{"description":"How to read the evidence table behind a location's status: probes, networks, fastest reply, distance limits, discarded probes, and why many results stay open.","language":"en","latest_measurement":"2026-10-17","method":"latency-v3","method_page":"https://crackthe.net/method/","source":"https://crackthe.net/","text":"Key points\nEvery measured location has an evidence table: one row per probe, with its city, network, fastest reply, and what that reply rules out. A row \u0026ldquo;rules it out\u0026rdquo; when the probe is farther from the listed country than its reply allows. 346 of 1,126 measured locations have no conclusion yet. That is honest, not a failure. Where to find the evidence Open a provider page, then a country in its location table. The country page of that provider shows each measured location with three boxes (what the list says, the measurement, the status) and the evidence table.\nThe columns of the evidence table ColumnWhat it means Probethe city and country of the measuring computer Networkthe network the probe sits in Fastest replythe shortest round-trip time of its pings Endpoint is withinthe distance limit from that reply Distance from probe to stated countryhow far the probe is from the listed country; \u0026ldquo;in the country\u0026rdquo; if inside Result\u0026ldquo;rules it out\u0026rdquo;, \u0026ldquo;rules nothing out\u0026rdquo;, or \u0026ldquo;discarded\u0026rdquo; Reading a row 800 km Probe reply after 8 ms Listed country farther than 800 km: ruled out If the endpoint is within 800 km of the probe, and the listed country is farther than that from the probe, this row rules the country out. A row rules the country out when the distance to the country is larger than the limit, plus the margin for imprecise probe locations. One such row is an observation; it takes at least 2 independent rows, and for locations the provider does not mark as virtual a second day, before the status becomes \u0026ldquo;server outside the country\u0026rdquo;.\nWhy many results stay open Most probes sit in data centers in a limited number of cities. If no probe is near the listed country, every reply is slow, every limit is large, and nothing can be ruled out or confirmed. The result is \u0026ldquo;no conclusion possible\u0026rdquo;. Measurements continue, and a later run may add a closer probe.\nMeasured locations that match their country, by provider hide.me86 % IVPN78 % Mullvad75 % OVPN75 % AzireVPN69 % Windscribe68 % AirVPN64 % Surfshark49 % Private Internet Access47 % NordVPN35 % Providers with many locations in regions with few probes have more open results.\nDiscarded probes A probe whose stated location conflicts with the others is discarded and shown crossed out. Two probes that both reach the same server cannot be farther apart than their two limits added. If they are, one of them sits somewhere else than its database says.\nThe path, too: when a location is ruled out, the archive traces the route from the nearest probe and shows the last named routers. Their names often hint at the city where the server really stands.\nNext step Read VPN server lists as data to see where the locations come from in the first place.","title":"Reading measurement evidence: probes, limits, and open results","url":"https://crackthe.net/learn/reading-measurement-evidence/"}