Durability tips for the docs: WAL off, flush levels, and page cache against the defaults, per wheel #354

Problem

I want measured durability advice for the docs: what WAL off, each flush level, and a larger page cache cost or buy against the defaults, for both wheels.

It is one row of the bindings study list. It was held on stale pins and re-pins after the 26.11.1 release wheels exist.

Priority: P3 (not on any deadline).

Acceptance

• The dry run is clean and its report is written
• The study runs on the release pair and is scored
• The docs gain a durability tips section whose figures come from the scorer output

Results and the scorer output go in this issue as comments. Numbers are generated by script from the result files, never typed.