Prediction Market Arbitrage: How It Works on Polymarket and Kalshi

Prediction market arbitrage means buying every side of an outcome for less than the $1 a winning contract pays. Price gaps do open between Polymarket and Kalshi, but trading fees, slippage, and mismatched resolution rules erase most of them before you can cash in. Run any spread through the calculator below, then see where the real opportunities come from.

Fee schedules checked October 9, 2026.

Prediction market arbitrage calculator

Enter the price of each leg in cents, pick the platform and order type for each side, and set how many contracts you plan to buy. The calculator applies each platform's published fee formula, then shows your total cost, guaranteed payout, and net result after fees. The bar underneath stacks both legs and their fees against the $1.00 payout line, so you can tell at a glance whether a spread clears it or falls short. Treat every result as an estimate, since your actual fill prices and fees can differ.

A positive result here is not a guarantee. Arbitrage still carries execution, settlement, and account risk, so only trade money you can afford to lose. If trading stops feeling like a choice, call 1-800-GAMBLER for free, confidential support.

What is prediction market arbitrage?

Prediction market arbitrage is a trading strategy that buys opposing outcomes of the same event, on one platform or across several, for a combined price below the $1 the winning contract pays. Because one side has to win, the difference between your total cost and $1 is your profit before fees.

Here's the simplest version. Polymarket lists YES on an outcome at 42¢, and the matching NO on Kalshi sits at 55¢. One of each costs 97¢ combined. However the event resolves, one of those contracts settles at $1.00, leaving a 3¢ gross margin on every pair you hold. Whether that 3¢ is still there once both platforms take their cut is a separate question, and usually the deciding one, which is why the fee math gets its own section below.

Why price gaps appear

In a perfectly efficient market, these gaps wouldn't exist. They show up because event contracts trade on separate order books run by different companies, each with its own users, liquidity, and pace. Three forces do most of the work:

• News shocks | When an injury report, poll, or court ruling drops, one platform's traders often reprice faster than the other's, and the gap lasts until the slower book catches up.
• Thin liquidity | In markets with only a few resting orders, a single large trade can push the price several cents away from where the rest of the market thinks it belongs.
• Split capital | Money can't move instantly between platforms. Traders funded on one site can't always jump on a mispricing at the other, so gaps survive longer than they would on a single shared exchange.

If terms like order book, maker, or taker are new to you, our prediction market glossary defines them, and our guide to how prediction markets work covers event contracts from the ground up.

Prediction market arbitrage strategies

Every arbitrage trade rests on the same idea, owning a set of contracts where at least one is guaranteed to pay $1, but the setups differ a lot in how often they appear and how hard they are to execute. These are the five main prediction market arbitrage strategies:

1. Same-market arbitrage | Buying complementary contracts on one platform for less than $1 combined.
2. Cross-platform arbitrage | Buying YES on Polymarket and NO on Kalshi (or the reverse) when the pair costs under $1.
3. Multi-outcome (negative-risk) arbitrage | Buying YES on every outcome in a multi-candidate market when the full set costs less than $1.
4. Calendar arbitrage | Exploiting a deadline ladder where an earlier date is priced above a later one.
5. Prediction market vs. sportsbook arbitrage | Pairing an exchange contract with the opposite side at a sportsbook.

Same-market arbitrage

Difficulty: Low to moderate. Realistic edge: Rare, and usually smaller than the fees.

On Polymarket and Kalshi, YES and NO on a single contract are two views of one order book: a bid for NO at 55¢ is the same order as an offer to sell YES at 45¢. That structure keeps the two sides of one contract from adding up to less than $1. Where same-market gaps do appear is between sibling contracts on one platform. If a game is listed as two separate yes/no markets, one for each team, and the YES prices total 97¢, owning both locks in $1, provided the game can't end in a tie or get voided under a rule that treats the two markets differently.

Cross-platform arbitrage (Polymarket vs. Kalshi)

Difficulty: Moderate. Realistic edge: A few cents at best, and only when the gap beats both platforms' fees.

This is the setup most people mean by Polymarket arbitrage or Kalshi arbitrage. Both platforms list many of the same sports, political, and economic events, and their prices drift apart whenever one crowd reacts faster than the other. Near a 50¢ price, taker fees on the two legs combined come to roughly 3.5¢ per pair, so a gap has to clear that hurdle before it's worth anything. Our Polymarket review and Kalshi review cover account setup on each side, and our Polymarket vs. Kalshi comparison breaks down how their markets differ.

How Polymarket and Kalshi arbitrage works, step by step

A cross-platform trade only works if you handle each stage in order. Skipping the rules check or the depth check is where most first attempts go wrong.

1. Find the same event on Polymarket and Kalshi, then read both rule sets, including the resolution source, deadline, and edge cases.
2. Note the best YES price on one platform and the best NO price on the other.
3. Add both prices plus each platform's taker fee, and move forward only if the total stays under $1.00.
4. Check that each order book has enough contracts at those prices to fill your full size.
5. Place the leg on the thinner book first, then fire the second leg immediately after it fills.
6. Hold both positions to settlement and log the final result, fees included.

Multi-outcome (negative-risk) arbitrage

Difficulty: High. Realistic edge: Small per set, since every extra leg adds another fee.

Some markets offer many mutually exclusive outcomes, such as a party nominee, an award winner, or a championship field. Exactly one outcome will win, so if the YES prices on every outcome add up to less than $1, buying one of each guarantees a $1 payout. Polymarket calls these linked groups negative-risk markets. The catch is completeness: the set only works if it covers every possible result, including any "other" or field option, and every leg has to fill at the price you planned.

Calendar arbitrage

Difficulty: Moderate. Realistic edge: Occasional, with capital tied up until the later deadline.

Markets often come in deadline ladders, like "by June 30" and "by December 31" versions of the same question. Anything that happens by June has also happened by December, so the earlier contract should never cost more. If "by June" YES trades at 30¢ while "by December" YES sits at 25¢, buying December YES and June NO (70¢) costs 95¢ and pays at least $1 in every scenario, and $2 if the event lands between July and December. Confirm the two contracts share identical wording apart from the date before trusting the ladder.

Prediction market vs. sportsbook arbitrage

Difficulty: Moderate to high. Realistic edge: Real at times, but sportsbooks limit accounts that arb consistently.

Sports contracts on Polymarket and Kalshi can be paired with the opposite side at a regulated sportsbook. The math requires converting American odds to implied probability and sizing the sportsbook stake so both outcomes return the same amount, which our arbitrage calculator handles for traditional bets. Sportsbooks are the weak link here: they can void palpable errors and cap stakes for winning players. Our breakdown of prediction markets vs. sports betting explains how the two products settle differently.

Prediction market arbitrage opportunities: where they show up

Arbitrage opportunities cluster around moments of uncertainty and markets with little money behind them. These are the places worth watching:

• Breaking news | Live sports prediction markets are the classic case, since a late scratch or in-game swing can leave one platform's price behind the other's for a short stretch.
• Data releases | Jobs reports, inflation prints, and Fed decisions move economy prediction markets within seconds, and the two platforms rarely reprice in perfect sync.
• Multi-outcome elections | Primaries and nomination markets in politics prediction markets spread money across many candidates, which leaves long-shot prices loose enough for outcome sets to sum below $1.
• Awards and entertainment | Thinly traded pop culture prediction markets, such as award shows, tend to have wide spreads and few market makers.
• New listings | Markets in their first hours haven't attracted much liquidity yet, so early prices often wander before the crowd settles on a number.

The same features that create these gaps also make them hard to trade. Thin markets mean shallow order books, so a mispricing you spot may only be available for a handful of contracts before the price moves against you.

The math after fees

Fees decide whether a spread is an arb or just a pair of trades. Polymarket and Kalshi both charge takers on a curve that peaks at a 50¢ price and shrinks toward the extremes, so the same 3¢ gap is far more likely to survive at 90¢ than at 50¢. All figures below were checked against each platform's published schedule on October 9, 2026.

Polymarket fees

Polymarket US, the CFTC-regulated platform open to US traders, charges takers 0.0695 × contracts × price × (1 − price), rounded to the nearest cent. That tops out around $1.74 per 100 contracts at 50¢. Maker orders earn a small rebate instead of paying a fee, and high-volume takers can get part of their fees back. This schedule took effect on October 1, 2026, per the Polymarket US fee schedule.

Polymarket's international platform, which isn't open to US persons, uses a separate category-based schedule. Since expanded taker fees took effect on March 30, 2026, rates run from 0.04 to 0.07 depending on the category, with geopolitics markets fee-free, according to Polymarket's fee documentation.

Kalshi fees

Kalshi's general taker fee is 0.07 × contracts × price × (1 − price), rounded up. Most markets charge resting maker orders nothing, though a handful carry their own multipliers, and there's no settlement fee. Buying 100 contracts at 30¢ costs 0.07 × 100 × 0.30 × 0.70 = $1.47 in fees. The current Kalshi fee schedule (PDF) has been in effect since July 7, 2026.

Worked example 1: a same-market gap that fees erase

A game on Kalshi is listed as two markets, one per team. Team A YES costs 42¢, Team B YES costs 55¢, and you buy 100 of each as a taker.

• Contract cost: $42.00 + $55.00 = $97.00
• Team A fee: 0.07 × 100 × 0.42 × 0.58 = $1.71 (rounded up)
• Team B fee: 0.07 × 100 × 0.55 × 0.45 = $1.74 (rounded up)
• All-in cost: $100.45 against a $100.00 payout, a $0.45 loss

A 3¢ gross gap at mid-range prices doesn't survive taker fees on both legs.

Worked example 2: a cross-platform arb that clears

Now a wider gap. YES costs 40¢ on Polymarket US and NO on the same event costs 54¢ on Kalshi. You buy 100 contracts of each as a taker.

• Contract cost: $40.00 + $54.00 = $94.00
• Polymarket US fee: 0.0695 × 100 × 0.40 × 0.60 = $1.67
• Kalshi fee: 0.07 × 100 × 0.54 × 0.46 = $1.74 (rounded up)
• All-in cost: $97.41 against a $100.00 payout, a $2.59 net profit (about 2.7% on capital)

That 6¢ gross gap leaves roughly 2.6¢ per pair after fees, assuming both legs fill at the quoted prices and both platforms settle the event the same way. Deposit costs matter too. Kalshi lists debit card deposit fees of up to 2%, so funding the Kalshi leg that way could cost about $1.11, close to half of this trade's profit.

Risks that kill most arbs

Arbitrage gets described as risk-free, but that label only holds in theory, because a real trade can fail at several points between spotting a gap and collecting $1. These are the risks that turn most apparent arbs into losses.

Resolution mismatch

Two contracts can ask the same question and still settle differently. In the January 2026 federal government shutdown markets, Polymarket's contract counted an official notice posted any time before 11:59 p.m. ET on January 31, while Kalshi's settled off a morning snapshot of the Office of Personnel Management website. A notice posted that afternoon would have paid YES on Polymarket and NO on Kalshi, so a trader holding Polymarket NO and Kalshi YES would have lost both legs. Always compare resolution sources, deadlines, and time zones word for word.

Leg risk and slippage

An arb is two trades, not one. If your first leg fills and the second platform's price moves before you can act, you're left holding an unhedged position. Slippage works the same way on a smaller scale: the quoted price may cover only a few contracts, and filling the rest at worse prices can quietly close the gap.

Speed

Automated traders scan both platforms constantly, so the obvious gaps often close within seconds. By the time a manual trader sees a spread, switches apps, and enters an order, the opportunity may be gone or half gone.

Capital lock-up and settlement timing

Your money stays tied up until the event settles. A 2.7% return in a week is excellent, but the same return on an election market that resolves months later is modest once you account for what else that cash could have done. You also need funds sitting on both platforms at once, and withdrawals aren't instant.

Account and access limits

Not every market is available in every state, and both platforms set their own eligibility rules, position limits, and terms of use. A trade plan that depends on a market you can't access, or on activity a platform's terms restrict, is no plan at all. Sportsbooks add another layer, since they can limit or close accounts they flag as arbitrage players.

Treat arbitrage as trading, not a sure thing. Set a budget before you start, never chase a loss from a leg that went wrong, and call 1-800-GAMBLER if you need help keeping trading under control.

Prediction market arbitrage scanners compared

A prediction market arbitrage scanner watches prices across several platforms and flags spreads as they appear, which saves you from refreshing two apps side by side. The tools below all cover both Polymarket and Kalshi. Prices and features were checked on each tool's own site on October 9, 2026, and they change often, so confirm the details before you pay for anything.

Prediction.com covers the widest range of venues, including Polymarket US, and its scanner costs nothing to use, which makes it the easiest starting point. Oddpool pairs its scanner with deeper market dashboards, though it stopped accepting new paid subscriptions, so newcomers are limited to its free tier for now. ArbBets is the most expensive option here; note that its entry-level Basic plan ($59 a month) covers sportsbook arbs only, and prediction market scanning starts at the Pro tier. EventArb is a free calculator and alert tool that also takes in Robinhood's event contracts.

No scanner removes the risks above. A flagged spread still needs a manual rules check, since automated market matching can pair contracts whose resolution terms differ. For analytics, AI research agents, and other trading aids beyond arbitrage, see our roundup of the best prediction market tools, and for wallet-level trader data, our Hashdive review covers one of the more quantitative options.

How prediction market arbitrage bots work

Because the best gaps close in seconds, serious arbitrage traders automate. A Polymarket or Kalshi arbitrage bot is built from five parts, and each one has to work for the whole system to make money:

1. Price feed | Live order book data from each platform, usually streamed over WebSocket connections rather than polled.
2. Market matching | Logic that pairs equivalent contracts across platforms. This is the hardest part to get right, since titles rarely match and resolution rules need checking.
3. Fee model | Each platform's current fee formula, applied at the actual fill price, so the bot only acts on spreads that survive fees.
4. Execution | Order placement through each platform's API, sequenced to limit leg risk.
5. Risk controls | Position caps, daily loss limits, and a kill switch that stops trading when one leg fails or a feed goes stale.

Should you build one?

For most traders, no. You'll be competing with well-funded firms on speed, the edge per trade is a few cents at best, and a single matching error can wipe out weeks of small gains. Building one makes more sense as a programming project with modest stakes than as a reliable income stream. If you want to explore, both platforms publish official documentation: the Polymarket US API docs cover REST, WebSocket, and official SDKs, and the Kalshi API docs include a demo environment for testing without real money. Manual traders can still put the same discipline to work, and our prediction market strategies guide covers position sizing and risk management.

Is prediction market arbitrage legal?

Buying contracts on two regulated platforms is generally allowed. Polymarket US and Kalshi both operate as CFTC-regulated exchanges, and holding opposing positions across them isn't prohibited in itself. A few limits still apply:

• Platform terms | Each platform sets its own rules on eligibility, account use, and trading conduct, so read them before you trade.
• Bots and APIs | Automated trading has to follow each platform's API terms and rate limits.
• State access | Availability varies by state, and some states have challenged sports event contracts in court.
• International Polymarket | The international platform is closed to US persons, so cross-platform trades for US traders should use Polymarket US.

This isn't legal advice. If you're unsure whether a strategy fits a platform's rules or your state's laws, check with the platform directly before trading, and see our ranking of the best prediction market sites for regulated options available in the US.