12. That 8-cent overround difference means a candidate priced at $0.15 on Kalshi has a normalized probability of 12.5%, while the same candidate at $0.14 on Polymarket normalizes to 12.5% as well. But if they are at $0.15 on both, the Polymarket contract is overpriced relative to its market (13.4% normalized) while the Kalshi contract is fairly priced relative to its market (12.5% normalized).
Where the Biggest Mispricings Hide
Long shots at the bottom of the board. A candidate at $0.03 in a $1.10 market has a normalized probability of 2.7%. If the true probability is 1%, the market is overpricing them by 2.7x. These are the most reliably mispriced contracts because no market maker bothers to correct a $0.03 price.
Newly added candidates. When a new candidate enters a race or a new outcome is added to a market, the existing prices do not immediately adjust downward. The total briefly jumps even higher than its usual overround, and the new contract is often priced with insufficient information.
Stale contracts in low-volume markets. A market with 12 outcomes where only the top 3 trade actively will have 9 stale-priced contracts. These prices reflect old information and present opportunities for traders with current analysis.
Cross-platform differences. The same multi-outcome market on Polymarket and Kalshi can have different overrounds. If Polymarket sums to $1.08 and Kalshi sums to $1.14, you can buy underpriced contracts on the cheaper platform. When the same candidate is $0.12 on one and $0.15 on the other, the gap is pure alpha after normalization. Use the arbitrage calculator to check if the cross-platform difference creates a risk-free arb. Read cross-platform arbitrage for the full execution framework.
Correlation with single-market mispricing. Multi-outcome market errors compound with the odds boost and promo dynamics on sportsbooks. When a sportsbook boosts a futures price, the prediction market equivalent may not adjust immediately. Cross-referencing boosted sportsbook prices against prediction market multi-outcome contracts surfaces edges that pure prediction-market-only analysis misses.
The Fee Trap on Thin Edges
Multi-outcome market edges are often small. A 2% normalized edge on a $0.10 contract is $0.002 of expected profit per contract. Kalshi's 7% winner fee on a $0.90 profit ($1.00 - $0.10) is $0.063. That fee alone is 31x your edge.
This is why fee-aware analysis matters. A contract that looks +EV before fees can be deeply -EV after them. The math flips on long shots where the potential profit is large but the probability of collecting is low.
Fee impact by price tier on Kalshi (7% winner fee):
| Buy Price | Max Profit | Fee | Fee as % of Profit |
|---|---|---|---|
| $0.05 | $0.95 | $0.067 | 7.0% |
| $0.20 | $0.80 | $0.056 | 7.0% |
| $0.50 | $0.50 | $0.035 | 7.0% |
| $0.80 | $0.20 | $0.014 | 7.0% |
| $0.90 | $0.10 | $0.007 | 7.0% |
The fee percentage is constant at 7%, but the absolute cost matters more when your edge is thin. A 3% edge on a $0.50 contract produces $0.015 of EV per contract. Kalshi's fee takes $0.035 on winning trades. On a contract that wins 53% of the time, your expected fee payment is $0.035 x 0.53 = $0.0186. That exceeds your entire edge.
Always run fee-adjusted EV before placing any multi-outcome trade. Read how prediction market fees eat your edge for the complete fee math across platforms.
Building a Multi-Outcome Portfolio
The best approach to multi-outcome markets is not picking one winner. It is building a portfolio of underpriced contracts across the board.
If three candidates in an 8-person field are each 2% underpriced after normalization, buying all three gives you positive expected value regardless of which one wins. You are not predicting the winner. You are exploiting pricing errors.
This portfolio approach connects directly to expected value fundamentals. Each individual contract might lose. But a portfolio of +EV bets produces profit over enough trials.
Size each position using the Kelly Criterion applied to the normalized probability and your estimated edge. Because these positions exist in the same market, they are inherently correlated. A strong result for one candidate means the others lose. Account for this correlation by reducing individual position sizes. The math for handling correlated positions applies directly here.
The pipeline is: normalize the market, identify underpriced contracts, calculate fee-adjusted EV, size with Kelly, and reduce for intra-market correlation. Run the full analysis through the multi-outcome calculator to see every contract's edge in one view.
Frequently asked questions
- What is overround in multi-outcome markets?
- Overround is the amount by which all contract prices in a multi-outcome market exceed 100%. If 8 candidates' Yes prices sum to $1.10, the overround is 10%. It represents the collective mispricing in the market and must be stripped out before comparing prices to your probability estimates.
- How do you find mispriced contracts in prediction markets?
- Sum all contract prices to calculate the overround. Normalize each price by dividing by the total. Compare normalized probabilities to your own estimates. Contracts where your estimate exceeds the normalized probability by 2% or more after fees are potential +EV trades.
- Does the favorite-longshot bias exist in prediction markets?
- Yes. Long-shot contracts are consistently overpriced relative to their true probabilities. This is well-documented in both sportsbook and prediction market data. Candidates trading at $0.03-$0.08 are the most likely to be overpriced, often by 2-3x their fair value.
- Should I sell overpriced contracts or buy underpriced ones?
- Both strategies work. Buying underpriced contracts is simpler because your maximum loss is the purchase price. Selling overpriced contracts (going short) has higher profit potential but exposes you to large losses if the long shot wins. Most traders focus on buying underpriced favorites and mid-tier candidates.
- How do platform fees affect multi-outcome market edge?
- Fees can eliminate thin edges entirely. Kalshi's 7% winner fee on a contract with a 3% edge will often cost more in expected fees than the edge produces. Always calculate fee-adjusted EV before trading. Polymarket's zero-fee structure preserves more of these thin edges.