From fa2169b9c73dd94d4f899b41d12c87744acbc7e7 Mon Sep 17 00:00:00 2001 From: Claude Date: Mon, 12 Oct 2026 18:34:55 +0000 Subject: [PATCH] docs(docs): List the shallower re-search after a root fail high as rejected MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit -4 ±6 over 6,000 games at 10+0.1 against d82638a, sprt [0, 5] failed at LLR -3.11. The candidate was 53ee5ec. Co-Authored-By: Claude Opus 5.5 Claude-Session: https://claude.ai/code/session_0133NvZMXKrttWcNCNU9xXjZ --- docs/ROADMAP.md | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/docs/ROADMAP.md b/docs/ROADMAP.md index c7da95d5..56be6346 100644 --- a/docs/ROADMAP.md +++ b/docs/ROADMAP.md @@ -709,3 +709,12 @@ of these again without saying what is different this time. games at 10+0.1 against 54fb9a6 (sprt [0, 10] failed at LLR -3.01) at equal time (end clock -16 ms). In practice it only moved a tenth of the time from the first ten moves after the book to the next thirty; from move 60 both rules sit at the floor of 10. +- Re-searching a root fail high a ply shallower for each fail high since the depth + opened or last failed low, a fail low restoring the full depth. The move a fail high + names was the depth's answer after a full depth re-search at 89 of 100 such depths on + the games suite, and replayed one depth at a time the shorter re-search made those + depths a median 0.53 to 0.55 of their nodes on every position of the games suite and + the bench. -4 ±6 over 6,000 games at 10+0.1 against d82638a (sprt [0, 5] failed at LLR + -3.11 on the last batch the cap allowed), with 1.04 times master's nodes at the same + time. The tactical count fell from 228 to 217 at depth six. One ply for the whole + re-search, however many fail highs, is a different change to measure.