logaritmisk
  • Joined on 2023-10-06
logaritmisk closed issue logaritmisk/trueskill-tt#36 2026-09-01 17:32:19 +00:00
just release-plan leaves the tree dirty, so just release always aborts
logaritmisk pushed to main at logaritmisk/trueskill-tt 2026-09-01 17:32:19 +00:00
87fca8dcca docs: correct drifted documentation and compile the README in CI
ef62b57a08 fix(release): skip the changelog hook during a dry run
Compare 2 commits »
logaritmisk closed issue logaritmisk/trueskill-tt#35 2026-09-01 17:32:19 +00:00
Documentation has drifted from the code: four README examples no longer compile
logaritmisk opened issue logaritmisk/trueskill-tt#36 2026-09-01 04:39:24 +00:00
just release-plan leaves the tree dirty, so just release always aborts
logaritmisk opened issue logaritmisk/trueskill-tt#35 2026-09-01 04:39:07 +00:00
Documentation has drifted from the code: four README examples no longer compile
logaritmisk pushed to main at logaritmisk/trueskill-tt 2026-09-01 04:34:39 +00:00
b2a7ade10c chore: Release trueskill-tt version 0.3.0
617bc07f6f feat: allow drift to vary per competitor via Member::with_drift_scale
Compare 2 commits »
logaritmisk pushed tag v0.3.0 to logaritmisk/trueskill-tt 2026-09-01 04:34:39 +00:00
logaritmisk closed issue logaritmisk/trueskill-tt#34 2026-09-01 04:34:38 +00:00
Allow drift to vary per competitor: Member::with_drift_scale
logaritmisk opened issue logaritmisk/trueskill-tt#34 2026-09-01 04:11:11 +00:00
Allow drift to vary per competitor: Member::with_drift_scale
logaritmisk pushed to main at logaritmisk/trueskill-tt 2026-08-28 06:27:26 +00:00
1a88678384 refactor!: remove ConvergenceReport::slices_skipped
logaritmisk closed issue logaritmisk/trueskill-tt#33 2026-08-28 06:27:25 +00:00
ConvergenceReport.slices_skipped is public, always zero, and nothing will ever populate it
logaritmisk opened issue logaritmisk/trueskill-tt#33 2026-08-28 06:02:10 +00:00
ConvergenceReport.slices_skipped is public, always zero, and nothing will ever populate it
logaritmisk closed issue logaritmisk/trueskill-tt#32 2026-08-28 06:01:55 +00:00
Ingestion discards all convergence state, so appending one event costs a full re-converge
logaritmisk commented on issue logaritmisk/trueskill-tt#32 2026-08-27 22:18:22 +00:00
Ingestion discards all convergence state, so appending one event costs a full re-converge

Spiked. The premise of this issue is wrong — the re-convergence is necessary work, not waste. Recommending we close it.

I filed this off the observation that appending one event costs a full…

logaritmisk closed issue logaritmisk/trueskill-tt#4 2026-08-27 21:50:46 +00:00
Cross-slice parallelism: per-slice dirty bits to skip clean slices
logaritmisk opened issue logaritmisk/trueskill-tt#32 2026-08-27 21:50:42 +00:00
Ingestion discards all convergence state, so appending one event costs a full re-converge
logaritmisk commented on issue logaritmisk/trueskill-tt#4 2026-08-27 21:50:20 +00:00
Cross-slice parallelism: per-slice dirty bits to skip clean slices

Closing: measured, and this approach cannot pay for itself. The design is also unsound in three independent ways, but the measurement alone settles it.

Work is on branch dirty-slice-skipping

logaritmisk pushed to main at logaritmisk/trueskill-tt 2026-08-27 16:41:51 +00:00
5f46296671 chore: ignore proptest regression seed files
2745fbb622 test: add property-based tests, a shared finiteness helper, and boundary inputs
6d2573b92e perf: make the per-slice SkillStore compact instead of dense
1ac3b21db5 fix: enforce EventBuilder weight/team length in release
4e043364fd perf: stop cloning inference inputs in OwnedGame and ingestion
Compare 8 commits »
logaritmisk commented on issue logaritmisk/trueskill-tt#26 2026-08-27 16:06:43 +00:00
Test suite has no coverage for degenerate inputs, release-mode behaviour, or non-finite results

Three of the four remaining items done — 2745fbb. Staying open for the fourth, which needs a decision rather than work.

Property-based teststests/properties.rs, proptest as a…

logaritmisk closed issue logaritmisk/trueskill-tt#17 2026-08-27 16:02:16 +00:00
Per-slice SkillStore is dense over the global index space — memory scales with the largest competitor index, not slice size