Why the screen is never the whole truth
In an online match, what you see is a slightly old picture of the game. Your opponent’s movements travel across the internet to the server, the server combines them with everyone else’s actions, and the result travels back to you. Every step takes time, so your screen shows a recent past, and so does your opponent’s, and the two pasts are not the same.
Most competitive online games use a server that is the final authority on what happened. Your computer sends inputs, the server decides the outcome, and your game then shows you that outcome. When two players act within a few thousandths of a second of each other, the order the server sees them in can decide the duel. That is not randomness in the dice sense, but it can feel like it, because the cause is invisible to the player.
Latency, ping and jitter
Latency is the time a message takes to travel. Ping is the usual way it is measured, as a round trip: your computer to the server and back. A ping of 40 ms means roughly 20 ms each way, though routes are not always symmetrical.
Two related ideas matter as much as the average:
- Jitter is how much the delay varies from one packet to the next. A steady 60 ms is often easier to play on than a connection that swings between 20 ms and 100 ms.
- Packet loss is when messages never arrive at all, forcing the game to guess or repeat information.
A player with a lower ping typically has a smaller gap between what they do and what the server learns about it. That is an advantage, but how large it is depends on how the game compensates for the gap.
What a tick rate means
A server does not update the game continuously. It processes the world in steps called ticks. The tick rate is how many of these steps happen per second, and it sets the finest timing the server can resolve.
The interval is simply one second divided by the rate:
| Tick rate | Time between ticks |
|---|---|
| 20 per second | 50 ms |
| 64 per second | 15.625 ms |
| 128 per second | 7.8125 ms |
Two actions that fall inside the same tick may be treated as simultaneous, or ordered by some rule the developers chose, such as which message arrived first. At 128 ticks the server can separate events that a 20-tick server would lump together. Higher rates are not automatically better in practice, though: they cost more server processing and bandwidth, and a player’s own frame rate, input hardware and connection can still be the bigger limits.
Lag compensation and its trade-offs
To make shooting feel responsive despite delay, many games use lag compensation. When your shot reaches the server, the server briefly rewinds the positions of other players to where you saw them at the moment you fired, then checks for a hit. This lets you aim at what is on your screen rather than leading targets by your ping.
The cost lands on the other side. The player who was hit may have already moved behind a wall on their own screen, and is told afterwards that they were hit. This is often called being shot around a corner. The effect grows with the shooter’s latency, because the server is rewinding further. Developers usually cap how far back the server will rewind, so a very high-ping player is eventually not compensated and has to lead their shots.
So the same system that helps the high-ping shooter can disadvantage the low-ping defender. Neither is cheating; it is a design choice about whose experience to protect.
What this means for close duels
Consider two players who see each other at the same instant and both fire. Several small factors can shift who the server credits first:
- Whose input reached the server earlier, which depends on ping and jitter.
- Which tick each input was assigned to.
- How far the server rewinds for lag compensation.
- Each player’s own display delay, since the monitor and game take time to show the enemy at all.
None of these is large on its own, and a difference of a few milliseconds is usually well below human reaction time, which is typically measured in the region of a couple of hundred milliseconds. But in a duel between equally skilled players, small timing edges can decide a share of the outcomes. It is hard to put a reliable number on it for any specific game, and anyone who claims a precise figure should be treated cautiously.
The effects of hit boxes, the shape of the area that counts as a hit, and weapon spread are separate from network timing, though players often blame the network for results that actually come from those systems. The underlying shooter mechanics are one layer; the network sits on top of them and adds its own uncertainty.
Reading your own results fairly
Because timing problems are invisible, they are easy to over-blame. A handful of strange deaths in one evening says very little, and the same short-run results that fool us about dice also fool us about lag. Remembering the odd trade that felt unfair is easier than remembering the many that felt normal.
If you suspect your connection is a factor, check measurable things: your ping and jitter, whether you are on a wired connection, and whether the game offers a network statistics overlay. Choosing nearby servers is usually the most reliable improvement. For judging whether your results have actually changed, you need more than a few matches, which is the point behind small samples in player statistics.
Fairness in online play is partly about the quality of the connection and partly about how developers choose to split the unavoidable delay between players. Matchmaking fairness often includes region and ping limits for this reason: a perfectly balanced skill match can still be tilted by a poor connection.