Brazil’s final presidential tally gives Luiz Inácio Lula da Silva a lead of more than 4.4 million votes, leaving a dispute over extended polling hours far removed from the scale of the recorded result. In an October 26 examination of election data, Folha reported that votes recorded after the customary closing time were too few to change the outcome.

The electoral court, known as the TSE, reported 64,256,232 votes for Lula, or 51.79% of valid votes, against 59,826,041, or 48.21%, for Senator Flávio Bolsonaro. Subtracting those official totals produces a margin of 4,430,191 votes. The percentages refer to valid candidate votes, rather than every eligible voter.

The result follows Sunday’s runoff and reverses the candidates’ first-round ordering. The Chamber’s account, updated October 26, confirms the final totals and describes Lula’s victory as an unprecedented fourth presidential mandate. Its first-round figures put Bolsonaro at 47.03% and Lula at 45.16%, showing why the runoff, rather than the opening vote, determined the winner.

The Chamber also reported abstention of 19.22% among eligible voters, placing the runoff among recent presidential contests with relatively low nonparticipation. Turnout, candidate vote shares and the victory margin answer different questions: participation measures who came to vote, while valid-vote percentages compare the choices made for the candidates. Mixing those denominators can exaggerate or obscure the size of a victory. The official figures therefore need to remain attached to their definitions when they are used in arguments about the election’s legitimacy.

What the extension actually covered

The underlying Supreme Court announcement describes a narrower action than reopening every polling station. Justice Cristiano Zanin extended voting until 7 p.m. in places where ballot machines had remained open after 5 p.m. The distinction is important when assessing claims about how many voters could have been affected.

The order arose from a complaint by Lula’s coalition concerning reduced public-transport services in Salvador, Feira de Santana and Lauro de Freitas. The court linked that complaint to an earlier ruling requiring free collective transport on election days. Its announcement also stated that both presidential candidacies had alleged transportation problems.

That primary explanation identifies the stated basis for the intervention. It does not mean the competing campaigns agreed about its scope or timing. Nor does an order extending access establish, by itself, how many additional people voted. Those are separate questions requiring election records and analysis, rather than inference from the announced closing time alone.

Operational criticism and the result

TSE President Kassio Nunes Marques acknowledged difficulties caused by the late decision. The court’s briefing said the change required adjustments to counting systems and procedures, together with additional security precautions for staff and polling workers. Marques warned that changes so close to closing could create hard-to-predict operational consequences.

Those criticisms should be distinguished from a finding that the result was altered. The Associated Press reported that Marques said the extension did not affect the result. AP reporters also observed polling places in Brasília and Rio de Janeiro closing shortly after the original deadline, illustrating that the formal extension did not produce identical operating hours everywhere.

Bolsonaro rejected the outcome on election night and objected to the timing of the change, AP reported. The distinction between that political objection and the announced count remains central: dissatisfaction with a procedure does not itself replace the electoral authority’s published totals.

What the late-vote analysis can show

Folha’s data examination counted 34,330 votes recorded after 5 p.m., using TSE records. That is less than one-hundredth of Lula’s margin. Even this broader late-vote category does not identify only people who voted because of the extension: it can include voters already waiting in line.

The newspaper described a second analysis using a later closing-time cutoff to estimate a smaller set of potentially affected votes. That approach relies on a methodological assumption about when ordinary queues would have cleared. TAQ has not independently reproduced either calculation. Their value is in testing the scale of the claim against recorded voting activity, with that limitation stated.

A separate allegation about observers

The electoral judiciary also issued a specific rebuttal to Bolsonaro’s claim that Liberal Party observers had been removed from the TSE’s secure room. It said that room was not open to visitors, while other areas, including the tabulation room, were accessible to authorized observers.

The statement said two party representatives attended the voting-machine integrity test and the tabulation and results presentation, with attendance records and video documenting their access. Folha separately reported the same official response. These accounts distinguish a restricted technical room from the spaces used for observing the count; they do not support treating the two as interchangeable.

For now, the verifiable national outcome is the final tally published by the electoral authority and confirmed by the Chamber. Procedural objections warrant evidence-specific scrutiny, but the available late-vote comparison and the court’s explanation do not establish that extended hours overturned the voters’ recorded choice.