logaritmisk
  • Joined on 2023-10-06
logaritmisk closed issue logaritmisk/trueskill-tt#37 2026-09-08 13:53:48 +00:00
EventBuilder and record_winner/record_draw cannot set prior or drift_scale
logaritmisk closed issue logaritmisk/trueskill-tt#5 2026-09-08 13:53:48 +00:00
Decide whether rayon becomes a default-on feature
logaritmisk pushed to main at logaritmisk/trueskill-tt 2026-09-08 13:53:48 +00:00
7c6965c6a9 Merge branch 'fix/ingestion-shape'
911b48faba feat: add EventBuilder::members for per-member configuration
f57784c141 docs: record the rayon opt-in deviation in spec section 6
8e4d6a637d fix: reject malformed events at the ingestion boundary
Compare 4 commits »
logaritmisk commented on issue logaritmisk/trueskill-tt#51 2026-09-08 08:37:09 +00:00
Cache the joint factorisation: repeat posterior_of queries re-pay the full O(n^3) solve

Shipped in 0.7.0 and wired up here. Measured on the real export, round-grain connectivity — 11 competitors, 55 pairs:

379 variables
factorise   4.3 ms
55 queries  1.6 ms   (~29 us each)
-----…
logaritmisk commented on issue logaritmisk/trueskill-tt#51 2026-09-08 08:27:42 +00:00
Cache the joint factorisation: repeat posterior_of queries re-pay the full O(n^3) solve

Released in 0.7.0, on the registry now.

Minor rather than patch: History::joint, Joint and Joint::variables are new public surface. Nothing was removed and no existing signature…

logaritmisk pushed to main at logaritmisk/trueskill-tt 2026-09-08 08:27:30 +00:00
82eff740b6 chore: Release trueskill-tt version 0.7.0
logaritmisk pushed tag v0.7.0 to logaritmisk/trueskill-tt 2026-09-08 08:27:30 +00:00
logaritmisk closed issue logaritmisk/trueskill-tt#51 2026-09-08 08:24:46 +00:00
Cache the joint factorisation: repeat posterior_of queries re-pay the full O(n^3) solve
logaritmisk pushed to main at logaritmisk/trueskill-tt 2026-09-08 08:24:46 +00:00
c1b1c6c7d7 Merge branch 'feat/joint-handle'
1bb6bb31d8 feat: factorise the joint once with History::joint
Compare 2 commits »
logaritmisk commented on issue logaritmisk/trueskill-tt#51 2026-09-08 05:55:25 +00:00
Cache the joint factorisation: repeat posterior_of queries re-pay the full O(n^3) solve

Implemented, on feat/joint-handle. Not merged or released yet.

Thank you for the flat four-solve trace — that is the measurement that makes this a bug report rather than a feature request, and…

logaritmisk opened issue logaritmisk/trueskill-tt#52 2026-09-08 05:54:35 +00:00
The joint precision matrix is 0.19% dense and we factorise it densely
logaritmisk commented on issue logaritmisk/trueskill-tt#43 2026-09-08 05:47:10 +00:00
No public way to ask "what is P(skill below x)?" — cdf is pub(crate), so callers approximate it with a z-multiplier

Adopted in lester-stash on 0.5.0 — thank you. One correction to this issue, since the reasoning in it was partly wrong and I'd rather not leave that standing.

**The claim that a z-band "silently…

logaritmisk commented on issue logaritmisk/trueskill-tt#51 2026-09-08 05:45:36 +00:00
Cache the joint factorisation: repeat posterior_of queries re-pay the full O(n^3) solve

Correction to the figure above, and a finding that is probably more useful than the original request.

I said our 90-cell grid is "about nine minutes". That is wrong. It took the 6 s rate from…

logaritmisk opened issue logaritmisk/trueskill-tt#51 2026-09-08 05:41:29 +00:00
Cache the joint factorisation: repeat posterior_of queries re-pay the full O(n^3) solve
logaritmisk commented on issue logaritmisk/trueskill-tt#47 2026-09-08 04:57:38 +00:00
A node's reported sigma is ~10x wider than the estimate's measured reproducibility

Run against 0.6.0. All 55 pairs resolve, and the answer is that my 1.30 was wrong — the marginal was already right.

The number

exact / marginal, 55 pairs, real export:  1.035

So…

logaritmisk closed issue logaritmisk/trueskill-tt#46 2026-09-08 04:49:07 +00:00
No joint posterior: anything derived from two or more competitors has to add marginals and overstate its width
logaritmisk commented on issue logaritmisk/trueskill-tt#46 2026-09-08 04:47:20 +00:00
No joint posterior: anything derived from two or more competitors has to add marginals and overstate its width

Closing as delivered — and no follow-up issue is needed, because the slice question is fixed rather than deferred.

0.6.0 makes the joint time-expanded: one variable per appearance,…

logaritmisk commented on issue logaritmisk/trueskill-tt#47 2026-09-08 04:47:01 +00:00
A node's reported sigma is ~10x wider than the estimate's measured reproducibility

Fixed in 0.6.0. The latest-slice limitation was my mistake, not a documentation gap, and your measurement is what caught it.

What was wrong

I built the joint over a single slice and…

logaritmisk pushed tag v0.6.0 to logaritmisk/trueskill-tt 2026-09-08 04:46:44 +00:00
logaritmisk pushed to main at logaritmisk/trueskill-tt 2026-09-08 04:46:43 +00:00
b113385c6f chore: Release trueskill-tt version 0.6.0