Discover how Edgegap works

Discover how Edgegap works

Latency

Latency is the time a packet takes to travel between a player’s device and the game server, usually quoted as the round trip and measured in milliseconds. It sets the floor on how responsive a match can feel, because no amount of client-side work removes physical travel time. Distance is often the largest contributor, and it sets a floor no software can remove.

Also called

lag, network latency, server latency, game latency

By

By

By

Philip Coté

Philip Coté

Philip Coté

,

CTO

CTO

CTO

Published

Published

Published

What Is Latency in a Game, and What Does Server Latency Mean?

“Latency” is used for several different delays, and they add up to what a player feels:

  • Network latency is the time data takes to cross the network. It can be quoted one way or as a round trip (RTT). Many in-game displays show the round trip, which is what ping measures.

  • Server latency is the time a message waits and is processed inside the game server. A server running at a tick rate of 128 Hz handles inputs about every 7.8 ms; at 30 Hz, about every 33 ms.

  • End-to-end latency runs from a player’s input to a change on screen. On a single machine, NVIDIA calls the local part system latency, “the delay between your mouse or keyboard actions and the resulting pixel changes on the display”, covering the input device, the computer and the monitor but not the network (NVIDIA, September 2020).

  • Lag is the player’s word for any of the above, and often for packet loss and jitter too.

For a multiplayer developer, the useful view is the whole budget, from one player’s input to another player’s screen, and which part of it the network adds. That is what “game latency vs server latency” usually comes down to: one number for the trip, one for the server, inside a longer chain.

Article's key insights

  • Latency has a floor, and everything else stacks on top of it: The physical travel time between two machines is the minimum delay any online game can have. Routing, tick rate, update rate and processing all add to it. Netcode cannot remove that delay. It can only decide where players feel it.

  • The shortest path is rarely the fastest: A direct peer-to-peer connection looks like the shortest route between two players, but packets still cross both players' ISPs and whatever routes those networks choose. A well-placed server between them is often both faster and fairer.

  • Lag compensation moves the cost, it doesn't erase it: Lag compensation rewinds the server so a high-ping player's shot counts, which means the target can be hit after reaching cover. The only real fix is keeping ping gaps small, which makes it a matchmaking and server placement problem as much as a netcode one.

  • Tick rate is a time budget: A server's tick rate sets how many milliseconds it has to simulate each step of the game. Higher rates cut delay and sharpen hit registration, but cost more CPU and bandwidth. A server that overruns its budget stutters for everyone in the match.

  • Two separate choices, not one: Where the game runs (peer-to-peer, relay or dedicated server) and how state is shared (lockstep, rollback or state replication) are independent decisions. Any sync model can run on any hosting model.

Where Does Game Latency Come From?

A round trip between a player and a game server picks up time in four places:

  1. Distance: Light in optical fiber covers about 5 µs per kilometre (CommScope, 2017), so every 100 km between player and server adds about 1 ms to the round trip. São Paulo to Munich is about 9,850 km in a straight line, a round-trip floor of about 98.5 ms in fiber before any equipment is counted. Toronto to Montreal, about 500 km, has a floor of about 5 ms.

  2. The player’s connection: Home connections differ by technology. Measuring round trips to a nearby test server, the FCC found internet providers’ median idle latency of 7 to 14 ms on fiber, 12 to 24 ms on cable and 23 to 34 ms on DSL in late 2022, and that latency under a download “generally is significantly higher than idle latency” (FCC, 2023). Wi-Fi and mobile networks add their own delay on top.

  3. Routing: Packets rarely travel in a straight line. Riot Games found that fiber routes between the same two locations could be “fully three times longer” from one company than from another (Riot, 2015 data). Researchers measured internet ping times at more than three times the time light would take on a straight line (APNIC, 2017).

  4. Time in the server: An input waits for the next tick, then is processed and sent back with the following update.

Across regions, distance is usually the largest part, and no software removes it. Within a region, the player’s own connection can weigh as much as the trip to the server.

Diagram of where one game round trip picks up time: player, home connection, internet routes and game server, with distance, the player’s connection, routing and server tick time labelled, and the fiber distance floor of about 5 ms from Toronto to Montreal against about 98.5 ms from São Paulo to Munich.

How Much Delay Does a Game Add on Top of the Network?

The network round trip is often a smaller share of what a player feels than they assume. In July 2020, Riot modelled how much longer a peeking player in VALORANT has to react than the defender they peek. Under its baseline assumptions, a 35 ms round trip, 128-tick servers and 60 FPS clients, the head start came to about 141 ms (Riot).

Most of that comes from inside the game: about two frames of buffering on the server, about three on the client, including time for the frame to reach the screen, and the client’s frame time. Riot estimates that its regional servers, its own routing network, its 35 ms target and 128-tick servers shave about 40 ms (28%) off that baseline advantage. At 144 FPS instead of 60, the total falls to about 71 ms. Our VALORANT deep dive covers how those choices fit Riot’s backend.

Diagram of the delay from one player’s action to another player’s screen in Riot’s VALORANT model: client frame, network, server buffering, network and client buffering, with the head start for a peeking player at about 141 ms at 60 FPS and about 71 ms at 144 FPS.

Netcode works on this budget, but none of it shortens the trip. Client-side prediction shows a player’s own actions before the server confirms them, interpolation adds a small, deliberate delay to smooth others’ movement, and lag compensation rewinds the server to what the shooter saw when they fired. Each decides where the delay shows up. The round trip itself only shrinks when the server is closer or the route is better.

How Do You Lower Game Server Latency?

A developer can’t move players or change their internet provider. Three choices still shape the round trip:

  1. Where each match’s server runs: Distance sets the floor, so the more locations a server can start in, the more often one sits close to every player in the match. In Edgegap’s case study of a AAA publisher, analyzing 600,000 transactions, average latency fell from 116 ms to 48 ms, and the share under 50 ms rose from 14% to 78% (Edgegap, 2023).

  2. Who is in the match: A matchmaker that measures each player’s latency before grouping them can keep distant players apart and widen its limits only when a match won’t fill. Server regions and ping cover the trade-offs, and our guide to latency through matchmaking walks through the setup.

  3. The route: Some studios work on the path itself. Riot built its own network and peered with internet providers to bring players’ traffic onto it “as close to the player as possible” (Riot), a choice at the scale of a large publisher. Hosting across many providers’ networks also gives placement more routes to choose from.

The rest of the budget, frame rate and buffering, is tuned in the game itself.

A word from our sponsor (ourselves!)

Most lag comes from distance, not from code. Edgegap's Edge Cloud is a distributed network with up to 615+ locations available on demand, so each match runs at the best available location for its players. Independent replays of live studio traffic measured a 58% drop in average round-trip time compared with public cloud.

Edgegap's Take (just our opinion, take it with a grain of salt!)

Players Feel Latency in Your Game, So Control Where Each Match Runs

Players notice latency more than most problems. In a survey of 2,000 players in the US and UK that Edgegap commissioned from Bryter, half ranked it their top issue while playing, and about 1 in 3 said they stop playing when it affects them (report, July 2022).

Much of that latency starts outside a developer’s code: the distance to the server, the player’s own connection and the routes their provider chooses. Players still feel it in your game, so the part you do control matters: where each match’s server runs.

Edgegap deploys each match’s server on demand at the best available location for its players on the Edge Cloud network, across 615+ locations on 17+ providers (platform data, as of 18 September 2026). Placing the server where its players’ combined latency is lowest lowers the delay they share and narrows the gap between them, so matches can be both faster and fairer. In the AAA publisher case above, moving servers closer to players cut average round-trip time by about 58%, from 116 ms to 48 ms.

A useful check: per region, the share of matches where every player stays under the latency your game can handle.

Philip Coté, CTO

-

-

-

Philip Coté

Philip Coté

Philip Coté

,

CTO

CTO

CTO

Frequently Asked Questions

What is good server latency?

Which game server hosting has the lowest latency?

How do you test game server latency?

Get your Game Online Easily & in Minutes

Start Integrating Now!

Get your Game Online Easily
& in Minutes