Jitter and packet loss are different problems
An uneven arrival and an absent arrival may both feel like lag, but they are not the same event.
Latency is travel time
Round-trip latency is the time for a journey out and back. Lag Arena divides the configured round trip evenly between the two directions before jitter. Real routes can be asymmetric and are not necessarily the routes used by your game server.
Jitter changes timing
Jitter describes variation in delay. This model adds a seeded value between minus and plus the configured jitter to each one-way packet delay, with a floor of zero. This is an explicit teaching distribution, not a universal definition or measurement method for jitter.
Loss removes a packet
Packet loss here independently drops a configured percentage of input and snapshot packets. A short run’s observed fraction need not equal the configured probability. The evidence panel therefore shows actual dropped and sent counts for this run.
Change one thing at a time
Start with Clean. Increase only latency and replay. Return to Clean, increase only jitter and replay. Then test loss alone. Keep seed and recorded input constant. Finally combine the variables and compare the input-arrival consequences.
A speed test is a different experiment
High throughput does not guarantee low delay under load. Tools such as Waveform’s bufferbloat test measure a real connection under defined conditions. Lag Arena sends no test traffic; its numbers describe the simulated world only.
Further reading
The experiment and explanation above were written for Lag Arena. For a deeper technical reference, read Waveform’s bufferbloat test.
Model review: October 31, 2026. This is an educational simulation; it does not reproduce a named commercial game.