Why Kalshi and Polymarket Price World Cup Teams Differently

The short answer

The same World Cup contract — “Will this national team win the 2026 FIFA World Cup?” — trades on both Kalshi and Polymarket, and the natural question is how far apart the two venues price it. The honest answer, measured rather than assumed, is: far less than the betting blogs claim. On a snapshot we pulled from both platforms’ own APIs in the same second — June 2, 2026, 10:09 UTC — France was priced at 17.1% on both venues, and every liquid contender sat within about 0.4 percentage points. The headline “5-to-8-cent gaps” that circulate on aggregator sites mostly evaporate the moment you compare like-for-like at a single point in time.

That is not a boring result. It is the interesting one. The liquid top of this market is efficient — two separate crowds, on two separate rails, arrive at almost the same number. The gaps that do persist live in the thinner, less-liquid corners, and they are a structural story (different audiences, different settlement rails, the cost of arbitrage) rather than a disagreement about soccer. This article shows the snapshot, explains where the dramatic-gap headlines come from, and lays out how to read cross-platform divergence without fooling yourself.

Risk note: Event contracts carry a real risk of loss, including total loss of capital. Prices quoted here are a single live snapshot and change with every trade. Nothing here is financial or betting advice.

We took the snapshot ourselves — synchronously

Most “Kalshi vs Polymarket odds” comparisons have a hidden flaw: the two numbers are captured minutes or hours apart. On a market that re-prices with every trade, that turns an ordinary time difference into a fake “spread.” So we did it the only way that means anything — both venues at once.

The method:

• Both prices from each platform’s own public market-data API, not a scraped page or a third-party aggregator. First source, both sides.
• Captured in the same second (the two API calls completed between 10:09:30 and 10:09:31 UTC on June 2, 2026).
• The same price point on both: the midpoint of each platform’s best bid and best ask, which is the standard way a venue’s displayed “probability” is derived. We did not compare one platform’s bid to the other’s ask.
• The same contract: Kalshi’s “Will [team] win the 2026 Men’s World Cup?” and Polymarket’s “Will [team] win the 2026 FIFA World Cup?” resolve on the identical event — the tournament’s champion, decided July 19, 2026.

Here is the top of the board, implied probability (midpoint), as of that snapshot:

Snapshot: June 2, 2026, 10:09 UTC. Prices are live and move constantly; by the time you read this they will have changed. Sources: Kalshi and Polymarket market data.

The largest gap on any liquid contender is 0.4 of a percentage point — well under half a cent on a $1 contract. France and Brazil are effectively identical. This is two independent markets, one denominated in US dollars and one in USDC, agreeing.

Why the favorites barely move between venues

When the same instrument trades at the same price on two unconnected venues, that is the signature of an efficient, well-arbitraged market — and the World Cup winner is about as liquid as prediction markets get. Polymarket’s World Cup Winner market alone had taken roughly $1.45 billion in cumulative volume since it opened in July 2025, by its own API at our snapshot. Deep, heavily watched markets are exactly the ones where prices aggregate dispersed information well; the foundational work of Wolfers and Zitzewitz frames a winner-take-all contract’s price as the market’s probability estimate, and decades of evidence show such prices are hard to beat (Wolfers & Zitzewitz, JEP 2004).

Two forces pin the favorites together. First, both crowds see the same world — the same injuries, the same form, the same draw — so their estimates converge on the same fundamentals. Second, even though true cross-venue arbitrage is hard (more on that below), a large gap on a liquid contract would invite one-sided pressure on the cheaper side until it closed. The result is what the table shows: at the top, there is no free money and no real disagreement. If you want a fuller side-by-side of the two platforms’ fees, settlement, and funding, see our Kalshi vs Polymarket comparison.

Where the “5-to-8-cent gap” headlines come from

If the venues agree this closely, why do “Kalshi vs Polymarket World Cup odds” roundups keep advertising fat gaps? Because the gap is usually an artifact of how it was measured. Four traps, all avoidable:

1. Prices snapped at different times. This is the big one. A market that moves on every trade will show a “spread” between any two timestamps that has nothing to do with the venues disagreeing — it is just the clock. Capture both at once, and most of it disappears.
2. Bid-on-one vs. ask-on-the-other. Quote Kalshi’s best ask against Polymarket’s best bid and you have manufactured a gap out of two bid-ask spreads. Comparing midpoint to midpoint removes it.
3. Thin-order-book “midpoints.” On a deep longshot, the book can be nearly empty. At our snapshot, Czechia on Kalshi showed a 0c bid and a 1c ask — there is no real price there, yet a naive script records a 0.5c “midpoint” and reports a gap against Polymarket’s 0.2c. That 0.3c is noise from an empty book, not a signal.
4. Payout framing. “Portugal pays $90 more per $100 risked on Kalshi” sounds enormous, but it is a sub-penny probability difference run through a leverage-like denominator. A 0.2-point gap at long odds becomes a big-looking dollar figure. The probability is what reflects belief; the payout framing is what sells affiliate clicks.

None of this means the aggregators are lying on purpose. It means measuring divergence is a skill, and the dramatic version usually fails the most basic control: take both prices at the same instant, the same way.

But the gaps are real — they just live in the corners

Here is the part the contrarian over-correction gets wrong: cross-venue divergence is real. It simply does not live where the headlines put it.

The most rigorous look at this to date is a January 2026 study by Jonas Gebele and Florian Matthes, which built the first human-validated, cross-platform dataset of aligned prediction markets — over 100,000 events across ten major venues from 2018 to 2025 (arXiv:2601.01706). Their findings map almost exactly onto what our snapshot shows:

• Only about 6% of events are concurrently listed across platforms at all — most markets are not even comparable.
• Where they are comparable, semantically equivalent markets show persistent execution-aware price deviations of 2-4% on average, “even in highly liquid and information-rich settings.”
• Crucially, these are “driven by structural frictions rather than informational disagreement.”

So the venues genuinely fail to converge — on average. The reconciliation with our near-zero World Cup gaps is the whole point: the 2-4% is an average across a vast range of markets, many of them thin or semantically fuzzy, and the World Cup winner is one of the most liquid and least ambiguous contracts in the entire ecosystem. It sits at the tight end of that distribution. The deviations the literature documents concentrate exactly where liquidity thins and the question gets blurry — the mid-table teams, the obscure props, the markets only one crowd cares about. That is not the favorites. That is the corners.

For a forecasting product, that distinction is everything. The liquid top is where the market is smartest and an edge is least likely; the durable edge lives in the less-liquid corners where audiences diverge and arbitrage cannot reach.

Why segmented venues can’t fully close the gap

The reason small deviations persist instead of being instantly arbitraged away is one of the most durable results in finance: the limits of arbitrage. Shleifer and Vishny showed that textbook “riskless” arbitrage is a fiction — in reality it “requires capital, and is typically risky,” is carried out by a small number of specialized players, and “becomes ineffective” precisely when prices diverge (Shleifer & Vishny, Journal of Finance 1997). Across prediction-market venues, those frictions are unusually high:

• You can’t net positions across venues. Buying a team cheap on one platform and selling it dear on the other does not offset into a clean, capital-free hedge. As the Gebele-Matthes work puts it, positions “cannot be netted across venues and must be held until resolution.” Your capital is tied up on both sides until July.
• Different rails, different money. Kalshi is US-dollar-native, funded by bank transfer and debit, under CFTC exchange rules (Kalshi Help). Polymarket’s global app runs on USDC on Polygon and resolves through the decentralized UMA optimistic oracle. Moving capital between a dollar exchange and an on-chain book is not instant or free, so the trade that would close a gap is itself costly.
• Fees and the spread. A 0.3-point “gap” is gone the moment you pay taker fees on both venues and cross two bid-ask spreads. Most apparent edges at the top are smaller than their own transaction costs.
• Different audiences. The two crowds are not the same crowd. Per Pew Research, Kalshi’s volume is roughly 80% sports with thin politics, while Polymarket skews global and diversified — about 39% sports, 32% politics, 20% crypto — and its international app dwarfs its US one ($9B vs $1.3B in April 2026). Different order flow incorporates different information at different speeds, which is precisely how small, real gaps open in the first place — and why they can briefly widen around news that reaches one audience before the other.

Put together: at the liquid top, shared information plus enough arbitrage-like pressure keeps the venues nearly identical; in the corners, thin liquidity plus expensive, unnettable arbitrage lets platform-local beliefs survive. That is the structural picture the data supports.

How to read divergence honestly

If you want to compare two venues without manufacturing a gap, the discipline is simple and it is the same discipline we hold ourselves to:

1. Same timestamp. Capture both prices as close to simultaneously as you can. If you can’t, say so explicitly — a time gap on a live market is a confound, not a spread.
2. Same price point. Midpoint to midpoint, or bid to bid — never one venue’s bid against the other’s ask.
3. Check the bid-ask width. If a book shows a 0c bid and a 1c ask, there is no meaningful “price” to compare. Ignore midpoints on near-empty books.
4. Net out costs before calling it an opportunity. Subtract fees on both sides and both spreads. A gap smaller than its transaction costs is not edge.
5. Distrust payout framing. “$90 on $100” is a probability difference in a costume. Convert back to the implied probability and judge the gap there.
6. Treat bigger relative gaps in thin markets as noise first, signal second. The wider the spread, the less any single number means — on both venues.

Run that checklist on the World Cup favorites and you get our table: agreement to a fraction of a point. Run it on the thin corners and you find the real, structural deviations the literature describes — the ones worth studying.

Where this points

The takeaway is not “the venues are identical,” and it is not “there are huge gaps to exploit.” It is the precise middle: the liquid top of a prediction market is genuinely efficient, and the durable mispricing lives in the less-liquid corners where audiences diverge and arbitrage is too costly to discipline the price. That is a more useful map than either the hype or the cynicism.

It is also exactly the territory MispriceHQ is built to study — the gaps between a market’s price and a model’s estimate of fair value, concentrated where the crowd is thinnest. To be explicit and honest, as always: our machine-learning fair-value engine is in development. It has no live track record and has resolved no markets, and we will publish its methodology and performance only once it is validated. If you are new to the World Cup market itself, start with how its winner odds work and how those odds compare to a sportsbook. For how we think about turning a price into a probability you can test, see reading implied probability from market prices; for the broader landscape, what prediction markets are; and for how we approach fair value, our methodology.

Until then, the most valuable thing we can do is measure carefully and report what we actually find — even when what we find is that the favorites agree, and the headline gap was never really there.

Reminder: Event contracts carry a real risk of loss. The prices in this article are a single live snapshot from June 2, 2026 and will have changed. This is research and education, not financial or betting advice.