Polling Station Capacity: How Fast Can People Really Vote?

Every polling station has a speed limit. When official turnout is higher than that limit, the votes could not have been cast the way the protocol says.

By Lyudvig Khachatryan · November 11, 2026

Imagine a polling station with one registration table. Each voter must show an ID, be found in the list, sign, and receive a ballot and an envelope. Even with no queue and no delays this takes time. Multiply that time by the number of voters the commission reports, and you get the time the station needed. If that time is longer than the hours the station was open, the reported turnout is physically impossible.

The capacity formula

The book defines the throughput of a polling station, ζ, as the largest number of voters it can serve in one hour. It is limited by the slowest of three steps: registration tables, voting booths and ballot boxes.

Here 3600 is the number of seconds in an hour, and each t is the shortest possible time one voter spends at that step. Two rules follow:

Using the three-hour turnout reports

Armenian law requires commissions to report turnout at 11:00, 14:00, 17:00 and 20:00. That gives four windows of three hours, 10,800 seconds each. Dividing 10,800 seconds by the number of new voters in a window gives the average time per voter. If that time is shorter than the minimum needed to register one person, the window is suspect.

For the monitoring, the minimum time to register one voter at one table was estimated at about 44 seconds. A station with one table therefore cannot register more than about 245 voters in three hours.

What happened in the 2015 referendum

Comparing the official turnout reports of the 2015 constitutional referendum with this limit, the monitoring found:

About 100 polling stations reported times close to or below the 44-second minimum. Some reported less than 20 seconds per voter while more than 250 people voted, or more than 500 at stations with two registration tables. At most of the stations with unlikely or almost impossible speeds, most votes were counted for YES.

There is one more sign that cannot be explained by speed at all: at 5 polling stations the reported number of participants went down from one three-hour report to the next. Turnout can only grow during the day, so these numbers were changed.

How to make the estimate stronger

The book suggests three levels of caution. The simplest uses the average minimum time to register one voter. A stricter one uses only the single shortest step. The most reliable one measures real queues: record when a group of voters enters and leaves the station and divide the time by the number of people. Even the most generous estimate leaves many stations above the limit.

The same check for 2017–2026

The monitoring pages compare turnout and speed for every polling station in the 2017, 2018, 2021 and 2026 elections, so you can see whether the pattern repeats or disappears.