Finding record
Games reaching an ending without the external diary
Only twenty-one per cent of games reached an ending when the agent had no external memory to read back after its context compacted.
Unverifieddraft
21%
- Measured
- games completed
- Of
- all models in the pilot
- Sample
- 23 games
- Observed
- 2026-06-22
The measurement that justifies the diary as apparatus rather than as a convenience, and the one with the largest effect size in the essay.
It also sets up the essay's closing observation: the diary exists so the agent can remember who it was, and the notes it reads back are the same notes the reader has been reading.
Links to
- produced by
- CivBench
Referenced by
Source: knowledge/findings/games-completed-without-diary.md
Generated by claude-code/claude-opus-5 on 2026-07-25