Choose video settings by comparing frame times and visibility on your own hardware. A preset described as “maximum FPS” cannot guarantee a target frame rate for every PC or game update.
Use the same map, route, resolution and test duration. Record the original values and change one option at a time. Track visible stutter and responsiveness as well as average FPS.
Lower resolution and reduced anti-aliasing can reduce rendering work, especially when GPU-limited. Compare the result with your ability to identify distant players. If the CPU is limiting performance, lowering resolution may do little.
Evaluate shadow-related settings for information visibility rather than switching every option low. Texture choices depend partly on available graphics memory. Effects and ambient-occlusion choices should be compared in a realistic scene; do not assume one setting directly controls input latency in every system.
V-Sync, variable refresh, FPS caps and supported in-game latency options interact. Compare the configuration appropriate to your monitor rather than using “cap below refresh” as a universal rule. On compatible hardware, consult the vendor’s documentation for supported Reflex behaviour.
A video preset does not fix a poor server route. The maximum acceptable matchmaking ping preference is not a guaranteed in-match latency cap. For network symptoms, see the ping guide.
For drivers, cooling and background load, use our PC optimization guide. For persistent frame-time spikes, see the stutter troubleshooting guide. NVIDIA’s Reflex reference checked on 6 November 2026. This guide does not claim benchmark results on your hardware.
There is no universal preset. Compare frame times, visibility and responsiveness on your own hardware.
It can reduce GPU work, but a CPU bottleneck may limit the benefit.
Check what information remains visible. A frame-rate gain is not useful if it removes information you need.
No. Graphics-memory capacity and the rest of the workload can affect the result.
No. The appropriate cap depends on synchronization, display behaviour and sustainable system performance.
They do not directly improve the network route. Diagnose rendering and network latency separately.
Repeat the same scene and test duration, change one variable and record the original values.
Check the wider PC and frame-time troubleshooting guides rather than adding unverified launch options.