A Quaest poll published August 15 put Lula at 43% and Flavio Bolsonaro at 40% in a simulated runoff - a two-point gap inside the poll's own margin of error, and tighter than the 44-39 spread Quaest found ten days earlier. Polymarket's "Brazil Presidential Election" market, which pays out on whoever actually wins the presidency, tells a different story: as of 2026-08-16 07:08 UTC, Lula's contract is at 65.5c (best bid 65c on $44,168, best ask 66c on $1,233) and Flavio Bolsonaro's is at 28.75c (best bid 28.7c on $6,103, best ask 28.8c on $5,118). A runoff that pollsters call a coin flip is still priced more than 2-to-1 in Lula's favor to win outright. Numbers below are from the Gamma and CLOB APIs, fetched at that time.
01A tightening runoff, on paperWhat the poll actually found
Quaest surveyed 2,004 people for Globo between August 10 and 13 and published the results August 15. In a simulated second-round matchup, Lula led Flavio Bolsonaro 43% to 40% - inside the poll's 2-point margin of error, and down from a 44-39 lead in Quaest's prior release on August 5. In the first-round scenario, the same poll had Lula at 38% and Flavio at 31%, with three other right-of-center candidates splitting the rest: Renan Santos at 4%, Ronaldo Caiado at 4%, and Romeu Zema at 2%. Flavio became the right's standard-bearer only because his father can't run: Brazil's electoral court (TSE) ruled Jair Bolsonaro ineligible for eight years in June 2023 over misuse of state media to allege the 2022 election was rigged, and a separate Supreme Court case convicted him on coup-related charges in September 2025, sentencing him to more than 27 years. Flavio - a sitting senator - launched his own campaign July 25 in Sao Paulo with Alfredo Gaspar as his running mate. Brazil votes October 4; if no candidate clears 50% of valid votes, the top two go to a runoff on October 25.
02Two weeks of Polymarket pricesWhat actually moved
Polymarket's market resolves on who wins the presidency, including any second round, so it's the cleanest single number to compare against a runoff poll. Lula's contract spent the first half of August in a narrow 63.5-66.5c band, dipped to 63.5c for most of the week the poll was in the field, then climbed back to 65.5c on the evening of August 15 - the day the poll came out - and held there into August 16. Flavio's contract did move, but slowly: it climbed from about 25c on August 1 to 27.25c by August 10, then ticked up again during the poll's field dates (27.25c to 28.15-28.45c, August 10-13) and kept grinding higher after publication to 28.75c now. Over two weeks, that's a roughly 1.5-2c net rise on Flavio and a wash on Lula - a market drifting in the poll's direction, not repricing to match it.
WIN PROBABILITY, AUG 01 -> NOW (AUG 16 07:08 UTC)
Aug 01 Aug 10 Now
Lula 64.5c 63.5c 65.5c
Flavio Bols. 25.45c 27.15c 28.75c
Why the gap doesn't close just because the poll did
A simulated runoff asks 2,004 people to pick between two names today. Polymarket's contract has to price the whole path: does Flavio actually reach a runoff, and does he win it, months from now, against a Workers' Party incumbent with the advantages of office. The first round is the gating step, and there the same poll has Lula at 38% to Flavio's 31% - a 7-point gap, not 3 - with Renan Santos, Ronaldo Caiado, and Romeu Zema splitting another 10% of the vote between them. All three are real declared candidates in their own right (Caiado is Goias's outgoing governor, Zema is Minas Gerais's sitting governor, Santos co-founded the Free Brazil Movement), and all three held campaign launch events on August 16 - the same day this post is written. Every point one of them holds is a point Flavio doesn't have locked up for a two-man runoff. A poll question that assumes the runoff already happened skips past that fragmentation; a market that has to pay out on the actual result can't.
04The rest of the ballot confirms itWhat the other contracts say
Polymarket lists eighteen names in this market, and the pricing outside the top two matches the fragmented-right story. Jair Bolsonaro himself is at 0.45c - the market treats his ineligibility as settled, not a live question. Michelle Bolsonaro, his wife, is at 0.15c. Among the minor right-of-center governors, Renan Santos is the best-priced at 5.05c, Ronaldo Caiado at 0.6c, and Romeu Zema at 0.25c - low, but not zero, which is consistent with a first round where none of them individually threatens Flavio but their combined vote share does. On the left, Workers' Party figures like Fernando Haddad and Camilo Santana sit near 0.15-0.5c, pricing Lula as the incumbent party's only real contender. Every number on the ladder is compatible with the same read: a competitive but not even race, with Lula the clear favorite to win outright and Flavio the clear runner-up, exactly what 65.5c against 28.75c says before you even look at the poll.
05Tracking a crowded field from the terminalWatching eighteen contracts move together
A fragmented race like this one is hard to read from a single poll or a single contract - the story is in how eighteen prices move relative to each other as candidates launch, poll, and drop out. In polymarket-tui, searching "Brazil" pulls up the whole slate together with 24-hour price change and volume on each name, so a governor's campaign launch or a new poll shows up next to Lula and Flavio's numbers instead of in isolation.
$ uv tool install polymarket-tui
$ polymarket-tui
No account is needed to browse markets, books, and charts. The install page has Homebrew and one-liner options.