Combo pricing breakdown
Leg comparison
Every active leg (label filled in), in the order entered above. Edge is your probability estimate minus the market's implied probability, in percentage points — positive means you think the leg is more likely than the market does.
| Leg | Market price | Your probability | Edge |
|---|
How it works
A combo/parlay position only pays out if every leg resolves YES — miss one and the whole combo is worthless, just like a sports parlay. Because all legs must hit together, the combo's fair price is the product of the individual leg prices (each expressed as a probability, price÷100), not their sum and not their average: three legs at 65¢, 72¢ and 58¢ multiply out to roughly 27¢, far below any single leg's price. That lower price is exactly why the payout multiple (1 ÷ cost per contract) is so much higher than any one leg could offer — you're being paid for clearing more hurdles, not for a mispriced edge. The default calculation assumes the legs are independent: your probability for leg B doesn't change depending on what happens with leg A. That's a simplification. Real Kalshi and Polymarket combo legs are frequently correlated — two legs both tied to the same election, the same Fed meeting, or the same macro data release tend to move together, because a shared underlying factor drives both. When legs are positively correlated, the naive independence product understates the true joint probability, because one leg resolving YES makes the correlated leg more likely to as well. The correlation-adjustment field lets you nudge the naive product up or down by your own rough estimate of how related the legs really are — it is a manual judgment call, not something this calculator computes from real data.
Reading the numbers
Take the worked example loaded by default: Leg A at 65¢ market price (your estimate 70%), Leg B at 72¢ (your estimate 75%), Leg C at 58¢ (your estimate 65%), 100 contracts, correlation adjustment at 0%. Cost per contract is 0.65 × 0.72 × 0.58 = $0.27144, so buying 100 contracts costs $27.14 total against a $100 payout if all three hit — a payout multiple of about 3.68x. Naive joint probability, multiplying your three estimates together (0.70 × 0.75 × 0.65), comes to 34.125%. With correlation left at 0%, that's also the adjusted joint probability, giving an expected value of $0.06981 per contract ($6.98 total) and an expected ROI of about 25.72% on capital risked. Now nudge the correlation adjustment to +15% — a modest assumption that these three legs tend to move together — and the adjusted joint probability rises to 0.34125 × 1.15 ≈ 39.24%. EV per contract jumps to about $0.12100 ($12.10 total), and expected ROI very nearly doubles, from 25.72% to about 44.58%. Nothing about the market prices or your raw probability estimates changed — only the correlation assumption did, and it moved expected ROI by almost 19 percentage points. That's the point of showing this field explicitly: correlation assumptions can swing parlay EV enormously, and blindly trusting the naive independence product on legs that are actually correlated in the real world is a fast way to misprice your own edge.
FAQ
Why is a 3-leg combo priced so much lower than any single leg, and why is the payout so much higher?
A combo only pays out if every leg resolves YES, so its fair price is the product of the leg prices, not their sum or average. Three legs priced at 65¢, 72¢ and 58¢ individually multiply out to about 27¢ combined (0.65 × 0.72 × 0.58 ≈ 0.271), because each additional leg is another independent hurdle the position has to clear. That lower entry price is exactly why the payout multiple is so much bigger — at a 27¢ cost per $1 of payout, a winning contract returns roughly 3.68x, far more than any single leg could offer on its own. The trade-off is proportional: win probability drops the same way the price does, so the bigger payout compensates for a much lower chance of actually collecting it, not free money.
What's wrong with just multiplying the leg probabilities together — why does correlation matter?
Multiplying probabilities together is only mathematically correct if the legs are truly independent — knowing the outcome of one leg tells you nothing about the others. Real event-contract legs are frequently not independent at all: two legs both tied to the same election, the same Fed decision, or the same macro data release tend to move together, because they're driven by a shared underlying factor. When legs are positively correlated, the naive product understates the true joint probability, because if one leg is trending YES the others sharing its driver are more likely to as well. This calculator's correlation-adjustment field is a rough manual way to nudge the naive product up or down for that effect — it is not a computed correlation from real data, just your own judgment call, and getting that judgment wrong is one of the biggest ways combo bettors misprice their own edge.
How is this different from the cross-platform prediction-market arbitrage calculator already on this site?
The arbitrage calculator compares prices for the SAME single event across two DIFFERENT platforms — for example Kalshi pricing an event at 45¢ while Polymarket prices the same event at 52¢, and the tool looks for a risk-free spread between the two. This combo/parlay calculator does the opposite kind of bundling: it prices MULTIPLE DIFFERENT events on ONE platform, packaged into a single all-or-nothing position where every leg has to hit. There's also a plain single-market calculator on this site that just converts one contract's price into payout, fees and breakeven odds with no bundling at all. All three tools answer different questions — same event two platforms, one event one platform, or many events one platform — and it's worth knowing which one actually matches the position you're looking at.
If my combo has positive expected value on paper, is it actually a good bet?
A positive expected value number is only as trustworthy as the probability estimates that produced it, and a multi-leg combo compounds that risk across every leg at once — being overconfident on any single leg's probability, or getting the correlation adjustment wrong, can flip a paper-positive EV into a real-world losing bet. Because the naive independence assumption tends to understate true joint probability when legs are correlated, a modest positive-correlation adjustment can nearly double the expected ROI on the same inputs, which cuts both ways: if you assume correlation that isn't really there, you'll overstate your edge just as easily as you'd understate it by ignoring correlation entirely. Positive EV also says nothing about variance — a combo with a payout multiple near 4x is winning well under half the time even when it's fairly priced, so bankroll sizing and being honest about how confident you really are in each leg matter as much as the EV sign itself.