A single review can tell you whether a game is good. It cannot tell you that eleven of the nineteen games in this file blame the wrong player for friendly fire, or that most co-op designs quietly assume nobody will ever miss a week. These logs exist for the findings that need the whole file.
Eleven games with friendly fire, and one consistent structural problem: the player who is standing still and doing their job correctly is the easiest one to hit, so the game reliably assigns fault to the careful partner.
Nine games tested by deliberately abandoning one player for a fortnight. Three handled the return gracefully, four made it awkward, and two made the returning player a liability in a save they had helped build.
Both logs are drawn from sessions that were played for the reviews rather than staged separately. Nothing here was set up to prove a point, which is why the samples are uneven — eleven games have friendly fire and nine were suitable for the absence test, and neither number was chosen.
Figures in the logs are re-checkable. Where a claim rests on a specific session it says so, and where it rests on an impression it is described as an impression rather than dressed up as a count.
An attempt to measure voice-channel load — how much of a session is spent talking — was abandoned after four games. Transcribing our own conversations turned out to change how we spoke to each other, and the numbers described the experiment rather than the games.
A second attempt to grade matchmaking quality was closed before it opened. Every game in this file was played with someone we already know, which means we have nothing honest to say about playing with strangers, and a log built on four hours of curiosity would have been worse than no log at all.
| Log | Games | Status | Why |
|---|---|---|---|
| L1 — Friendly fire and blame | 11 | Open | Finding held across every game with the mechanic |
| L2 — Drop-in and absence | 9 | Open | Clear separation between three design approaches |
| L3 — Voice channel load | 4 | Closed | Measurement changed the behaviour being measured |
| L4 — Matchmaking quality | 0 | Not opened | No honest sample; we always play with people we know |
The friendly-fire log currently claims that blame allocation is a structural property of the system rather than of the players. That is drawn from eleven games and two people, and two people is a very small number of personalities to generalise from.
If it reverses, it will be because a third player joins the file for long enough to show that we were describing our own dynamic. That is noted at the top of the log itself rather than buried here.
