Remote electronic voting (REV)

It has been with us since 2019, and this time the map was: 32 regions. List them.

Why these regions#

What is interesting is that the regions were chosen on purpose. Almost all of them had a low United Russia result at the election five years ago.

Interestingly, the region where the remote-voting experiment was first held in 2008 never introduced it in full. In 2026 not a single new region appeared on the list of those using remote electronic voting.

And in 4 regions that had used it before, they did not introduce it this year. A summary table by year and region is in the Google Sheets table.

Did everything go well in testing and voting#

No. Testing of the federal remote-voting system before the election showed errors, which Viktor Tolstoguzov wrote about on his channel. Even on voting day voters complained and forwarded screenshots of errors; they did not understand whether they had voted.

An entire TEC in Belgorod Oblast accidentally stopped voting#

On the first day of voting, observer Alexander Shukshin noticed in the official dump that the Rovenskaya TEC had published a finishVoting transaction: for voters across the whole district the remote-voting system stopped issuing ballots for single-mandate district No. 80 in the State Duma election. Only a third of the voters had managed to vote.

It is unknown whether the other two thirds who had signed up for remote voting but were in their precinct area on election days managed to vote, because after that the dump from the remote-voting site showed no removeFromVotersList transactions in the district — transactions that must appear when a voter casts a ballot at the polling place.

Loss of the vote due to network errors#

If you plot ballots issued minus ballots dropped into the electronic box for the State Duma single district, about 15 thousand ballots were never received by the system. Voters were probably deprived of the right to cast their vote, including because of the failures.

The CEC had adopted a resolution in advance allowing a person to come to a precinct and vote with a paper ballot even if they had signed up for electronic voting and had not withdrawn the application. Failure cases do not fit that rule, because the electronic ballot had already been issued, and an attempt to get a paper one must be refused, to prevent voting twice.

In Magadan the voting pages were blocked#

Magadan is the first time zone where voting starts. And that is exactly where the problem appeared at once. The number of electronic ballots that were issued but never landed in the electronic box grew.

Issuing ballots was broken. On the chart of ballots issued there was a huge rise: ballots were issued, but the system could not accept them. The ballot page would not load for users. The blocking system, the Technical Means of Countering Threats, simply would not let the voter onto the site to vote. It is roughly as if a voter were handed a ballot, but a police officer blocked the path to the booth and the box. And people left with the ballot in their pocket (in the browser).

In the end the organisers contacted the security agencies to turn the blocks off in the region. On 18 September in Magadan they were turned off. Twitter and Facebook loaded for people. It was hard not to notice. To solve the problem of voters being unable to vote, the security agencies temporarily lifted all the blocks. And the chart shows a drop: people who could not vote earlier came back in and voted. In the analogy with ordinary voting, those who tried the precinct again saw that the officer had left and they could vote normally. Not everyone came back. More than 3% ended up losing their ballot and were deprived of the right to vote.

A strange failure in the morning#

In the morning there was a strange failure in the number of ballots issued per unit of time. The curves for every region went down, as if the system periodically stopped working or, possibly, there was a failure in a boosting system. The activity curve was going down smoothly and naturally, and this drop was unnatural. It affected many regions. That is possible only with a centralised failure, yet there were no reports of anything like it.

The peak at 10:10, five time zones#

Synchronous; the cause is unclear, but it was centralised.

The peak at 16:10, two time zones#

At 16:10 a peak is visible in only two time zones: Moscow and Kaliningrad.

Likewise, the cause is unclear, but it was centralised.

A ban on voting#

Under the remote-voting procedure, ballots issued before 20:00 must be accepted until 20:15. But the blockchain shows a finishVoting record before 20:15, and two transactions were found in which a voter's ballot was rejected. The record contained FAIL_.

Moscow Oblast, State Duma, Dmitrov district 119

Perm Krai, Perm city duma

Acceptance of ballots was closed 15 minutes before that time, so voters could not drop the electronic ballot into the electronic box. Expert Viktor Tolstoguzov wrote about this in detail.

Quite a lot of election "contracts" on the blockchain were closed at 20:00, more than a thousand. One of the FAIL ballots was around 20:00, the second at 20:14:57, which is also before the official closing time.

The remote-voting TEC even adopted a separate decision to fix a protocol problem with control ratios that did not match (these 2 issued ballots were counted as valid, but they were not part of the votes, so the number of valid ballots did not equal the sum for candidates and parties). These 2 people also lost their vote, like the voters in Magadan.