How we analyse PlayStation performance

Source hierarchy

1. Official announcements and developer documentation.
2. Independent technical analysis (video breakdowns, frame-rate benchmarking).
3. Reviews and community-reported findings, marked as such.
4. Facts declared on the PlayStation Store product page.

We never estimate a frame rate or a resolution. When a value isn't verified it is shown asUnknown orNot verified — an empty field is information too.

Outlets we read

Measured technical analysis: Digital Foundry, Eurogamer, VG Tech, NX Gamer, El Analista De Bits, Cycu1. Some of these publish only on YouTube — when they do, the video itself is the cited source and appears in the video section of the game page.

Editorial impressions used when no measured data exists: IGN, GameSpot, Push Square, GamesRadar, PC Gamer, Polygon, VGC, GamingBolt, always shown as a quote with a link, never converted into a PSFoundry Score.

Declared facts: the PlayStation Store product page and official publisher pages.

Three levels of coverage

Not every game has been measured. Instead of leaving a page blank, we publish the strongest level of evidence available for that title.

• Full technical analysis

  Measured data exists: graphics modes, frame-rate targets, internal and output resolutions, upscaling and ray tracing. These pages get a PSFoundry Score.

• Editorial impressions

  No measured numbers are available, but credible reviews describe how the game looks and performs. We quote them and never invent figures.

• Declared support

  Only the facts declared by the platform holder or publisher (120 Hz output, VRR, HDR, PS5 Pro patch). No frame rates, no resolutions, no PSFoundry Score.

The PSFoundry Score

A single 0–100 number describing how a game runs — never how good the game is. It is console-specific: a title can score differently on PS5 and PS5 Pro. It measures technical delivery only, and is kept separate from the PS5 Pro upgrade magnitude and from data confidence, so a big generational jump never inflates a technically shaky version.

• Frame rate

  26%

  Target and measured FPS, weighted towards the mode the game recommends. Frame-generated 60fps modes are scored as their native base frame rate. Holding a target is rewarded, but a game whose best mode never exceeds 30fps on this console is capped, and one that never reaches 60fps is capped lower than a 60fps title. When only some modes have a measured minimum, the score is capped: undocumented modes are never assumed flawless, and modes reported as unstable pull the score down.

• Image quality

  20%

  Output and internal resolution, reconstruction method (PSSR, FSR, TAA) and how it holds up in motion — never how detailed the art direction looks.

• Stability

  18%

  Measured minimum frame rates versus target, plus verified reports of stutter, tearing and crashes.

• Frame pacing

  8%

  Whether delivered frames are evenly spaced, judged separately from raw frame rate. The worst documented mode governs this category: one smooth mode does not offset another that judders.

• Responsiveness

  8%

  Measured input latency plus VRR / 120 Hz support. Without a latency measurement it is estimated from the natively rendered frame rate — interpolated frames never count — and reported input lag is penalised.

• Graphics modes

  8%

  Whether the options fit the game. One well-executed locked mode is not penalised against a title that ships five, but a single mode that exists only because the game runs through backwards compatibility is, and a 60fps mode that only reaches its number through frame generation does not count as a genuine high-frame-rate option.

• Visual features

  7%

  Ray tracing and PSSR are scored on execution when a game targets them, and simply excluded when it does not.

• Loading

  5%

  Measured load and fast-travel times, when a source publishes them.

The score is context-aware. A category a game never set out to deliver — ray tracing in a stylised 2D title, for instance — is excluded rather than scored as zero, and the remaining weights are re-normalised over what actually applies. Missing evidence is likewise never a penalty: it is left blank. When coverage is too thin to be comparable, no score is published and the page explains which fields are still missing. Data confidence (high, medium, low) is shown next to the score and never added to it.

90–100 Exceptional · 80–89 Great · 70–79 Good · 60–69 Fair · 40–59 Below average · below 40 Poor. Scores of 80 or more are shown in green, 70–79 in blue, below 70 in yellow.

Catalogue scope

PSFoundry covers games available on PlayStation 5 and PlayStation 5 Pro. DLC, season passes and bundles are excluded.