One number per game, out of ten, to one decimal place. Roughly half of it is the quality of the input layer, which is unusual enough that it needs stating explicitly rather than being buried.
We do not publish sub-scores, because averaging categories produces numbers nobody can interpret. What we can describe is the weighting in our heads when we argue about a final figure. Response, commitment and readability together account for about half. Level and encounter design accounts for roughly a quarter. Everything else — writing, art direction, sound, structure, technical stability — makes up the remainder.
That weighting is a genuine editorial position rather than a claim about how games should be judged universally. A site built around narrative criticism would weight it completely differently and would be right to. Ours is stated so that a reader can adjust for it.
The response number in each review is the median of 100 camera-measured samples on our reference wired pad, with the game locked to 60 frames per second on a 60 Hz display. It includes everything: the pad, the cable, the operating system, the engine and the panel. That makes it a whole-chain figure rather than an engine figure, which is what a player actually experiences.
What it does not tell you is how the game will feel on your setup. A wireless pad typically adds six to ten milliseconds. A high refresh rate display typically removes eight to twelve. A television in a processing mode can add sixty. We publish our figure so you have a fixed reference point, not so you can predict your own.
Response describes how quickly a game acknowledges you. Commitment describes how long it then ignores you. A game with a 30 millisecond response and a 40-frame animation lock feels far worse than one with a 60 millisecond response that lets you cancel out of everything. Because these two properties are frequently confused, we time them separately and report them separately.
A published score changes only when a patch materially alters something we measured. Taste does not revise a score, nor does a change of reviewer, nor does the passage of time. When a revision happens, the original number stays visible on the page alongside the new one and the reason is stated in a dated note.
| Game | Original | Revised | Reason |
|---|---|---|---|
| Heavy Frame | 7.4 | 7.8 | Patch corrected recovery cancels on two weapons |
| Ring Depth | 8.6 | 8.9 | Hitbox display added in a post-release update |
| Glass District | 7.4 | 7.4 | Requested review after patch; camera events unchanged |
Our numbers may not be used in advertising, on storefront pages or in any promotional material. A score here is inseparable from the measurements that produced it, and a figure quoted without them tells a reader nothing they can act on. We treat that separation as a misrepresentation of the work rather than a marketing convenience.
Readers disagree with these numbers regularly and that is the expected outcome of publishing an opinion. What we would ask is that the disagreement engage with the measurement: if a figure looks wrong for your setup, the likely explanation is a difference in device, refresh rate or display processing, all of which are described above and none of which make either of us mistaken.
