#Definition
A parlay trader in prediction markets constructs positions that combine multiple correlated outcomes, where all components must succeed for the trade to profit. By linking related markets together, parlay traders accept lower win rates in exchange for potentially outsized payoffs when their combined thesis proves correct.
This approach differs from simple diversification by intentionally selecting outcomes expected to move together, amplifying both potential gains and losses.
#Why Traders Use This Approach
Parlay trading attracts traders because:
- Amplified returns: Combined positions multiply potential profits beyond single-market limits
- Thesis expression: Complex views about related events can be expressed in single positions
- Capital efficiency: Smaller capital can achieve larger exposure to specific scenarios
- Correlation exploitation: Identifying underpriced correlation creates edge
- Asymmetric payoffs: Small investments can produce significant returns
However, parlay strategies require accepting that most individual parlays will lose.
#Tools of the Trade
- Parlay Calculators: To verify payouts match combined probabilities.
- Correlation Matrices: Visual grids showing how strongly different assets move together.
- Platform Combiners: Built-in tools like "Same Game Parlay" features on sportsbooks.
#Estimating Correlation
Accurately estimating correlation between outcomes is the core skill of parlay trading:
#Correlation Coefficient Interpretation
| Correlation (ρ) | Relationship | Parlay Impact |
|---|---|---|
| +1.0 | Perfect positive | Outcomes always occur together |
| +0.7 to +0.9 | Strong positive | High chance of co-occurrence |
| +0.4 to +0.7 | Moderate positive | Notable relationship |
| +0.1 to +0.4 | Weak positive | Slight connection |
| 0 | Independent | No relationship |
| -0.1 to -0.4 | Weak negative | Slight inverse relationship |
| -0.4 to -0.7 | Moderate negative | Notable inverse relationship |
| -0.7 to -1.0 | Strong negative | Outcomes rarely occur together |
#Practical Correlation Estimation
Since historical data is often unavailable, estimate correlation qualitatively:
| Scenario | Estimated Correlation | Reasoning |
|---|---|---|
| Same party winning 2 Senate seats | +0.5 to +0.7 | National environment affects both |
| Jobs report beat + Consumer confidence up | +0.4 to +0.6 | Economic indicators move together |
| Team wins Game 1 + Team wins series | +0.6 to +0.8 | Momentum and skill correlation |
| Rain tomorrow + Umbrella sales up | +0.7 to +0.9 | Direct causal relationship |
| Candidate wins primary + Same candidate wins general | +0.3 to +0.5 | Primary success predicts but doesn't guarantee |
#Correlation-Adjusted Parlay Calculation
Standard (Independent) Parlay:
P(A and B) = P(A) × P(B)
Example: 0.60 × 0.50 = 0.30
Correlation-Adjusted Parlay:
P(A and B) = P(A) × P(B|A)
Where P(B|A) = P(B) + ρ × (1 - P(B)) × sqrt(P(A)/(1-P(A)))
Simplified approximation for moderate correlation (ρ ≈ 0.5):
P(B|A) ≈ P(B) + 0.15 to 0.20
Example with ρ = 0.5:
P(A) = 0.60, P(B) = 0.50
P(B|A) ≈ 0.50 + 0.15 = 0.65
P(A and B) ≈ 0.60 × 0.65 = 0.39 (vs 0.30 independent)
#Parlay Vig Analysis
Vig (vigorish) compounds across parlay legs, often making multi-leg parlays -EV even when individual legs are fair:
#How Vig Compounds
| Legs | Individual Vig | Compounded Vig | Total Edge Lost |
|---|---|---|---|
| 1 | 5% | 5% | 5% |
| 2 | 5% | 9.75% | 9.75% |
| 3 | 5% | 14.3% | 14.3% |
| 4 | 5% | 18.5% | 18.5% |
| 5 | 5% | 22.6% | 22.6% |
#Vig Calculation Example
Single Bet Vig:
- True probability: 50%
- Market price: 52.5% (5% vig)
- Edge lost: 2.5%
Two-Leg Parlay with 5% vig per leg:
- True combined: 50% × 50% = 25%
- Market combined: 52.5% × 52.5% = 27.56%
- Edge lost: 27.56% - 25% = 2.56% absolute, ~10.25% relative
Key insight: On a 25% probability bet, losing 2.56% absolute
is losing 10.25% of expected value!
#Breaking Even on Parlays
For parlays to be +EV despite vig, you need:
| Number of Legs | Required Edge Per Leg | Or: Required Correlation Underpricing |
|---|---|---|
| 2 legs | >5% per leg | >10% combined correlation edge |
| 3 legs | >5% per leg | >15% combined correlation edge |
| 4 legs | >6% per leg | >20% combined correlation edge |
#Platform Parlay Support
Different platforms handle parlays differently:
| Platform | Native Parlays | DIY Parlays | Notes |
|---|---|---|---|
| Polymarket | No | Yes (buy multiple markets) | Must construct manually |
| Kalshi | Limited | Yes | Some linked markets available |
| Traditional Sportsbooks | Yes | N/A | Built-in parlay builders, high vig |
| Betting Exchanges | Some | Yes | Varies by exchange |
#DIY Parlay Construction on Polymarket
Since Polymarket doesn't offer native parlays:
- Buy YES on each correlated market separately
- Calculate your effective parlay cost: Multiply all purchase prices
- Track positions together: Treat them as one trade mentally
- All must win: Only profitable if ALL positions resolve YES
Example DIY Parlay:
- Buy 100 shares YES on Market A at $0.60 = $60
- Buy 100 shares YES on Market B at $0.50 = $50
- Total invested: $110
- Payout if both win: $200 (100 + 100)
- Profit if both win: $90 (82% return)
- Loss if either loses: $110 (full loss, minus any partial recovery)
#How It Works
Strategy Complexity: High
Parlay trading follows a multi-leg construction process:
-
Identify correlated outcomes
- Find multiple markets whose outcomes are logically connected
- Assess correlation strength and direction
- Determine whether markets properly price the correlation Correlation Analogy: Rain and Umbrellas are positively correlated. If you bet on "Rain" and "High Umbrella Sales", you should get a lower payout than if they were independent, because one predicts the other. Finding where the market doesn't account for this connection is the edge in combinatorial markets.
-
Construct the parlay
- Select 2-5 legs that share a common thesis
- Calculate combined probability assuming independence
- Compare to true probability accounting for correlation
-
Size appropriately
- Accept that most parlays will fail
- Risk only amounts you can lose entirely
- Build multiple smaller parlays rather than single large bets
-
Monitor component legs
- Track each market separately
- Reassess if early legs resolve favorably or unfavorably
- Consider exiting remaining legs if thesis invalidated
#Parlay Mathematics
For a simple two-leg parlay:
- Leg 1: YES at $0.60 (60% implied probability)
- Leg 2: YES at $0.50 (50% implied probability)
Independent probability: 0.60 × 0.50 = 0.30 (30%)
But if outcomes are positively correlated:
- When Leg 1 hits, Leg 2 probability increases to 70%
- True combined probability: 0.60 × 0.70 = 0.42 (42%)
Parlay cost: $0.60 × $0.50 = $0.30 per $1 payout
Expected value (if correlation priced as independence):
EV = (0.42 × $1.00) - $0.30 = $0.12 (+40% edge)
Potential payout: $1.00
Risk: $0.30 (full loss if either leg fails)
#When to Use It (and When Not To)
#Suitable Conditions
- Multiple markets with identifiable positive correlation
- Correlation underpriced by independent market pricing
- Small position sizes you can afford to lose completely
- Clear thesis connecting the outcomes
- Sufficient liquidity in each component market
#Unsuitable Conditions
- Unrelated markets forced into parlays for higher payouts
- Large position sizes that create significant loss exposure
- Markets with negative correlation (outcomes unlikely to occur together)
- Thin liquidity that prevents fair entry prices
- Situations where correlation is already properly priced
#Examples
#Example 1: Political Parlay
A trader believes a political party will perform broadly well:
- Leg 1: Party wins Senate race in State A
- Leg 2: Party wins Senate race in State B
- Leg 3: Party wins Governor race in State C
If national political environment drives all races, outcomes are correlated. Buying YES on all three creates a parlay that profits maximally if the party sweeps, costing less than three separate full-size positions.
#Example 2: Economic Cascade
A trader expects strong economic data across indicators:
- Leg 1: Jobs report beats expectations
- Leg 2: GDP growth exceeds threshold
- Leg 3: Consumer confidence rises
Economic indicators often move together. A parlay captures the thesis that the economy is stronger than markets recognize, with all legs needed for full payout.
#Example 3: Sports Championship Path
A market on championship outcomes:
- Leg 1: Team wins first playoff series
- Leg 2: Team wins second playoff series
- Leg 3: Team wins championship
Each subsequent leg is conditional on prior success. The parlay expresses confidence in the team's complete championship run at combined odds.
#Risks and Common Mistakes
- Overestimating correlation: Assuming outcomes are more connected than they actually are
- Ignoring negative EV: Parlays often have worse expected value than single bets due to vig stacking
- Size creep: Increasing parlay sizes after wins, amplifying eventual losses
- Too many legs: Each additional leg compounds the probability of failure
- Forcing correlation: Combining unrelated markets just for higher potential payouts
- Survivorship bias: Remembering big parlay wins while forgetting numerous losses
#Practical Tips
- Limit parlay legs: 2-3 legs balance amplification with reasonable win probability
- Require genuine correlation: Only parlay outcomes that logically connect
- Track overall results: Maintain records of all parlays, not just winners
- Use parlays sparingly: They should be small portion of overall trading activity
- Calculate EV carefully: Account for how each leg's vig compounds
- Consider partial exits: If early legs win, sometimes exiting remaining legs locks in profit
- Accept low hit rates: Success might mean 20-30% win rate with sufficient payoff magnitude
#Risk of Ruin for Parlay Strategies
Parlay strategies have higher risk of ruin than single-bet strategies due to concentrated exposure and low win rates:
#Risk of Ruin by Strategy Type
| Strategy | Win Rate | Avg Payout | Risk of Ruin (5% bets) | Risk of Ruin (10% bets) |
|---|---|---|---|---|
| Single bets (55% edge) | 55% | 1.9x | <5% | ~15% |
| 2-leg parlays | 25-30% | 3.3x | ~20% | ~45% |
| 3-leg parlays | 12-15% | 6-8x | ~40% | ~65% |
| 4-leg parlays | 5-7% | 12-16x | ~55% | ~80% |
#Calculating Sustainable Parlay Sizing
Kelly Criterion for Parlays:
Kelly % = (p × b - q) / b
Where:
- p = probability of winning parlay
- q = probability of losing (1 - p)
- b = net payout odds (payout / stake - 1)
Example 3-leg parlay:
- Win probability: 15%
- Payout: 7x stake
- b = 7 - 1 = 6
- Kelly = (0.15 × 6 - 0.85) / 6 = 0.05 / 6 = 0.83%
Recommendation: Even with edge, bet <1% per parlay
#Parlay Portfolio Guidelines
| Bankroll Size | Max Single Parlay | Max Total Parlay Exposure | Parlays per Month |
|---|---|---|---|
| $1,000 | $20 (2%) | $100 (10%) | 5-10 |
| $5,000 | $75 (1.5%) | $400 (8%) | 5-15 |
| $10,000 | $100 (1%) | $700 (7%) | 10-20 |
| $50,000+ | $250 (0.5%) | $2,500 (5%) | 10-25 |
Key principle: Parlays should never be the primary strategy. Use them for high-conviction correlated opportunities, not as a way to chase outsized returns.
#Related Terms
- Prediction Market
- Expected Value
- Risk Management
- Combinatorial Markets
- Conditional Markets
- Liquidity
#FAQ
#Are parlays negative expected value in prediction markets?
Not necessarily. In traditional sportsbooks, parlays typically carry negative EV due to compounded vig. In prediction market exchanges, if you identify underpriced correlation between legs, parlays can be positive EV. The key is whether your combined probability estimate exceeds what the parlay costs, not whether parlays are inherently good or bad.
#How many legs should a parlay have?
Most experienced parlay traders limit themselves to 2-4 legs. Each additional leg dramatically reduces win probability: a 50% per-leg parlay wins 25% with 2 legs, 12.5% with 3 legs, 6.25% with 4 legs. Beyond 4 legs, even strongly correlated outcomes become very unlikely to all succeed.
#How do parlay traders handle partial wins?
When some legs win and others remain pending, traders face decisions about remaining exposure. Options include: holding to see if remaining legs hit, selling remaining positions if prices improved, or exiting if the original thesis feels invalidated. Some platforms allow trading of partially-complete parlay positions.
#Is parlay trading suitable for prediction market beginners?
Parlay trading is generally unsuitable for beginners because it requires understanding correlation, accepting frequent losses, and managing complex multi-leg positions. Beginners benefit from simpler single-market strategies that provide clearer feedback for learning. Parlays are better suited to experienced traders with established frameworks for evaluating correlated positions.