VALORANT · PRACTICAL SETUP GUIDE
VALORANT NVIDIA Reflex: On vs On + Boost
Choose a testable baseline, distinguish latency from FPS, and summarize your own On and On + Boost readings without inventing a benchmark.
Published and sources checked: · ProConfigVault editorial
A useful starting point
On compatible GeForce hardware with the setting available, use Reflex On as a comparison baseline. Try On + Boost as a separate experiment. Keep it only when the result and power tradeoff suit your setup. This is an editorial testing approach, not a promise that one mode wins on every PC.
What On and On + Boost change
NVIDIA describes Reflex as game-integrated scheduling that reduces render-queue delay. Boost additionally holds higher GPU clocks in CPU-bound conditions, with a power tradeoff. See NVIDIA’s Reflex technical introduction. The feature’s benefit depends on the workload; this page does not assign a fixed percentage reduction.
| Mode | Role in your test | What to record |
|---|---|---|
| Off | An optional diagnostic control if you need to compare Reflex enabled versus disabled. | Record the same metric and scene; do not treat an Off run as equivalent to an On run. |
| On | A baseline with the game’s Reflex latency control enabled. | Save the setting and record its result before adding Boost. |
| On + Boost | A comparison that adds the higher-clock behavior described by NVIDIA. | Record the latency metric, frame-time behavior and any power, temperature or fan-noise change. |
If the option is missing, first check your GPU, the running game client and the official driver reference. This guide does not establish that every GPU or platform exposes it. Do not install a third-party “Reflex unlocker” to make the menu match a screenshot.
Choose the measurement before choosing the mode
More FPS does not by itself prove lower click-to-display delay. NVIDIA distinguishes PC timing from complete system latency; see its Reflex Analyzer measurement explanation. Keep the exact metric label in your notes.
| Observation | What your record can say | What it cannot establish alone |
|---|---|---|
| FPS or a frame-time reading | Whether frame delivery changed in the sampled task. | The complete delay from a physical click to a visible screen response. |
| A named in-game or software latency statistic | How that specific statistic behaved under the same measurement method. | An interchangeable result with another tool or a different timing scope. |
| A correctly configured end-to-end measurement setup | A click-to-display observation within that setup’s stated method and limitations. | A universal result for other PCs or all VALORANT scenes. |
| Ping, packet loss or a server symptom | A separate network observation to investigate. | That switching Reflex modes repaired the network. |
| “Feels faster” in one round | A subjective note that may guide a further test. | A measured number of milliseconds or a repeatable improvement. |
The calculator below never reads the game, the GPU or an input device. You supply the numbers. If all you have is FPS, keep an FPS record instead of entering it into a milliseconds field.
A small A/B test you can reproduce
- Save the conditions. Write game build, driver, CPU/GPU, resolution, graphics profile, frame cap, refresh rate, VSync/VRR state, power mode and whether the laptop is plugged in. Do not change those while comparing Reflex.
- Pick an exact metric and a fixed task. Write the tool and complete label. Use one location, movement route and duration. Let the system settle under the same conditions before each recorded run.
- Alternate the mode order. Try On → On + Boost → On, then On + Boost → On → On + Boost. Record one comparable summary per run using the same method. These six samples are a small check, not a statistical certification.
- Keep the raw record. Include mode, run order, scene, duration, metric and value, plus any hitch, thermal change or background interruption. Mark interrupted runs and repeat them rather than quietly selecting a favorite result.
- Look for repeatability. Use the medians and ranges as a compact summary. If the result flips between sessions or the sample ranges overlap, treat the evidence cautiously. Even non-overlap in a tiny sample is not proof that a difference is reliable.
Keep a stable setting if you cannot distinguish the modes with your available method. If Boost adds an unwanted power or noise cost without a repeatable useful change, return to On. That decision is about your observations and priorities, not a professional player’s copied preset.
Compare your own latency samples
Enter three comparable per-run readings for each mode, all in milliseconds and all from the same metric. The tool reports medians, ranges and simple arithmetic. It does not declare a winning mode or perform a significance test.
No samples have been entered. The median is the middle of three sorted readings; the range runs from the smallest to the largest. A lower sample median is not proof of a repeatable latency improvement.
Questions that can otherwise spoil the comparison
Should I change NVIDIA Control Panel Low Latency Mode too?
Keep it fixed while testing. NVIDIA says in-game Reflex takes priority over its driver Ultra Low Latency mode when both are enabled (technical reference). Changing both controls in the same run makes your own comparison harder to interpret.
Should I uncap FPS or change VSync for this test?
Record and hold those settings constant first. A new frame cap or display-sync policy is a different experiment. This page does not prescribe one universal VSync/VRR configuration for every display.
Can I compare this result with a mouse polling-rate number?
A polling interval is a calculation about reporting frequency, not the same measurement as PC or complete system latency. Use the polling-rate interval calculator to understand that arithmetic; keep the two records separate.
Sources, date and limits
- NVIDIA — Introducing Reflex (2020 technical introduction): scheduling, Boost’s clock/power tradeoff and driver-mode priority. This historical explanation is not presented as a current compatibility list.
- NVIDIA — Reflex Analyzer explanation: measurement scope. Historical overlay shortcuts and supported-device counts are not reproduced here.
Sources checked 18 October 2026. The test routine and sample calculator are original editorial tools. We have not benchmarked the modes in the native game for this article, verified your menu, or measured a guaranteed percentage benefit.
Continue with the general VALORANT graphics guide, mouse input checks, or ProConfigVault tools hub.
