Fix trump-void1 move ordering to restore v2.9 search efficiency#213
Open
tameware wants to merge 2 commits into
Open
Fix trump-void1 move ordering to restore v2.9 search efficiency#213tameware wants to merge 2 commits into
tameware wants to merge 2 commits into
Conversation
The heuristic extraction refactor changed weight_alloc_trump_void1's first branch from `lead_suit == trump` to `suit == trump`. Since that is exhaustive with the following `else if (suit != trump)`, the three ruffing branches (using the `24 - rank + ...` formula) became dead code, and trump ruffs were scored with side-suit discard weights instead. This mis-ordered ruffs, costing alpha-beta cutoffs. The effect is small for solve but compounds heavily in calc's warm-TT iterative deepening: calc explored ~34% more nodes than v2.9. Restoring the original `lead_suit == trump` pitch branch makes the ruffing branches reachable again and cuts calc time ~25% (gap to v2.9: 1.37x -> 1.02x). Ordering-only change; double-dummy results are unchanged. Co-authored-by: Cursor <cursoragent@cursor.com>
Contributor
There was a problem hiding this comment.
Pull request overview
Restores the intended move-ordering behavior in weight_alloc_trump_void1() by fixing the pitch/ruff branch condition so that trump-ruff scoring logic is reachable again (aligning with the original v2.9-era heuristic behavior and improving alpha-beta cutoff efficiency).
Changes:
- Switches the first-branch condition in
weight_alloc_trump_void1()fromsuit == trumptolead_suit == trumpto correctly model the “lead is trump → we must pitch (void in trump)” case. - Re-enables the subsequent trump-ruff weighting branches that were previously made unreachable by the refactor.
Per Copilot. Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Opus 4.8's fix for 3.0's performance regression v 2.9.
The heuristic extraction refactor changed weight_alloc_trump_void1's first branch from
lead_suit == trumptosuit == trump. Since that is exhaustive with the followingelse if (suit != trump), the three ruffing branches (using the24 - rank + ...formula) became dead code, and trump ruffs were scored with side-suit discard weights instead. This mis-ordered ruffs, costing alpha-beta cutoffs.The effect is small for solve but compounds heavily in calc's warm-TT iterative deepening: calc explored ~34% more nodes than v2.9. Restoring the original
lead_suit == trumppitch branch makes the ruffingbranches reachable again and cuts calc time ~25% (gap to v2.9: 1.37x ->
1.02x). Ordering-only change; double-dummy results are unchanged.
Benchmarks. Compare is the PR code, branch is the develop head.