VPN and gaming latency
A tunnel adds latency equal to the round trip to the server, on top of the round trip to the game. On a nearby server that is 5 to 25 milliseconds; on another continent it is over 150. No setting recovers it, because the packets genuinely travel further.
4 min read
Start with the honest version: for competitive play, a VPN makes things worse. The packets take a longer path, and there is no configuration, protocol or provider that makes a longer path shorter.
That said, the cost is quantifiable and there are cases where it is worth paying, plus one narrow case where a tunnel genuinely helps.
What it costs
Your traffic goes to the VPN server, then to the game. The added latency is the round trip to the server, and it is added to every packet in both directions.
| Server location | Added round trip |
|---|---|
| Same city | 2 to 10 ms |
| Same country | 5 to 25 ms |
| Neighbouring country | 15 to 40 ms |
| Another continent | 120 to 250 ms |
For a game where the difference between 30 ms and 50 ms is noticeable, a nearby server is the only tolerable option and a distant one is unplayable.
Jitter matters more than average ping
Jitter is how much the round trip varies between packets. A stable 60 ms feels better than a ping that swings between 25 ms and 120 ms, because prediction and interpolation in the game engine can compensate for consistent delay and cannot compensate for inconsistent delay.
A well-connected VPN server on a clean path can actually reduce jitter compared with a congested residential route. That is the narrow case where a tunnel helps, and it is real — just much less common than the marketing suggests.
Measure both with the speed test, which reports jitter alongside latency.
NAT type, the other cost
Multiplayer games use peer-to-peer connections for voice and sometimes for gameplay. A tunnel adds a layer of address translation, which typically moves your NAT type from open to moderate or strict.
The effects are slower matchmaking, voice chat relayed through servers rather than direct, and occasionally an inability to host. Fixing it means port forwarding through the tunnel, which most consumer providers do not offer.
That is often a bigger practical problem than the latency.
Where a VPN does help
Against connection-based attacks. In games where opponents can obtain your address, a tunnel means they get the server's instead. This is a genuine problem in some competitive scenes and a VPN addresses it directly.
Against ISP routing that is worse than the alternative. Occasionally the path from your provider to a game server is poor, and routing via a well-connected datacentre is better. Rare, worth measuring, never worth assuming.
Against bufferbloat you cannot fix. If a shared connection saturates and your latency triples, a tunnel does not fix it — but it does not cause it either, and diagnosing it is worth more than either choice.
What does not help
"Gaming VPNs" promising lower ping. Some sell optimised routing, and on a genuinely bad path that can be real. Most are selling the same tunnel with different marketing, and the physics is unchanged.
Choosing a server near the game rather than near you. Your traffic goes to the VPN server first, so a distant VPN server plus a short hop to the game is still a long path. Distance from you to the exit is what counts.
The practical answer
If you need the tunnel on for privacy reasons and you play casually, pick the nearest server from the server list and accept the 10 to 20 milliseconds.
If you play competitively, use split tunnelling to exclude the game, or turn the VPN off for those sessions. Both are reasonable, and there is no configuration that gets you a tunnel and unchanged latency at the same time.
If the reason for the VPN is protection from address-based harassment, keep it on and pick the nearest server. That is a case where the trade clearly favours the tunnel.
Cloud gaming is a different question
Streaming a game from a remote service rather than running it locally changes the calculation entirely. Every frame is a round trip, so the latency budget is much tighter and there is no client-side prediction to hide it.
A tunnel in front of a cloud gaming session adds its round trip to an already-marginal path, and the result is usually visibly worse. It also adds bandwidth pressure, since these services push 15 to 40 Mbps continuously and are sensitive to jitter.
If you use cloud gaming, exclude it from the tunnel or turn the tunnel off. This is the one gaming case where the answer is not "it depends".
Measuring it properly
Ping the game server directly with the VPN off, then ping the VPN endpoint. The sum of the second and the path beyond it is roughly what you will pay, and comparing that with the first tells you the cost before you commit to a session.
In-game latency displays are more useful than any external test, because they measure the actual path to the actual server. Take a reading over several minutes rather than one glance: the average matters less than how much it moves.
