Arbitrage Between Polymarket and Kalshi: What It Is and Why It’s Harder Than It Looks

What prediction market arbitrage is

A binary prediction contract settles at either $1 or $0. If you can buy YES on one platform for 43 cents and buy the economic equivalent of NO on another platform for 54 cents, your total cost is 97 cents. At resolution, one side pays $1 and the other pays $0. Your gross profit is 3 cents per share set.

That is the clean textbook version.

Real markets are messier. Sometimes the opposite side exists as a direct binary contract. Sometimes you have to build it synthetically by buying every other outcome in a multi-outcome market. Sometimes the headline looks identical but the resolution rules differ just enough to break the hedge. That is why arbitrage starts with contract design, not price alone.

For beginners, this is where calculators become useful. You often need to convert quoted odds into probabilities, compare apples to apples across platforms, and then ask whether the spread is still attractive after fees. A good workflow is to use an odds converter or implied probability calculator first, then check whether the position is truly an arbitrage or just a low-edge trade that belongs in an expected value calculator.

How arbitrage works between Kalshi and Polymarket specifically

Kalshi and Polymarket often list markets on similar themes: elections, macro data, crypto, sports, and other headline-driven events. But the order books are separate, the user bases are different, and the platforms are built under different regulatory and product structures. Kalshi is a CFTC-regulated exchange in the U.S. Polymarket operates globally through separate legal entities, and its U.S. product is run by QCX LLC d/b/a Polymarket US, a CFTC-regulated designated contract market that is being rolled out via waitlist.

Those differences matter because they affect who can trade, how quickly prices update, and how much capital is available in any given market. They also affect costs. Kalshi’s current fee schedule says the general taker fee formula is:

with no settlement fee. Polymarket says most markets are fee-free, but certain market types, including crypto, NCAAB, and Serie A, now have taker fees.

So the arbitrage question is never just, “Which platform has the lower displayed number?” It is, “Which platform has the better executable price, after fees, for the exact exposure I need?”

Worked example with real numbers

The cleanest real-world cross-platform arbs are often synthetic, not one-click binary pairs.

As of March 21, 2026, Kalshi’s “Number of rate cuts in 2026?” market showed “Exactly 0 cuts” at 34%, while Polymarket’s corresponding multi-outcome market showed “0 (0 bps)” at 36%. On Polymarket, the other outcomes in that market collectively represented the non-zero-cuts side.

That creates a synthetic cross-platform setup:

• Buy Exactly 0 cuts on Kalshi at 34 cents.
• Buy the rest of the field on Polymarket for about 64 cents.

Your total pre-fee cost:

If there are exactly 0 cuts, the Kalshi leg pays $1 and the Polymarket basket pays $0. If there is at least 1 cut, one of the Polymarket non-zero outcomes pays $1 and the Kalshi leg pays $0. In either case, the gross payout is $1.

That sounds like a 2-cent arbitrage. But now we add the frictions.

Kalshi’s published fee formula implies that buying 100 contracts at 34 cents would incur about $1.58 in taker fees:

So for 100 shares, the rough math becomes:

The apparent 2-dollar gross edge shrinks to roughly 42 cents before you even deal with the hardest part: actually buying the Polymarket basket at those quoted levels. Because Polymarket’s displayed probabilities are usually the midpoint of the bid-ask spread, not necessarily the actual ask you can hit, a tiny edge can disappear the moment you try to execute. If the spread is wider than 10 cents, Polymarket may show the last traded price instead, which can be even less helpful for live arbitrage decisions.

That is why many “arbs” are really just screen arbs. They exist in a spreadsheet, not in your account.

Why these arbs exist

The first reason is different user bases. Kalshi and Polymarket do not have one unified order book. They have separate pools of traders reacting to the same news at different speeds.

The second reason is fee structure. A headline mispricing can survive longer when platforms charge different fees or display prices differently. A 2-cent difference may be meaningless on one platform and meaningful on another depending on how fees are assessed. Kalshi’s standard fee curve is most punishing around the middle of the probability range, while Polymarket’s fee burden is market-dependent and often zero, though not always.

The third reason is liquidity fragmentation. In event-driven markets, especially elections and macro, one community may move faster than the other. That is one reason traders watch Polymarket election odds so closely: they can move sharply on narrative and flow, not just on new hard information. When that happens, Kalshi and Polymarket can temporarily disagree.

The fourth reason is product design. Even when two markets appear similar, one may be binary and the other multi-outcome. That creates temporary pricing gaps that sophisticated traders can sometimes stitch together into a synthetic hedge.

The practical barriers

• Fees. This is the most obvious barrier, and beginners still underestimate it. A spread is not a real spread until you compare it with the actual fee burden. That is exactly why it helps to check a fees comparison page before assuming a quoted gap is tradable.
• Execution risk. Arbitrage usually involves multiple legs. One side can fill while the other moves away. Once that happens, you no longer have a locked trade. You have directional exposure.
• Capital lockup. Even a valid arbitrage can tie up money until resolution. A 1% gross return may sound attractive until you realize the capital is trapped for months.
• Access. Polymarket’s official materials say the international platform is separate from Polymarket US, the U.S. app is being rolled out via waitlist, and the platform restricts order placement from blocked regions. Polymarket’s help docs also say it strictly prohibits using VPNs or similar tools to bypass geographic restrictions. For U.S. retail traders, that makes cross-platform hedging less straightforward than many online examples imply. Meanwhile, Robinhood’s event contracts offer a third U.S.-accessible venue with its own fee structure and market selection — but also its own limitations on what markets are available and how orders work.
• Rules mismatch. “Fed cuts in 2026” sounds simple, but serious traders still read the rule pages. Resolution source, timing, and edge cases matter. A good arbitrage setup is based on economic equivalence, not just similar wording.

Is arbitrage realistic for retail traders?

Sometimes, but only episodically.

Retail traders can occasionally catch real cross-platform mispricings, especially when markets are moving fast. But as a repeatable strategy, arbitrage is tougher than it sounds. The people most likely to monetize it consistently tend to have alerting systems, pre-funded accounts, fast execution, and the discipline to ignore spreads that look good but are not executable.

For most users, the best reason to study arbitrage is educational. It teaches you how to think in probabilities, how to compare market structures, and how to treat fees as part of the trade rather than an afterthought. Those habits make you better at ordinary trading too, even when you are not locking in a risk-free position.

Frequently Asked Questions

Common questions about prediction market arbitrage.

Takeaway

Arbitrage opportunities between Polymarket and Kalshi are real, but they are episodic, not constant. Fees often eat the spread. Execution risk turns many apparent arbs into ordinary trades. Capital can stay locked up longer than people expect. And for U.S. traders, access restrictions can make the second leg harder than the spreadsheet suggests. Still, understanding arbitrage makes you a better prediction market trader, because it forces you to think clearly about pricing, liquidity, and market structure. Even if you never execute one, that skill is worth having.

Start Evaluating Your Trades

The tools below are free and designed specifically for prediction market math:

• Fees Compared — see what each platform actually charges at any contract price
• Expected Value Calculator — find out if a trade has positive EV after fees
• Implied Probability Calculator — convert contract prices to implied probabilities
• Odds Converter — translate between contract prices, American odds, and implied probability
• Robinhood Prediction Markets — how Robinhood’s prediction market offering compares to Kalshi and Polymarket