Placar 2026

On the night of October 25, this dashboard will project the result from the first ballot boxes

Starting at 5 p.m., the dashboard reads the TSE results CDN every minute and compares each municipality and zone with its own first round. With about 2% of polling stations counted, the projection stabilizes; in the 2022 replay it called the winner at 6:00 p.m., with 28% counted, while the TSE partial count still showed the loser ahead.

Lula's headline chance went from 40% on Sep 17 to 30% on the eve of the first round, fell to 17% with the counted vote and stands at 18%

In one sentence: 80 model runs have been published since Sep 17. After the first round, Lula's headline chance ranged between 11% (Oct 13) and 21% (Oct 22); polls only, without the market, it is at 17% today, and Polymarket at 20%.

Since the first round, 57 polls from 19 pollsters: the aggregate went from Flávio +2.4 to Flávio +0.9 in valid votes

In one sentence: the summary line combines all polls with each pollster's correction and today stands at Lula 49.5% × Flávio 50.5% of valid votes (Flávio +0.9, ±1.3 on the margin), vs. Flávio +2.4 on Oct 7 — 1.5 points toward Lula. In the simple average of recent polls, without correction, the margin is Flávio +2.0: the pollsters that polled most in this phase show Flávio doing better than those with a calibrated track record.

Runoff polls are moving toward Lula: 1.2 points per week

In one sentence: with the 57 polls conducted after Oct 4, the Lula − Flávio margin is moving +1.2 points per week (±0.4; a strong signal, with the prior that holds the line steady while there are few polls). Three days ago the reading was +1.4 (45 polls); it is election day: Lula goes to the ballot box starting from −0.9 at poll level, and what decides it now is poll error, not pace.

Daily trend-pointer reading: Oct 8: +0.0 · Oct 9: +0.0 · Oct 10: +0.0 · Oct 11: +0.0 · Oct 12: +1.7 · Oct 13: +1.7 · Oct 14: +1.8 · Oct 15: +2.7 · Oct 16: +3.0 · Oct 17: +2.4 · Oct 18: +2.0 · Oct 19: +2.1 · Oct 20: +1.8 · Oct 21: +1.7 · Oct 22: +1.4 · Oct 23: +1.3 · Oct 24: +1.2 · Oct 25: +1.2. With few polls the slope is sensitive to each new wave; the uncertainty band says by how much.

Wave by wave, runoff polls have stopped moving: +0.4 point per week within each pollster

In one sentence: across the 30 differences between consecutive waves in 16 series with fieldwork after Oct 4, the Lula − Flávio margin (valid votes) moves +0.4 point per week (±0.9; sign indistinguishable from zero), with more weight on the newest difference and on the one with the smallest error. Movement slowing down (long window +0.8, short +0.4). 17 of 30 differences toward Lula, 8 toward Flávio, 5 flat. In the literal version, using only the latest difference for each of the 16 pollsters, the average is −0.0 per wave (+0.0 with more weight on the newest).

Daily aggregate (pp per week, using the differences that existed on each day): Oct 16: +3.1 · Oct 17: +1.8 · Oct 18: +1.7 · Oct 19: +1.7 · Oct 20: +1.2 · Oct 21: +1.2 · Oct 22: +0.8 · Oct 23: +0.5 · Oct 24: +0.4 · Oct 25: +0.4. It is election day: the campaign is over and the pace no longer moves the margin.

In one sentence: the gap "rejection of Flávio − rejection of Lula" moves +1.0 point per week (±1.0) across the 14 differences in 8 series with rejection measured after Oct 4; positive favors Lula. Rejection tends to move ahead of the vote: it measures the ceiling, not the choice. Available room (people who neither vote for the candidate nor reject him) in the latest wave from the most recent pollsters: Datafolha Oct 24: Lula 7%, Flávio 7%; Futura (v2) Oct 24: Lula 6%, Flávio 5%; Alfa Oct 22: Lula 9%, Flávio 6%; Ipespe Oct 22: Lula 6%, Flávio 4%; Vox Brasil Oct 22: Lula 10%, Flávio 9%; Gerp Oct 21: Lula 6%, Flávio 4%. Excluded from the base: Real Time Big Data asks "whom would you not vote for" with an "I reject both" option (28% on Oct 15–19), which is not comparable to the others.

Between 10 and 14 million votes are still in play, 8 times Flávio's current lead; the contest is concentrated among women, high-school-educated voters and pardos

In one sentence: averaging the latest wave from 19 pollsters with fieldwork between Oct 19 and Oct 24, 2.0% are undecided and 4.0% vote blank or spoiled; adding the share who say they could still change (PoderData and Quaest), 7.9% to 11.2% of those who will vote are in play, 10 to 14 million voters, against a gap of ~1.5 million valid votes in Flávio's favor in the model's adjusted margin (−1.3 pp). The segments where neither candidate leads by more than 6 points in valid votes (Datafolha Oct 22): women (49 × 44); high school (44 × 49); pardos (46 × 47); 60 or older (49 × 44); 45 to 59 (47 × 48).

What has moved since the first runoff wave (Datafolha Oct 7 → Oct 22, valid-vote margin): toward Lula among 16 to 24-year-olds (−8 → +6), high school (−15 → −5), Northeast (+27 → +37), Center-West/North (−22 → −13); toward Flávio among 35 to 44-year-olds (−4 → −12), college (−3 → −9). Lula's largest net gains come from the Northeast (+12.0M), Catholics (+8.8M), up to 2 minimum wages (+8.6M); Flávio's, from evangelicals (+16.2M), 2 to 5 minimum wages (+6.2M), Southeast (+5.2M) (net = segment size × Lula − Flávio difference × first-round turnout; the breakdowns overlap and do not add up).

How many voters chose another candidate, voted blank or spoiled, or did not vote on Oct 4 (TSE), and what they say now (Lula × Flávio, with blank/spoiled and undecided; margins of error of 6 to 24 points in these breakdowns). The net multiplies the size by the side that takes more, without discounting those who will not vote; for those who stayed home on Oct 4 there is no net, because abstention tends to repeat between rounds (20.9% and 20.6% in 2022) and only part of it swaps places with those who will stop showing up.

If each voter turns out with the probability they report (certain = 1, maybe = ½), Flávio's voters turn out at 98.0% and Lula's at 96.5%: the valid-vote margin would go from −1.8 to −2.6 (−0.8 pp). Certainty about voting is lower precisely in the groups where Lula does best: 90% among 16 to 24-year-olds and 93% among those earning up to 2 minimum wages, versus 97% among those 45 or older (PoderData Oct 19–21). Among those who did not vote in the first round, only 69% say they will definitely vote; among those currently voting blank or spoiled, 56%. That is why Sunday mobilization matters more for Lula than for Flávio, and abstention in the 2022 runoff (20.6%) stayed close to the first round's (20.9%): it rarely changes much between rounds.

• The tied groups decide: women, high school, pardos, 60 or older, 45 to 59, 25 to 34 and 16 to 24 are within ±6 points in valid votes.
• Non-aligned voters (neither PT supporters nor Bolsonaro supporters) are ~16% of the electorate and 28% of them have no candidate (Datafolha); in Quaest, independents are 31% undecided in the unprompted question.
• Potential: 5% say they could vote for Lula and 5% for Flávio without him being their only option (PoderData): these are the voters still listening to both.

Four scenarios for October 25: the most likely is a narrow Flávio win, in 57% of simulations; Lula wins in 18%

In one sentence: today's adjusted margin (Flávio +1.3) falls in the "Narrow Flávio win" card; adding the polls' historical error and the uncertainty until Oct 25, 82% of simulations end with Flávio ahead. Polls taken after the first round, in valid votes: Futura (education weighted by PNAD) Oct 24, Flávio +0.3; PoderData Oct 24, Flávio +2.1; BTG/Nexus Oct 24, Flávio +1.1; Quaest Oct 24, Flávio +1.1; Datafolha Oct 24, Flávio +1.1; Palver (v3 calibration) Oct 24, Flávio +1.0; Gerp Oct 24, Flávio +3.2; Colectta Oct 23, Lula +0.5 (the 8 most recent of 57).