PROTOCOL
How we rate and rank
A rating is a compressed opinion. This page decompresses it: what we score, how we weigh it and where the data comes from.
01// Five criteria
Each game gets a score from 0 to 5 in steps of 0.1 for each criterion. The overall rating is the weighted average, rounded to one decimal. There is no editor's bonus and no hidden adjustment.
Plot, characters, dialogue and how well the game uses its setting to say something.
Moment-to-moment play: controls, combat, movement or whatever you spend most of your time doing.
The place itself: art direction, sound, density and whether it feels inhabited.
Builds, routes, consequences and how much the game lets you play your way.
Performance, settings, stability, controller support and Steam Deck behaviour.
02// Worked example: Katana ZERO
- Story & writing ×25%
- 4.7
- Play feel ×25%
- 4.8
- World & atmosphere ×20%
- 3.8
- Systems & choice ×15%
- 3.6
- PC tech ×15%
- 4.5
4.7×25% + 4.8×25% + 3.8×20% + 3.6×15% + 4.5×15% = 4.350 → rating 4.4
03// Grades
Overclocked
4.5 and above
Best in class. Buy at full price.
Prime
4.0 and above
Excellent, with small flaws. Easy recommendation for fans of the genre.
Stable
3.5 and above
Good, with clear weaknesses. Wait for a sale or check the 'best for' line.
Glitch
below 3.5
Significant problems. Only for specific tastes.
04// Charts
Charts rank by a single criterion or apply a filter (price, Steam Deck Verified, playtime). Ties are broken by overall rating. Rules are stated at the top of each chart.
05// Data
Release dates, developers, publishers, US list prices and Steam Deck ratings come from Steam and SteamDB as of 3 Oct 2026. Playtimes are typical main-plus-extras estimates. Price per hour is list price divided by that estimate.
06// Independence
Ratings and positions are not for sale. We do not take payment from publishers, and we have no affiliation with gambling operators. Reviews are written by the editorial desk with AI-assisted research and editor review.