Frame rate is only meaningful with its test conditions. A note that says “the game runs at 60 FPS” is incomplete if it omits the device, resolution, graphics preset, display mode, refresh rate, measurement tool, and scene. A new user needs a reproducible baseline and a safe way to change one variable at a time.
Identify the whole test system
Record the computer or console model, operating-system version, game title and build, graphics driver, CPU and GPU, memory, display model, cable or connection type, resolution, display mode, and selected refresh rate. Note whether the device is on external power and which power profile is active. For laptops, record whether the game uses integrated or discrete graphics.
Do not treat display refresh rate and rendered frame rate as the same value. Microsoft defines display refresh rate as how many times per second the screen refreshes and shows the selected rate, resolution, and variable-refresh-rate capability in Advanced display settings (Microsoft Support). The game can render above or below that rate. State both values and how each was observed.
Preserve the graphics and display settings
Export settings when the game supports it; otherwise capture readable screenshots and transcribe only the settings that affect the test. Include resolution scale, quality preset, frame cap, vertical sync, upscaling or frame generation, ray tracing, motion blur, field of view, fullscreen or borderless mode, HDR, and relevant operating-system graphics options. Record whether a restart is required.
Windows 11's windowed-game optimization can change presentation behavior for compatible DirectX 10 and 11 games, reduce frame latency, and enable features such as variable refresh rate on supported displays (Microsoft Support, Optimizations for windowed games). Include its state when the test uses windowed or borderless mode. Do not enable it during a baseline run unless that change is the item being tested.
If variable refresh is active, name the technology and configuration. NVIDIA describes G-SYNC as matching monitor refresh rate to GPU frame rate and requires a compatible display, GPU, operating system, and application (NVIDIA G-SYNC). Record whether it is enabled for fullscreen only or also for windowed applications, if that choice exists.
Define a repeatable measurement
Name the measurement tool, version, overlay or logging mode, sampling duration, warm-up period, and output location. State whether the number is an instantaneous overlay, an average, a percentile, or frame-time data. Microsoft's DXGI documentation notes that presentation counts and refresh counts describe different events and that presenting an image is not necessarily identical to calling the Present API (Microsoft Learn). That distinction is one reason not to compare metrics from different tools without describing their method.
Choose two or three reproducible scenes: a saved benchmark, a fixed tutorial path, or an in-game benchmark. Record the location, camera movement, duration, weather or player count if relevant, and whether shader compilation or asset loading occurred. Run the same scene several times after warm-up and retain every run, not only the best one.
Report more than an average
Keep average frame rate, a low-percentile or frame-time stability measure when the tool provides it, visible stutters, and the display refresh state. State the run count and variation. A high average can hide uneven delivery, while a stable result can still feel wrong if input latency, tearing, or motion blur is the actual concern. Capture the user's symptom separately from the measurement.
Avoid unsupported comparisons. Results from different resolutions, patches, drivers, power states, scenes, or tools are not direct before-and-after evidence. Label estimates and manual observations clearly. Do not claim a setting improved performance from a single nonrepeatable run.
Include troubleshooting boundaries
Provide a one-change-at-a-time order: reproduce the baseline, verify resolution and refresh rate, check the frame cap and synchronization settings, confirm power and GPU selection, then change one graphics setting. Save the previous state and define the rollback. Stop if a driver update, firmware change, overclock, BIOS setting, or hardware modification requires separate approval.
Leave a usable handoff
The notes should include the known-good configuration, baseline table, raw log location, screenshots, measurement instructions, test scenes, observed limitations, last verified date, and owner. Have the new user reproduce one baseline run and one reversible setting change. The handoff is ready when another person can obtain a comparable result and explain why refresh rate, frame rate, and presentation statistics are related but not interchangeable.
Sources
- Change the refresh rate on your monitor in Windows — Microsoft Support
- Optimizations for windowed games in Windows 11 — Microsoft Support
- NVIDIA G-SYNC — NVIDIA
- DXGI_FRAME_STATISTICS structure — Microsoft