Credit, EPA/Shutterstock

The result of the polls that gave victory to President Luiz Inácio Lula da Silva (PT) was recognized even by the Liberal Party (PL), the party of the defeated candidate, Senator Flávio Bolsonaro.

Despite this, narratives circulating on social media that the result would be "mathematically impossible" and insinuations of unproven fraud.

BBC News Brasil interviewed political scientists and data analysts to explain what made Lula's unprecedented victory possible — it was the first time since redemocratization, when the presidential election began to take place in two rounds, that the loser in the first round of voting managed to win the second stage.

Those interviewed also contested arguments circulating among Bolsonarists that extending the vote until 7pm would have allowed some irregularity in favor of Lula.

The closing of the polls was scheduled for 5pm, but was extended by decision of STF minister Cristiano Zanin following requests from the Lula and Flávio Bolsonaro campaigns, due to difficulties faced by voters in accessing polling places, whether due to the lack of public transport in some cities or due to weather reasons.

The thesis that the extension would have affected the legitimacy of the election was refuted by data from the Superior Electoral Court (TSE) itself, a Court presided over by minister Kássio Nunes Marques, appointed to the STF in 2020 by then-president Jair Bolsonaro.

According to electronic voting records, only 34,393 votes were registered after 5 pm, a much smaller number than the advantage obtained by Lula of 4.4 million voters.

"On Sunday (25/10), the most inclusive elections in the history of the Brazilian Republic were held. Even in the face of instability in the final stretch of voting, thanks to the measures adopted and the competence of the Court's technicians, it was possible to move forward with the investigation and deliver the election safely", declared Nunes Marques in a note on Wednesday (28/10).

End of Most read

Watch news from Brazil and the world on the BBC World Service

Download here

End of app promotion

The statement came shortly after Flávio Bolsonaro released a statement claiming to be the "true elected president". Afterwards, however, the senator retreated and released a new statement declaring himself the "leader of Brazil".

In his note, Nunes Marques also said that "all questions raised within due legal process will be analyzed and answered." This statement does not refer to disputes over the result of the vote, but to complaints that each presidential campaign raised against the opponent about alleged illegalities in the electoral race.

Flávio Bolsonaro accuses Lula, for example, of having used the public machine for electoral purposes, while the PT member says that the senator was favored by Meta, a group that controls Facebook, Instagram and WhatsApp.

The various complaints will still be analyzed by the TSE — if any illegality is proven serious enough to interfere with the integrity of the elections, the election may be annulled by the Electoral Court.

Understand below the mathematics behind Lula's election and what experts say about the suspicions of fraud raised without evidence by Bolsonaro supporters.

Where did the extra votes for Lula and Flávio come from in the 2nd round?

The 2026 elections had 158,745,502 people eligible to vote. To understand Lula's victory, let's break down how these voters behaved in the first and second rounds.

On October 4, about 125.2 million showed up to vote, while about 33.5 million did not go to the polls.

Of those who voted, 56.104 million chose Flávio Bolsonaro (47.03% of valid votes), while another 53.879 million preferred Lula (45.16% of valid votes). The other candidates were well behind and totaled 9.316 million.

Blank and invalid votes, which are not included in the count of valid votes, totaled almost 6 million votes.

In other words, Flávio ended the first round with an advantage of around 2.2 million voters over Lula. The scenario was very challenging for the PT member for two reasons, political analysts pointed out.

The first is that a turnaround depended on attracting a significant part of the voters of the other presidential candidates, who, in theory, were mostly in the right-wing camp.

The second was the need to increase the participation of voters who did not opt ​​for any candidacy in the first round, reducing abstention and blank and invalid votes — historically, however, there had never been a strong drop in these groups between the first and second rounds.

As the vote is secret in Brazil, it is impossible to know how each voter changed or not changed their position between the two votes. But the balance of the second round shows that both phenomena occurred, notes political scientist Jairo Nicolau, professor at Fundação Getúlio Vargas.

In other words, a relevant part of voters for third-way candidates opted for Lula and there was an unprecedented increase in voter participation that also favored the PT member.

Furthermore, the political scientist believes that the PT member may have even turned over some votes that went to Flávio in the first round, not from voters convinced of Bolsonarism, but from those who supported the candidate on the eve of the vote and gave him a boost in the first round.

"This was an unstable vote. The vote of the Bolsonarista aunt who wears yellow could not be turned around", he highlights.

Let's look at the numbers from the second round. On October 25, around 128.2 million showed up to vote, that is, almost 3 million more than in the first round, which reduced abstention to 30.5 million.

Furthermore, blank and invalid votes fell by around 1.8 million to 4.1 million.

In other words, these two phenomena represented another 4.78 million valid votes to be disputed by Lula and Flávio. Adding this amount to the 9.316 million voters who opted for third-way candidates in the first round, we reach 14.09 million new votes that could be captured by Lula or Flávio.

And that's exactly what happened: adding the increase in votes for PL and PT candidates in the second round, both grew by 14.09 million in total. But it was Lula who won the majority.

The PT member managed to attract the support of 10.376 million voters, totaling 64.256 million (51.79% of those valid).

Flávio obtained only 3.7 million new voters, finishing behind the PT member, with a total vote of 59.8 million (48.21%).

These numbers only reveal the total balance of voter migration, highlights Jairo Nicolau. In practice, notes the political scientist, it may have happened that part of the third-way voters migrated to the null vote in the second round. On the other hand, a larger portion of those who canceled in the first round may have opted for Flávio or Lula in relation to what the balance of the drop in voids reveals, for example.

The numbers released by the TSE, however, do not allow us to detail this flow of voters due to the secrecy of the votes.

For Nicolau, the migration of a relevant part of the electorate from candidates such as Ronaldo Caiado (PSD), Renan Santos (Missão) and Romeu Zema (Novo) to Lula occurred because there is a portion of the Brazilian voter that is quite fickle and moves between right-wing and left-wing candidates.

This had already happened in 2022, when Lula won, recovering part of the voters who had given up voting for the PT in 2018 and opted for Jair Bolsonaro that year.

"The mathematics of Brazilian politics is not ideological mathematics. People forget how uninformed Brazilian voters are, how depoliticized they are. A large part of them are not guided by ideology. So, there is a contingent of voters in Brazil who fluctuate, who are not that faithful," he states.

Nicolau notes that Lula's performance in the 2026 second round (51.79%) was less than a percentage point above his performance in 2022 (50.9%). In his assessment, what stands out was how the PT member weakened in the first round.

"Basically, Lula recovered what he lost [in the first round]. He simply returned to his level. And Flávio became more or less the size of his father [in 2022]. The noise of the campaign gives the impression that the result was very different. It wasn't. There were different campaigns for similar results [in 2026 and 2022]."

In his view, the president's growth in the second round reflects a more energized campaign, after the scare of the first vote, both on social media, with a coordinated action through the Lula's Army channel, organized by deputy André Janones (PT-MG), and on the streets, with the carrying out of large events across the country by supporters of the PT member.

"I'm talking about the biggest rallies and street mobilizations that the country has ever seen in an electoral campaign. There's no equal, there wasn't anything like it in 1989, there's nothing like what happened there in those three weeks."

For Pedro Burgos, professor at Insper and consultant in Artificial Intelligence, there was an organic movement of Lula supporters working to turn votes between the first and second rounds — something that needs to be better investigated with qualitative research, in his view.

"I started running a little robot to capture testimonials on social media with the words 'I became a vote' from 'Zema', 'Caiado', 'Cury'. And, as far as I saw, the proportion was 20 to one. In other words, for every twenty testimonials from people saying 'I became my mother's vote from Cury to Lula' there was one saying that it turned from Cury to Flávio", for example.

"The effort to turn the vote, especially among young people, was very big. What happened, and I think this is what the numbers [of the polls] have the most difficulty in capturing, is a lot of people trying to convince father, mother, mother-in-law, uncle, colleague one to one. I think this effect on the micro had a lot of impact", he ponders.

Credit, Reuters

Accusations of fraud

The report analyzed hundreds of videos on TikTok searching for terms linked to electoral fraud. Twelve of them raised suspicions about the result or endorsed accusations of fraud or scam. Together, they had just over 3 million views.

Among them were montages and excerpts from Flávio Bolsonaro's speeches in his first statement after the defeat, in which he called for the election to be annulled.

Some publications accused the Electoral Court of a coup by extending voting hours. Others questioned Lula's vote.

One of the videos with the most views stated that the PT member could only have reached the vote released by the TSE if he had won the majority of voters from the candidates eliminated in the first round (9.316 million).

This conclusion is false. Lula won around 10.4 million votes between rounds, but the count treats the voters of the eliminated candidates as the only possible source of these votes.

Voters who abstained or voted blank or null in the first round could also vote for Lula in the second. And abstention fell from 21.08% to 19.2% of the electorate. The totals for each round do not allow us to know how each voter voted in the two rounds.

How the extension of the polls was used to attack the legitimacy of the election

Among the content that talks about fraud in the election, a document entitled "Independent Expert Report" began to circulate on Bolsonaro profiles with accusations of fraud in the presidential contest.

According to the document based on false conclusions, the PT victory was "fabricated" at the polls that closed after 5pm.

The report was denied by the TSE. In a note, the Electoral Court says that the report "distorts data to suggest fraud in the elections."

One of the document's arguments is that the polls closed by 5pm (30% of the total) had a majority of votes for Flávio Bolsonaro. In these sessions, says the report, the PL candidate obtained 52.39% of the valid votes, opening an advantage of almost 1.8 million votes over Lula.

In the polls that closed after 5pm (70% of the total), the document continues, "Lula obtained an atypical explosion of more than 6.1 million votes, enough to annul the opposition leadership and turn the national result around."

According to the TSE, the first data that challenges this conclusion is that only 34,393 votes were cast after 5pm. In other words, the ballot boxes that gave Lula victory recorded the vast majority of votes within normal voting hours.

The Court also emphasizes that there is nothing unusual about registering votes after 5 pm.

Even if there was no extension of the vote by decision of Minister Zanin, the electoral rules allow voters who were already in line for their session until the normal closing time (5 pm) to vote.

Because of this, 31,714 people also voted after 5pm in the second round of 2022, when there was no extension of the time.

Statistician Rafael Izbicki, professor at the Federal University of São Carlos, also analyzed the polls' closing times and dismantled the accusations of fraud.

He points out that, of the 348,000 ballot boxes that closed after 5pm in the second round of 2026, the vast majority (340,000) closed by 5:15pm, possibly because of people who were still in line at 5pm. Only 8,814 sessions (1.8% of the total) closed as of 5:16 pm.

In the case of the polls that closed until 5pm, Izbicki concluded that these electoral sessions gave a majority of votes to Flávio not because of the closing time, but because of the location in which they were, especially in states where the senator had a more expressive vote.

This happened, for example, in six states whose voting ends before 5pm local time due to the time difference in relation to Brasília — Acre, Amazonas, Mato Grosso, Mato Grosso do Sul, Rondônia and Roraima.

In these states, the polls officially close at 3 or 4 pm. Therefore, even when there was an extension in these locations, practically all election sessions closed before 5pm. Of these six states, Lula only won in Amazonas.

"For this reason, practically all 32 thousand sections of these States fall into the 'until 5pm' group. These six States [combined] voted more for Flávio: 59.7% of valid votes. They account for 1.44 million of the 1.75 million votes of Flávio's advantage in the 'until 5pm' group", notes Izbicki, when analyzing the Bolsonarist report.

At the request of BBC News Brasil, the professor also analyzed where the other polls were that closed until 5pm and gave, on average, a favorable vote for Flávio.

According to his survey, the majority of sessions closed by 5pm, outside of those six states with a different time zone, were in nine other states in which Flávio won.

In other words, of the 30 million voters who voted in polls that closed by 5pm in States with the same time zone as Brasília, 16.8 million (56.2%) were in São Paulo, Rio Grande do Sul, Paraná, Santa Catarina, Goiás, Rio de Janeiro, Espírito Santo, Tocantins and the Federal District

Izbicki also analyzed how Lula and Flávio's votes went in the first round in the two groups of ballot boxes created in the Bolsonarist report: those that closed until 5pm in the second round and those that remained open beyond that.

The data shows that Lula had already performed better on October 4th in the set of polls that closed later in the second round, even without voting being extended on that occasion.

"The sections that closed after 5pm already gave Lula six points more in the first round compared to the percentage obtained by the PT member in those that closed until 5pm", highlights Izbicki.

His survey also showed that the PT party had similar growth between the first and second round in the polls that closed until 5pm or later.

"From the first to the second round, Lula grew practically the same in both groups. Furthermore, the votes registered after 5pm would not change the winner, even if they were all for Lula", he highlights.

"The polls that closed until 5pm voted more for Flávio because of where they are, not because of the extension", reinforces the professor.

Accusations of fraud in votes without biometrics are also contested

The report with false allegations of fraud raises other questions about the election results.

One of them treats as suspicious the fact that 8.6 million voters were allowed to vote without authorization through biometrics.

The report highlights that the number is "almost double the total difference" in votes between Lula and Flávio (4.4 million).

The TSE disputes any irregularity. According to the Court, when the enablement of voting by biometrics fails, there are other safe mechanisms to authorize the voter to vote.

The procedure consists of confirming the voter's year of birth, "data that needs to be provided by the voter himself, as it is not available in any document to which the poll worker has access." Furthermore, states the TSE, there is a signature on the voting book and the identification of the responsible poll worker.

The Court also emphasizes that it is impossible to identify how a qualified voter voted without biometrics, as these votes are also secret.

The analysis carried out by Rafael Izbicki reveals other weaknesses in the argument.

According to its survey, the proportion of voters who voted without biometrics was practically stable in the two rounds this year — there were 8.3 million on October 4th (6.66% of the total) against 8.6 million (6.73% of the total) on October 25th — and fell compared to the second round of 2022 (10 million or 8% of the total).

The data analyzed by Izbicki also shows that Lula grew more in the polls with fewer votes qualified without biometrics.

"The report claims to have tested the correlation between these qualifications and Lula's vote, but it does not show the result. If they had been used to add votes to Lula, it would be expected that he would grow more between rounds in the sections with more qualifications without biometrics", notes Izbicki.

"The data shows the opposite: where less than 6% of voters were qualified like this, Lula grew by around 3 points; where there were more than 15%, 2 points or less", he further said.

Graphics by Caroline Souza and Laís Alegretti, from the Visual Journalism team at BBC News Brasil

Discover more

Watch

End of Watch

How US elections should seal Trump's fate and the future of the relationship with Brazil in Lula's 4th term

'Bolsonarism is not defeated, it is widespread. Just see the bosses firing employees who voted for Lula', says philosopher Renato Lessa

Doctors are accused of persecuting colleagues and suggesting a boycott of patients for political reasons: 'I felt a target on my head'

Top news

LIVE, US elections: economy is Americans' main concern when voting, survey shows; first polls close soon; follow

When will the results of the US midterm elections be released?

Lula's victory explained mathematically

read more

Most clicked