Hell Let Loose: Vietnam went live on 13 August and pulled more than 30,000 concurrent players on Steam within days. It also landed a “Mixed” rating — around 53% positive across its first 1,706 reviews — while critics scored it 78 on Metacritic. Read the negative reviews and a familiar pattern appears: crashes, voice chat, and a server browser that kept handing people a seat that was already taken. Almost all of it got filed by players under one word. Lag.
It usually is. “Lag” is where three unrelated problems go to be misdiagnosed, and the fix for each one is different. Since a chunk of our support inbox is people convinced their ping is our fault, their router’s fault, or their internet plan’s fault, it is worth laying out what actually determines the number in the corner of your screen.
Your internet speed is almost never the problem
This is the single most expensive misunderstanding in gaming, because it is the one that gets people to upgrade their plan for nothing.
Bandwidth and latency measure different things. Bandwidth is how much data fits through the pipe at once, measured in megabits per second. Latency is how long one small packet takes to make the round trip, measured in milliseconds. A competitive shooter sends tiny packets very frequently — a few hundred kilobits per second is plenty. It has never been short of room.
So going from a 100 Mbps plan to a 1,000 Mbps plan does approximately nothing to your ping. You widened a road that was never congested. The delay is in the distance and the route, and no amount of megabits shortens either.
Bandwidth is how wide the road is. Latency is how far away the destination is. Buying a wider road does not move the destination closer.
What actually sets your ping
Four things, roughly in order of how much they matter and how little control you have over them.
1. Physical distance
Light in fibre travels at about two-thirds of its speed in vacuum, which works out near 200 km per millisecond. A round trip to a server 1,000 km away costs you roughly 10ms before anything else happens, and no technology sold to consumers will beat physics. This is the floor.
2. The number of hops
Your packets pass through a series of routers, and each one costs a little time to process and forward. A route with eight hops beats an otherwise identical route with eighteen. Hop count is not something you choose — it is decided by your ISP’s routing table.
3. Peering — who your ISP talks to directly
This is the big one, and almost nobody outside networking thinks about it. If your provider peers directly with the network hosting the game, your traffic takes a short path. If it does not, your traffic gets handed to a transit provider and may travel to another city, or another country, before coming back.
The measured effect is larger than most people expect. Within a single country, median ping can vary by 20 to 40 milliseconds between one ISP and another — purely from routing decisions, with geography held constant. Two neighbours on the same street, same speed tier, can have very different ping because their providers made different peering choices. Networks that peer directly with the major hosting platforms consistently come out ahead.
4. The last few metres
Wi-Fi adds latency and, worse, adds variance — and inconsistent ping feels far worse than consistently high ping, because prediction code cannot compensate for a number that will not sit still. Ethernet remains the cheapest upgrade in gaming. Bufferbloat is the other local culprit: an oversized queue in your router that holds packets during congestion, so ping stays fine until someone starts a large upload and then spikes into the hundreds.
Some of it is just where you live
Median latency to regional game servers in 2026 makes the geography painfully clear:
- Singapore 8ms, South Korea 9ms — dense domestic fibre, publishers hosting locally, and in Singapore’s case a central position with direct routes across Asia-Pacific.
- Netherlands 11ms, Germany 12ms — AMS-IX and DE-CIX Frankfurt, two of the largest internet exchanges in the world, sitting next door to most European game hosting.
- Japan 13ms, US East 18ms, US West 24ms, Brazil 28ms.
- Australia 52ms, South Africa 64ms, New Zealand 71ms — an unavoidable distance tax. New Zealand has to reach Australia or the United States for most things.
- Pakistan and Bangladesh 80–120ms — not distance so much as missing local exchange infrastructure, which forces traffic out through Singapore or Mumbai and back.
That last case is the important one, because it is not physics. It is infrastructure that has not been built yet. Countries in that position see their latency fall sharply when a local exchange opens and providers start peering there, without a single household changing their plan.
Server capacity is a different problem entirely
Back to the launch. The complaints about the server browser — seats that looked open and were full by the time you connected, or slots reserved for VIP players — were not latency at all. That is a capacity and stale-state problem: the browser reports a snapshot, then reality moves on before your connection completes.
The developers shipped a hotfix shortly after launch that improved AMD GPU stability and, tellingly, added more official servers across regions. That is the correct fix for a capacity problem and it would have done nothing whatsoever for a routing problem. Worth remembering when you are trying to work out which one you have: if you get in and play smoothly, your ping is fine and the queue was the issue. If you get in and the game feels like it is underwater, that is latency.
What actually helps
- Use a cable. Ethernet, every time, if the room allows it. Nothing else on this list is as cheap or as reliable.
- Pick the closest server region manually where the game lets you. Automatic matchmaking optimises for queue time, not for your ping.
- Run a traceroute or mtr to the game server. You are looking for where the delay is introduced. If your packets leave your city and reappear three hops later in another country, you have found a routing problem, and you now have evidence.
- Test for bufferbloat with a loaded-latency test. If ping is fine idle and terrible during an upload, enable Smart Queue Management or SQM on your router. This is a genuine fix, not a placebo.
- Take the traceroute to your ISP. Routing complaints backed by data occasionally get acted on. Routing complaints phrased as “my game lags” never do.
- Consider switching providers. If a competitor in your area peers better, that 20–40ms is available to you and no equipment upgrade will match it.
And the honest part about VPNs
Plenty of VPN marketing promises lower ping. We are not going to make that promise, because in most cases it is false.
A VPN adds a hop. Your traffic goes to the VPN server first and reaches the game from there. If your ISP was already taking a sensible route, that detour costs you latency. That is the normal outcome, and anyone claiming otherwise as a general rule is selling something.
There is a narrow set of cases where it goes the other way, and they are all versions of the same situation — your ISP’s route was worse than the VPN’s:
- Your provider has poor or missing peering with the game’s host, and the VPN network peers well.
- Your provider throttles or deprioritises traffic it identifies as gaming, and encryption stops it being identified.
- Your traffic is being routed through a distant exchange when a shorter path exists.
If you are in one of those situations the improvement can be substantial. If you are not, you will add 5 to 30ms. The only way to know which applies to you is to measure your ping with and without, on the same server, at the same time of day — not to trust a claim on a pricing page, including ours.
What we will claim is what a VPN reliably does: it encrypts every packet leaving your device so the networks in between cannot read or log what you connect to, and it keeps your home IP address off the servers you talk to — which, in games with player-hosted lobbies, is not a purely theoretical concern. Those are the reasons to run one. A better number in the corner of the screen is a maybe, and it depends entirely on how bad your provider’s routing was to begin with.