Redesign lazy/super-lazy eval margins with position-specific percentile

tables, and fix four real search/recognizer bugs found while validating it

Main theme: replace the flat, heavily-padded LAZY_EVAL_BASE_MARGIN /
SUPER_LAZY_MARGIN_BY_ARMY constants with per-material-bucket percentile
tables (king safety by attacker army, residual by own army, separate
tables for the lazy vs super-lazy baselines) derived from a 30k-position
calibration pool, targeting a measured ~1-1.5% exceeded-margin rate
instead of an unmeasured "add lots of headroom" floor. QSearch's
FUTILITY_BASE_MARGIN_LAZY/SUPERLAZY were re-derived to match the new,
now-meaningful (p50-based) iPositional this produces. Self-play confirms
the working tree beats both HEAD, a more liberal (~5% miss target)
variant, and a more conservative (near-zero miss) variant of the same
margin mechanism.

Along the way, an initial self-play regression (this margin work bundled
with several changes in one pass, one of which broke things) led to a
day of ZFS-snapshot bisection and then targeted DEBUG-build fuzzing
against a stratified 200k-position corpus (tests/cross_sectional_200k.ep_,
new), which is what actually found and fixed four distinct, real,
pre-existing engine bugs -- none of them related to the margin redesign
itself:

- searchsup.c (CommonSearchInit): three early-return paths (timer
  abort, draw, too-deep bailout) left a node's own PV[] row stale
  instead of clearing it, letting an ancestor's UpdatePV splice a
  leftover move from a completely unrelated node into the displayed
  PV -- root cause of a WalkPV/MakeMove assert ("moving a piece to
  capture, but nothing was on the target square").

- dynamic.c (RecordEnprisePieceAtPly): the two-slot en-prise dedup
  compared (square, piece) instead of square alone, so a same-square
  re-record with a different piece (legitimate -- a fail-high
  capture's victim is restored by UnmakeMove and can differ across
  calls at the same ply) shifted a duplicate square into both slots,
  tripping ValueOfMaterialInTroubleDespiteMove's two-distinct-squares
  assert.

- recogn.c (_RecognizeKBKP): the wrong-colored-bishop corner-fortress
  draw shortcut fired even when the weak side has its own pawn(s),
  which can hand the strong side real zugzwang resources the fortress
  logic doesn't account for -- produced a false hard draw bound where
  EGTB shows a near-forced loss.

- eval.c (_EvalPasserRacesAgainstLoneKings): a rook pawn's race was
  credited as winning even when the strong side's own king sits
  exactly on its own pawn's queening square (distance 0 specifically --
  adjacent is fine, the square is empty) -- a classic stalemate
  resource the race heuristic can't see. Only reachable through
  _RecognizeKPK's pawn-hash-cache-hit path, which is why
  testrecogn.c's TestRecognExhaustiveKPK reported "0 wrong" the whole
  time; extended it to cover both color assignments and the full board
  (no mirror-shortcut) and documented the remaining cache-hit coverage
  gap for a future session to close.

Also: match_play.py gains a local_ebf metric (geometric-mean of
depth-normalized node-growth ratios between consecutive completed
iterations) alongside the existing cumulative nodes**(1/depth) ebf,
which turned out to be misleading when comparing engines with
different lazy-exit aggressiveness -- and probe.c/hash.c pick up two
small, independently-found correctness fixes (TB root-probe promotion
piece missing its color bit; ASSERT(g_pHashTable) crashing on the
legitimate "--hash none" config) noticed along the way.

Co-Authored-By: Claude Sonnet 5 <[email protected]>

1 files changed, 27 insertions, 8 deletions

diff --git a/src/gamelist.c b/src/gamelist.c
index 6adf151..3af04a8 100644
--- a/src/gamelist.c
+++ b/src/gamelist.c
@@ -289,15 +289,34 @@ Return value:
 
 **/
 {
-#ifdef DEBUG
-    static POSITION p;
-
-    memcpy(&p, &g_RootPosition, sizeof(POSITION));
-    
-    return(&p);
-#else
+    // 2026-09-16: used to hand DEBUG builds a pointer into a single
+    // shared `static POSITION p`, freshly overwritten from
+    // g_RootPosition on every call -- presumably meant as a defensive
+    // "give callers a copy, not the real thing," but since it's the
+    // SAME static buffer every call, any caller that stashes the
+    // returned pointer and expects it to stay stable across a later
+    // SetRootPosition() call (e.g. ScriptCommand/_ProcessSuiteLine,
+    // which captures GetRootPosition() once at the top of a script run
+    // and reuses that pointer across every setboard/go line in the
+    // file) silently gets a pointer that can be overwritten out from
+    // under it by any OTHER GetRootPosition() call anywhere in the
+    // program in the meantime -- confirmed as the root cause of a real,
+    // deterministic bug: a scripted "setboard <FEN>; go" pair would
+    // update the real g_RootPosition correctly but Think() would still
+    // see the position from BEFORE the setboard, because its captured
+    // `pos` pointer pointed at this shared buffer, snapshotted before
+    // the setboard ran and never refreshed. Confirmed via split.c: a
+    // split created deep in the (wrongly-rooted) search captured its
+    // OWN root snapshot via a LATER, independent GetRootPosition() call
+    // that correctly saw the real (already-updated) position, producing
+    // a split whose recorded move path and recorded root position
+    // belonged to two different positions entirely. DEBUG-only bug --
+    // release builds always used this same direct-pointer path and were
+    // never affected. No actual defensive value survives removing the
+    // indirection: matching the release build's behavior exactly (a
+    // stable pointer to the one real g_RootPosition, always current) is
+    // strictly more correct, not less safe.
     return(&g_RootPosition);
-#endif
 }
 
 void