The truth about elections, held to a single standard.

Next up

Five categories, never conflated

The recorded numbers do not reflect the votes cast.

Unequal media access, repression, illegal financing, vote buying, misuse of state resources.

Cyberattacks, hack-and-leak, covert foreign financing, impersonation of media or institutions by a foreign state or proxy.

Troll farms, fake local-media networks, covert political ads, coordinated inauthentic amplification — foreign or domestic.

Patterns in the data worth examining — explicitly not proof of fraud.

A finding sits in exactly one category. Colors and icons carry the same meaning everywhere on the site.

From the case files

The clearest test of a trustworthy record is whether it can hold both statements at once.

Process violationsConfirmed
The provisional-results app displayed incorrect data for roughly four hours on election night.
More detail
The election authority's review described a human transfer error: the wrong dataset was copied from the results-management system into the public app. The app was taken offline and restored with corrected provisional data. Even though it was not the legal tally, the incident was a real communication and verification failure in a close election.
Sources:1 2

Result falsificationDebunked
The app anomaly did not provide evidence that officials re-tabulated or changed the official result.
More detail
The provisional display and the official count were separate processes. The MOG observed the official tabulation from original polling-station protocols, which used blind double entry, audit triggers and party access to records. It assessed speculation that the app incident created an opportunity to re-tabulate the official result as unfounded.
Sources:1
Statistical anomaliesDebunked
The abrupt change in the app's provisional trend was a corrected display error, not a vote-swing time series.
More detail
Provisional totals were reset when the wrong transferred dataset was removed. Treating the before-and-after screen values as successive batches of counted votes creates a false statistical series. The legal result came from a new final tally based on original protocols, not from continuing the erroneous app display.
Sources:1 2