Brazil runoff poll average, Oct. 20: Flávio 51.5%, Lula 48.5%

RaisLife's average of Brazil's presidential runoff polls puts Flávio Bolsonaro at 51.5% of valid votes and Lula at 48.5%, as of Oct. 20. Valid votes exclude blank and spoiled ballots. The figure averages 24 polls covering 72.790 respondents and was calculated by RaisLife; it is not a registered poll. The gap is 2.9 percentage points, within twice the polls' average margin of error (±1.7 points). Flávio stood at 51.4% in the average through Oct. 19, a change of +0.1 point.

The average includes polls registered with Brazil's Superior Electoral Court (TSE) and released after the first round on Oct. 4, weighted by sample size. One new runoff poll came out on Oct. 20, from Real Time Big Data: it put Flávio at 52.0% and Lula at 48.0% of valid votes, with 2.000 respondents and a margin of error of ±2.0 points (TSE registration number BR-02952/2026). Its details and sources are below.

New poll today

• Real Time Big Data (TSE registration number BR-02952/2026): Flávio 52.0%, Lula 48.0% of valid votes (of all respondents: Flávio 48%, Lula 45%). 2.000 respondents, margin of error ±2.0 points, conducted Oct. 15 to 19. This is a statistical tie by RaisLife's standard. Sources: Gazeta do Povo, Exame.

What the runoff polls show

Twenty-one polls have Flávio numerically ahead (13 of them statistical ties) and two have Lula ahead (both statistical ties). Five days remain until the runoff. Polls differ in method, sample and fieldwork dates, which is why RaisLife lays out all of them, by date and by pollster, rather than picking one. The full numbers and their history are on the presidential polls page.

Which pollsters are in the average

Today's average draws on polls registered with the TSE from AtlasIntel, CNT/MDA, DataTrends, Datafolha, Futura/100% Cidades, Gerp, Indexa/Broadcast, Ipespe/Abrapel, Nexus/BTG, PoderData/Aya, Quaest, Real Time Big Data, Veritá and Vox Brasil. All were released after the first round and each was reported by two different news outlets. Every pollster uses its own method and sample, so the numbers differ; the presidential polls page shows each pollster's series. An AtlasIntel poll conducted Oct. 3 to 8, partly before the first round on Oct. 4, is included because it was released afterward and accounts for 7% of the respondents in the average; without it, the average would be Flávio 51.4% and Lula 48.6%. Vox Brasil and Quaest published figures based on all respondents; the conversion to valid votes was calculated by someone other than the pollster: Poder360 for Vox Brasil and RaisLife for Quaest. For results and vote counting on election day, see the 2026 elections page.

Flávio Bolsonaro's campaign: what it could do to win

Bruno Rizzi, a political scientist at Fatto Inteligência Política, said in an Agência do Rádio report published by Brasil 61 that the main task for Flávio Bolsonaro's campaign is to keep up the strategies it has followed so far. He projected that the candidate would absorb most of the votes of right-wing voters whose candidates lost in the first round. (Brasil 61, Oct. 6)

Felipe Nunes, director of the Quaest polling firm, told the newspaper Estadão, in an interview republished by TNH1, that Flávio Bolsonaro needs to win over voters who backed Augusto Cury, Ronaldo Caiado and Renan Santos. According to Nunes, that requires making democratic, moderate commitments and avoiding radical gestures. (TNH1, Oct. 10)

Luciana Santana, a political scientist at the Federal University of Alagoas, wrote in a Movimento Econômico piece that Flávio starts the runoff with an advantage but that room for change remains. She pointed to anti-PT sentiment, meaning rejection of Lula's Workers' Party, among Renan Santos voters and called the way Augusto Cury's votes will split uncertain. (Movimento Econômico, Oct. 11)

Lula's campaign: what it could do to win

Leandro Gabiati, director of Dominium, said in a Brasil 61 report that Lula needs to turn the runoff into a plebiscite-style comparison between the two candidates, since, in his assessment, a defeated candidate's formal endorsement of another does not automatically transfer votes. (Brasil 61, Oct. 6)

Luciana Chong, director of the Datafolha polling firm, told Estadão, in an interview republished by TNH1, that reversing the first-round result is a major challenge for Lula. She said he would need to mobilize voters who stayed home and win the backing of politicians in states where he lacks a local political base, such as Minas Gerais and São Paulo. (TNH1, Oct. 10)

Arthur Leandro, of the University of Pernambuco, wrote in Movimento Econômico that abstention is uneven across social groups, and that targeting lower-income, less-educated voters would be a rational move for Lula's campaign. He said the runoff will be decided by a combination of several sources of votes. (Movimento Econômico, Oct. 11)

What comes next

The runoff is on Sunday, Oct. 25, five days away. Exit polls may be released only after voting ends. Readers who want to adjust the first-round numbers and build scenarios for the governorships still in play can use the governor runoff simulator. The official result is the TSE's.

A poll captures voters' preferences at the time it was taken and carries a margin of error. The average and the statistical-tie standard are RaisLife calculations. Information checked against the sources cited. Updated Oct. 20.

Text by Rai, RaisLife's virtual reporter (artificial intelligence), with automated editorial review.