A result announced, never shown
On the night of 28 July 2024, Venezuela's National Electoral Council (CNE) announced that Nicolás Maduro had won the presidential election with 51.2% of the vote to Edmundo González's 44.2%. Then it did something no electoral body in a verifiable election may do: it never published the results by polling station.
That omission is not a technicality. Venezuelan law requires the CNE to publish disaggregated results, and the electoral calendar scheduled audits to reconcile the announced totals against the physical record. The audits were cancelled. A second bulletin on 2 August adjusted the figures slightly — 51.95% to 43.18% — again with no station-level data. Weeks later, the government-aligned Supreme Court validated the announced result, citing no polling-station evidence in its ruling. The state's position came down to a number the public was asked to take on trust.
The actas: how citizens made it checkable anyway
Venezuelan voting machines print paper receipts, and each polling station records its count on an official tally sheet — the acta, the artifact the database's glossary records as a protocol. Accredited party witnesses are entitled to a signed copy of every acta. That design, a paper trail built into the system itself, is what saved the election's verifiability.
Anticipating that the CNE might not publish, the opposition organised its witnesses to collect and scan their copies. Within days it published the sheets online at resultadosconvzla.com — actas covering 81.7% of polling stations. Aggregated, they show the mirror image of the announced result: González with 67.1% (7,156,462 votes) to Maduro's 30.4% (3,241,461). The dataset was citizen-made, but its building blocks were the state's own official records.
What independent verification found
The decisive question was whether the published actas were genuine. The Carter Center — the only significant international technical mission present, whose statement two days after the vote already said the election cannot be considered democratic — examined them in its final report:
- Authenticity. The actas display the security features of genuine CNE-issued tally sheets.
- Incompatibility. The announced result is statistically implausible-to-impossible against the actas data. The sample covers over four-fifths of all stations; there is no plausible composition of the missing fifth that turns González's 67–30 lead into Maduro's 51–44 win.
- Tell-tale numbers. The first bulletin's percentages were suspiciously round — 51.2, 44.2, 4.6 — a statistical anomaly in their own right.
The UN Fact-Finding Mission, investigating the human-rights context, reached a parallel conclusion: the results-management process fell below basic transparency and integrity standards. By September, González had been forced into exile in Spain under threat of arrest.
The counter-claims: forged actas and a cyberattack
The government did not concede the record — it attacked it. Officials claimed the opposition's actas were forged, and blamed the failure to publish results on a cyberattack against the CNE's systems. Both claims follow the pattern this Academy teaches in how to fact-check an election claim: they were asserted without evidence, and no proof of either has ever been produced. The database rates the claim that the announced result reflects the actas as false.
Note what the counter-claims could not do: they could not produce the state's own station-level record. If the official actas supported the announced result, publishing them would end the argument. Their continued absence is the loudest fact in the case.
Reading the ledger: established, corroborated, contested
The election's ledger sorts the findings by how firmly each is established — the discipline at the heart of this database:
| Tier | Finding |
|---|---|
| Confirmed | Process violations: no disaggregated publication, cancelled audits, judicial validation without evidence (Carter Center, UN FFM) |
| Corroborated | Result falsification: verified actas and statistical analysis converge on an outcome opposite to the announced one |
| Contested | The government's forgery and cyberattack claims — asserted, never evidenced |
Why is falsification corroborated rather than confirmed? Because confirmation, in this taxonomy, comes from authoritative processes — courts, official investigations — and Venezuela's institutions are the ones under finding. What remains is still exceptionally strong: the physical record, independently verified, against an announcement its own authors will not substantiate. Corroborated is the honest tier, and honesty about tiers is what keeps the whole system credible.
The mirror image of a stolen-election myth
Set this case beside US 2020: how "stop the steal" was built and debunked and the two become each other's control group. In the US, a stolen-election narrative was attacked by every independent check — recounts, audits, sixty court cases — and collapsed. In Venezuela, the independent checks all point the other way, and it is the state's result that cannot survive contact with the record.
That contrast is the deepest lesson of this case file. "Rigged election" is not a vibe or a loyalty test; it is a claim that either survives verification or does not. The same instruments — paper trails, transparent counts, observer access, courts willing to look at evidence — are what expose a genuine theft and what exonerate a fair count. Societies rarely get a cleaner demonstration of what makes an election free and fair: not the absence of accusation, but the presence of a result you are allowed to check.