Arbitrage is trading on price differences across multiple platforms for the same product. Venue differences can create opportunities to seemingly guarantee a return, even if it’s slim. But it’s rarely useful for most traders, as margins are small, transaction fees can eat up profits, and it largely only makes financial sense at large-scale purchases where earning a penny or two on the dollar provides material payoffs. Also, platforms tend to share information with each other to flag accounts doing so. It’s not illegal, but it’s not appreciated either.
Nevertheless, it’s still a common practice. Certain websites track these opportunities for traders. On Tuesday, September 22, 2026, PicksByOdds tracked 123 cross-listed pairs and found the widest live splits at 6.5 percentage points: the Open section of the 46th FIDE Chess Olympiad and Ciro Gomes (running for Governor of Ceará in the 2026 Brazilian elections) winner market.
Kalshi priced the FIDE Olympiad hosts Uzbekistan to win at 60¢. Polymarket priced the same outcome at 54¢. Kalshi traded Gomes at 52¢ and Polymarket at 46¢. The margin therein represents an arbitrage opportunity.
The Olympiad Split After Six Rounds
The Open event is an 11-round Swiss-system tournament from September 16–27, with a rest day on September 22. After six rounds, the hosts, Uzbekistan, led with 12 match points. China and Armenia sat on 11.
Kalshi’s Olympiad market showed about $87,634 in volume, with the hosts hovering at 63%, China near 16%, and India near 11%. Polymarket’s matching contract had about $200,938 in volume, with the hosts near 58% and China near 13.8%.
Round 6 did the damage. The hosts beat defending champion India 2.5–1.5 after Nodirbek Yakubboev won on board three and the top boards drew. India dropped to 10 match points. Traders who marked off that pairing had a reason to elevate the Uzbekistan price. Traders who waited for a full rest-day recap had a reason to lag. Round coverage is in chess24’s Olympiad broadcast;
The rulebooks add some friction to the arbitrage opportunity. Kalshi cites Chess.com, FIDE, and Olympiad reporting as settlement sources, and the market can stay open until a champion is named or until mid-October. Polymarket points to the organizer at chessolympiad2026.fide.com and may use a consensus of credible reporting. Five rounds remain. Tiebreaks can still decide gold.
India’s week already had noise. ChessBase India reported a Round 1 scare when Arjun Erigaisi lost on board one to Thailand, but the team still won 3–1. The later direct loss to the hosts gave the Kalshi 60¢ price more support than the Polymarket 54¢ price.
The Same Gap on Gomes Versus the Incumbent
The second 6.5-point split sat on Gomes for Governor (Brazil) ahead of the October 4 first-round elections, with a possible runoff on October 25. Incumbent Elmano de Freitas ran the other way: 44¢ on Kalshi and 49¢ on Polymarket, a 5.5-point gap. Polymarket’s winner board showed about $121,937 in volume in one capture.
Polls don’t offer a unified view on the race. Ipsos-Ipec, via O Globo, had Gomes at 44% and Freitas at 33% in June. Datafolha, also on O Globo’s Pulso blog, later put Gomes at 52% and Freitas at 33%.
AtlasIntel tightened it to 49.3% versus 44.6%. Instituto Paraná, published September 18, had Gomes at 47.4% and Freitas at 39.9% among 1,352 voters. Those readings disagree, so two trading venues were never going to lock either, since published polls drive much of the prediction market implied-probability analysis.
Nevertheless, event market settlement is cleaner than the surveys: official counts from Brazil’s Superior Electoral Court. Temporary placeholders or suggested outcomes will not count. If the winner is still unknown by mid-2027, leftover names can go to “Other.”
Why Traders Should Not Treat Cross-Venue Spreads as Free Money
Fees, cash-transfer time, and position limits chew into the potential proceeds from a cross-venue spread. PicksByOdds publishes the mismatched pairs as a research feed, not a trading recommendation, and it refreshes about four times a day. The Tuesday read was stamped 1:03 p.m. ET. These spreads change constantly, and what looked like a lock for automatic profits can and will fade quickly. These are often opportunities only available to bots constantly scanning for quickly-appearing opportunities. Also, as previously mentioned, a couple of points of guaranteed return is still a guaranteed return, but unless you’re buying in at large volumes, which may not even be possible in these smaller markets, your nominal profits will be minimal
Volume and price both matter. Polymarket’s Olympiad book turned over more than twice Kalshi’s dollars in the snapshots above, yet Kalshi still showed the richer Uzbekistan price. Gomes attracted tens of thousands of dollars on his own leg. That is enough activity to be real, but not enough to guarantee every pairing or poll clears in seconds on both sides.
Time zones are another practical issue related to individuals conducting arbitrage trading. The global book can move when overseas traders are awake, so it changes 24 hours a day. Again, this points to an activity perhaps best left to automated bots.
Rest day is the test. No games land on September 22, yet the title keeps trading. If the 6.5-point Olympiad gap closes overnight, the books were only staggered. If it holds into Round 7, the split is a genuine fight about remaining matchups and whether 12 match points after six rounds is as safe as it looks.
References
1. PicksByOdds cross-listed pairs
2. Kalshi 46th FIDE Chess Olympiad Open Tournament Winner
3. Polymarket 46th FIDE Chess Olympiad Open Tournament Winner
4. Chess.com Olympiad event info
5. Kalshi chess markets
6. ChessBase India Round 1 report
7. Polymarket gubernatorial winner market
8. Futuur elected governor contract
9. PolyInsider market snapshot
10. ClearMarket event record
12. PicksByOdds daily digest, September 22, 2026
