How we test Where each number comes from

Server coverage

Once a day we read the server list each of seven providers publishes and count the locations per country. The result is a row per provider, per country, per day, with the collection date attached. We do not connect to those servers to check they answer — the count is what the provider claims, dated.

Speed

Speed runs are done by hand, not by a script: 14 cities, 46 paired runs so far, comparing NordVPN and Surfshark against the same raw line in the same session. Eleven of those cities also have a baseline measurement of the line itself, which is what lets us say how much of it survived.

We only call a winner in a city when we have five or more paired runs there and the gap is at least 10%. That threshold is met in 4 of the 14 cities. In the other ten we publish the numbers and say the sample is too small — we do not rank.

Prices

Prices are read from each provider's public pricing page daily. Where a provider does not publish a price in a currency, we do not convert one for them.

What none of this proves

That a provider lists a server in a country does not mean a streaming service will work through it, or that the country permits VPN use. We publish what we can check and mark the rest as unmeasured.

The files are at open data, under CC BY 4.0.