Methodology

Each graphics card and processor carries a single performance rating on a 1–10.5 scale. Ratings come from published review benchmarks and generation-over-generation comparisons, normalised so that cards and chips of comparable real-world gaming performance land on the same number. Video memory is tracked separately, because an 8 GB card and a 16 GB card of the same speed behave very differently at high texture settings.

Every game has a demand rating on the same scale plus a baseline framerate that represents what a mid-range system achieves at 1080p. These come from launch and post-patch benchmark coverage and community reports. Competitive titles are flagged separately: for those we bias recommendations toward framerate and clarity instead of maximum visual quality.

The gap between your graphics card rating and the game's demand picks a quality level from Low to Ultra+. Resolution shifts that level — 1440p costs roughly one step and 4K roughly two. Texture quality is then adjusted for available video memory and system memory, upscaling (DLSS, FSR or XeSS) is enabled where the game supports it and the hardware benefits, and ray tracing is only recommended on cards that can absorb it.

The estimate starts from the game's baseline framerate and is scaled by your graphics card rating, then multiplied by factors for resolution, the chosen quality level, ray tracing, upscaling, processor rating and system memory. A processor that is far weaker than the graphics card caps the result, which is how the bottleneck warning is derived. Framerate caps built into a game (Elden Ring's 60 fps lock, for example) are applied last.