A review can only describe one game. These two pieces count across every game the publication has logged, which is where the findings live.
Both pieces
Counting across games rather than within one
Each piece is revised as new frame logs arrive, and each carries the specific evidence that would overturn it. One of the two has already been narrowed once.
The frames that survive a week have consistently less in them than the frames that do not, and photo modes are almost universally built to encourage the opposite. Fourteen games' worth of counting, and one narrowing.
Nineteen games with a time-of-day control, and in seventeen of them a single forty-minute window produces the majority of usable frames. What that concentration says about the rest of the day.
19 gamesTwo revised in a patchOpen
Standing figures
What both pieces currently rest on
Summarised so the direction of the counting is visible without reading either piece in full.
Sort that table by tools and it looks like noise. Sort it by world, as it is sorted above, and it is almost monotonic: every game scoring eight or nine for its world sits in the top eight rows, and every game scoring three or four sits in the bottom three.
The strongest single line in the table is the bottom one against the top one. Causeway has nine for tools and four for world and keeps one frame in eighty. Glasshouse Rose has one for tools and nine for world and keeps one frame in ten. That is the whole publication in two rows, and the essay is an attempt to say why more carefully than that.
What the table cannot show is the interaction, which is where Moss And Marble matters. It is the only game in fourteen where a specific tool — a 135mm lens — measurably rescued frames that would otherwise have failed, and its internal figures rise from 1.8 per cent at the wide end to 16.7 at the long. Tools are not irrelevant. They are just enormously less important than a sky that changes.
Across all logs
4772frames logged to date
Fourteen games across three issues, every frame recorded at the time of capture and reviewed a week later.
Kept by world score
World 8–97.8%
World 6–76.4%
World 3–52.2%
Median kept ratio within each band. The gap between the top two bands is small; the gap to the third is not.
Kept by tools score
Tools 8–103.7%
Tools 5–76.4%
Tools 1–47.1%
Inverted, which is the finding that started the essay and the one most likely to be an artefact of how frames get taken.