GUIDES

Can a VPN lower your ping? Routes, relays and how to test it

Sometimes: a VPN can't shorten the distance to the server, but it replaces part of your provider's route with its own, which can be shorter or longer. How to tell which on your line, with a test you run with and without it.

"My friend in the next town gets 40 ms on the same server and I get 70." "A VPN took 20 off my ping one evening and added 30 the next." "Is a gaming VPN worth paying for?" All three come down to one question: is your traffic taking a longer road to the server than it has to, and is someone else's road shorter?

Ping is distance plus the road

Ping is the round trip to the game server and back. Part of it is distance, which nothing on your PC or in a VPN shortens; part is the road your packets take, chosen by networks you don't control. The queue in your own router is covered in high ping and lag spikes.

That road is not picked for speed. Riot Games, explaining why it built its own network: "Backbone providers and ISPs route traffic to the lowest cost path, not the lowest latency path" (Riot Games, November 2015). Its example sends traffic from San Francisco to Portland by way of Los Angeles, Denver and Seattle: "A direct trip may have taken 14ms, but the less efficient route takes a full 70ms." Since providers route differently, one player can have "extremely low latency while their friend, who's just next door but on a different ISP, fights against significant lag." A longer road also passes more routers, and a router that "gets more packets than it can handle" is "forced to simply drop them."

What a VPN changes

A VPN client wraps your traffic in a tunnelling protocol and sends it to a VPN server, which passes it on; Microsoft's VPN documentation describes the client making "a virtual call to a virtual port on a VPN server". The trip now has two legs, your provider's road to the VPN server and the VPN's road from there to the game, and the ping you see is both together plus whatever the VPN server adds. Three outcomes follow:

Valve describes the same trade for its own relay network, Steam Datagram Relay: for "a surprisingly high number of players" its relays find a faster route, but where a server is not near one of the data centres running relays, "the best relayed route might be slower than the default IP route, and there will be at least some players whose latency is worse than the status quo." A VPN server is a relay too.

Is the game inside the tunnel?

For the VPN client built into Windows, Microsoft documents two configurations (Microsoft Learn). In a force tunnel, the default when no routes are specified, "all traffic will go over VPN". In a split tunnel, only the routes you specify use it, "and all other traffic will go over the physical interface". If your VPN sends only some apps or addresses through the tunnel, make sure the game is one of them, or your test measures your normal road.

Some games already run their own network

With a VPN on, that traffic reaches the game's network from the VPN server instead of from your provider, so you are comparing the game's own road with the VPN's road into it. Only a test on your line settles that.

Test it with and without

  1. Pick the region you play in and the hour you usually play.
  2. Without the VPN, note the ping the game shows. If you have w0a, its Game Server Ping page gives three numbers per region, measured to a reference endpoint rather than the match server: the lowest of several probes, the jitter, and the share of probes left unanswered (how it measures).
  3. Connect the VPN to a server between you and that region, check that the game is in the tunnel, and take the same readings.
  4. Compare the lowest value, the jitter and the loss, not one reading, and repeat on another day before deciding.

Reading tracert and pathping

Both need an address. If your game doesn't show the match server's, use one in the same region, as the Game Server Ping page does.

tracert followed by an address lists the routers on the way, 30 hops at most by default. Microsoft's reference notes that "some routers don't return time Exceeded messages": their hop shows as a row of asterisks, meaning the router stayed silent, not that the path is broken. The times on each row can be affected, Microsoft notes, by factors such as "network congestion, the quality of network links, and the distance between hops".

pathping adds loss. By default it sends 100 queries to each router and shows a busy message for "approximately 90 seconds", varying with the hop count, while it collects them (Microsoft Learn). Loss appears in two places:

The row that matters for your game is the destination. Run pathping to the same address with the VPN off and on, and compare that row.

Rules and side effects

What does not lower ping

Related

Did this page help?

What was missing, or what did not work? We read every one of these.

If this page did not solve it, write to us — say which article you read, and we will fix the article.