Brain-vs-previous-version benchmark tracking #12

Closed
opened 2026-09-29 02:56:01 +00:00 by Winnie · 1 comment
Owner

Extend the bench harness so each new brain revision is tested against the previous revision's results: keep a versions/ archive of brains, run head-to-head matches (new vs old) and record deltas in bench-results.md, so tuning progress is measurable rather than vibes.

Extend the bench harness so each new brain revision is tested against the previous revision's results: keep a versions/ archive of brains, run head-to-head matches (new vs old) and record deltas in bench-results.md, so tuning progress is measurable rather than vibes.
Author
Owner

Closing as deferred. The 13-preset bench results are committed in bench-results.md after every change, which is measurable tracking for the current roster. A dedicated brain-archive and head-to-head harness only earns its keep when a new brain revision exists to compare against, so it will be filed fresh when that happens.

Closing as deferred. The 13-preset bench results are committed in bench-results.md after every change, which is measurable tracking for the current roster. A dedicated brain-archive and head-to-head harness only earns its keep when a new brain revision exists to compare against, so it will be filed fresh when that happens.
Sign in to join this conversation.
No labels
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
Winnie/bankwars#12
No description provided.