Who plays
Two regular contributors, occasionally a third for games that require it. Everybody involved knows everybody else outside this publication, which is deliberate: a co-operative game played with a friend and the same game played with a stranger are different products, and mixing the two would make every score meaningless.
The direct consequence is that we have nothing to say about matchmaking, lobby quality or playing with the public. That is a real gap and it is stated in every review rather than glossed over.
The together score
The headline number. The game played with the number of players it was designed for, by people who did not rehearse and did not read guides, for a minimum of six hours before anything is written.
Six hours is the floor and it is rarely the figure. The median game in this file received twenty-two hours, and the two games we abandoned fastest are scored with an explicit note that a short sample produced them.
The solo score
The same game played alone, by one contributor, where a single-player mode exists. It is scored on its own terms — we are not asking whether it feels lonely, we are asking whether it is a good game.
Where no solo mode exists the entry reads no score rather than zero. A game that declines to pretend it works alone is not being penalised for honesty.
| Requirement | Threshold |
|---|---|
| Hours in company before writing | 6 minimum |
| Hours alone before a solo score | 3 minimum |
| Players present | As the game intends |
| Guides or wikis consulted | None during scoring |
| Sessions logged as they happen | All |
Partnership survival
The figure this publication exists for. It is the number of hours played together before the arrangement stopped being enjoyable, and it is recorded at the time rather than decided afterwards.
A partnership is treated as ended when any of four things happens: we finish the game; we stop scheduling sessions; one player becomes a passenger for a sustained period; or a session ends in real friction and neither of us proposes another. Whichever occurred is named in the review.
Two things this figure is not. It is not a measure of length — a game we finished at thirty-four hours and a game that bored us at thirty-four hours would show the same number, so the reason is always stated. And it is not a prediction about your friendships.
Skill gap grade
The most practically useful thing we publish and the hardest to justify, because it depends on a specific pair of people with a specific difference in ability.
The two of us differ noticeably at anything requiring vehicle handling or precise timing, and are close to equal at planning and observation. That is our instrument, and a game is graded on how visible that difference became.
| Grade | Means |
|---|---|
| A | Gap never surfaced; roles need different competences |
| B | Gap visible but managed by rotating roles |
| C | Gap affects outcomes; weaker player notices |
| D | One player carries; the other becomes an observer |
| F | Unplayable by a mismatched pair at any setting |
What we record during a session
Start and end time, who played which role, anything that produced friction, and — for games with friendly fire — who caused each incident and who the game blamed. That last column is why the first session log exists.
Notes are written during play or immediately after, never days later. Where a review makes a claim about how something felt at a particular hour, that claim comes from a note written at that hour.
Survival hours are recorded as the sessions happen. Nothing in this file is a figure we reconstructed afterwards from memory, because memory would flatter the games we liked.
Purchasing and independence
Every copy is bought at retail with the publication's own money, and co-operative games mean buying two or three copies, which is why this file grows slowly. No review copies, no early access, no embargoed coverage, and no game scored while it is still in a beta state.
The four limits
First, two people. Every finding here is filtered through one specific pair, and the session logs say where that matters most.
Second, no strangers. Nothing here describes public matchmaking.
Third, our schedules. Several partnerships in this file ended because of calendars rather than games, and where that is the case the review says so instead of blaming the design.
Fourth, one hardware configuration and one household. Latency between two people in the same building is not the latency most co-operative play involves, and we cannot speak to how these games behave across a poor connection.
Corrections
Where a published figure turns out to be wrong it is struck through with the replacement beside it and the date of the change. Where a whole finding reverses, it is rewritten at the top of the relevant log rather than quietly edited, and the previous version is described.
Summary
Minimum six hours in company before writing, three alone before a solo score. Partnership survival recorded live, with the reason for its end always named. Skill gap graded A to F against one specific pair of players. Everything bought at retail, nothing played with strangers, no guides during scoring.