qvm: recover leaderboard selection

docs/ASSET_PIPELINE_TODO.md

@@ -973,3 +973,6 @@ Research progress: RE-771 [report](research/QVM_RANKING_COMPLETE.md); original a
 
 
 Research progress: RE-772 [report](research/QVM_LEADERBOARD_LAYOUT.md); original assets unchanged.
+
+
+Research progress: RE-773 [report](research/QVM_LEADERBOARD_SELECTOR.md); original assets unchanged.

docs/FINDINGS.md

@@ -1319,3 +1319,6 @@ RE-725–734 batch: [health, contact and Rust impulse findings](research/QVM_HEA
 
 
 **RE-772 — Leaderboard slot mapping sparse layout and sentinels:** 278587 maps20 weapon-like identifiers to slots0..19;284555 initializes nine metric fields and preserves record holes. [Evidence](research/QVM_LEADERBOARD_LAYOUT.md).
+
+
+**RE-773 — Complete leaderboard joint thresholds sentinel updates and ties:** 284610 requires every candidate metric to strictly exceed its threshold, then updates only nonsentinel fields and records the candidate entity. [Evidence](research/QVM_LEADERBOARD_SELECTOR.md).

docs/PROVISIONAL_TODO.md

@@ -1317,3 +1317,6 @@ Research progress: RE-771 [report](research/QVM_RANKING_COMPLETE.md); full nativ
 
 
 Research progress: RE-772 [report](research/QVM_LEADERBOARD_LAYOUT.md); full native/platform/game scope remains open.
+
+
+Research progress: RE-773 [report](research/QVM_LEADERBOARD_SELECTOR.md); full native/platform/game scope remains open.

docs/RESEARCH_LEDGER.md

@@ -1518,3 +1518,8 @@ Exactly ten task commits follow baseline `27a7f96f`. Every completed task passed
 ## 2026-10-21 — RE-772 Leaderboard slot mapping sparse layout and sentinels
 
 278587 maps20 weapon-like identifiers to slots0..19;284555 initializes nine metric fields and preserves record holes. Focused `cargo xtask test-group leaderboard-layout` passes (3 Python tests); selected reports regenerate twice byte-identically, with storage, links and whitespace guards. New report: 14920 bytes. Full-suite validation is deferred to the tenth batch commit. No original assets, Rust runtime or shared VM evaluator changed.
+
+
+## 2026-10-21 — RE-773 Complete leaderboard joint thresholds sentinel updates and ties
+
+284610 requires every candidate metric to strictly exceed its threshold, then updates only nonsentinel fields and records the candidate entity. Focused `cargo xtask test-group leaderboard-selector` passes (3 Python tests); selected reports regenerate twice byte-identically, with storage, links and whitespace guards. New report: 697619 bytes. Full-suite validation is deferred to the tenth batch commit. No original assets, Rust runtime or shared VM evaluator changed.

docs/REVERSE_ENGINEERING_TODO.md

@@ -1363,3 +1363,6 @@ Purpose: turn the supplied installation into an evidence-backed specification of
 
 
 - [x] RE-772 Leaderboard slot mapping sparse layout and sentinels ([report](research/QVM_LEADERBOARD_LAYOUT.md)).
+
+
+- [x] RE-773 Complete leaderboard joint thresholds sentinel updates and ties ([report](research/QVM_LEADERBOARD_SELECTOR.md)).

docs/research/QVM_LEADERBOARD_SELECTOR.md

@@ -0,0 +1,7 @@
+# Complete leaderboard joint thresholds sentinel updates and ties
+
+RE-773. Reproduce with `python3 scripts/analyze_qvm_leaderboard_selector.py`; validate with `cargo xtask test-group leaderboard-selector`. See the [machine report](qvm-leaderboard-selector.json).
+
+Recovered complete284610 with actual278587 and corrected repeated-call reentry over77 configurations. All nine metrics jointly gate a candidate, including sentinel-valued inactive fields; updates preserve inactive sentinels. All-sentinel records never select despite candidates passing gates. Scan order and strict ties are explicit. Invalid keys, live-null clients and original runtime parity remain open.
+
+No original executable/full VM/native world collision or rendering runs. Supplied callback/native responses and fixture domains remain explicit; complete macOS/Linux Rust/Bevy/Solari gameplay remains open.

docs/research/README.md

@@ -909,3 +909,6 @@ The extension collector adds five focused tests for native tail calls, command-r
 
 
 [RE-772: Leaderboard slot mapping sparse layout and sentinels](QVM_LEADERBOARD_LAYOUT.md).
+
+
+[RE-773: Complete leaderboard joint thresholds sentinel updates and ties](QVM_LEADERBOARD_SELECTOR.md).

scripts/analyze_qvm_leaderboard_selector.py

@@ -0,0 +1,15 @@
+#!/usr/bin/env python3
+"""Complete284610 strict joint metrics with actual lookup and sentinel rules."""
+from pathlib import Path
+from qvm_research_support import source
+from qvm_leaderboard_support import FIELDS,RECORD
+from qvm_leaderboard_selector_support import fixture,select,KEYS
+from analyze_qvm import write_json
+
+def main():
+    q,raw,h,ins=source();rows=[]
+    for active in [[o] for o in FIELDS]+[FIELDS,[]]:
+        for pattern in range(7):
+            m,candidates=fixture(ins,raw,h,active,pattern);_,_,p=select(ins,m,{8:RECORD,12:KEYS,16:2});p.update(input={'active':active,'pattern':pattern},candidates=candidates);rows.append(p)
+    write_json(Path('docs/research/qvm-leaderboard-selector.json'),{'source':q['source'],'sha256':q['sha256'],'probes':rows,'limits':'Complete284610 with actual278587 and corrected repeated-call tracing.64 entity slots,4 live clients,2 valid keys, finite float metrics. All nine candidate metrics must strictly exceed record thresholds; only nonsentinel fields update. Invalid keys/live-null clients and native/runtime parity remain open.'})
+if __name__=='__main__':main()

scripts/qvm_leaderboard_selector_support.py

@@ -0,0 +1,33 @@
+"""284610 source strict joint-threshold selector with repeated-call tracing."""
+from qvm_health_effects_support import H
+from qvm_notification_support import POOL,CPOOL
+from qvm_leaderboard_support import mapping_seed,initialize,mapping,RECORD,FIELDS
+from qvm_trace_support import resume
+from analyze_qvm_acceleration import word
+SEL=H+0x20000
+KEYS=SEL+0x1000
+SLOTS=[0,2,5,63]
+
+def select(ins,m,args):
+    for o,v in args.items():m[SEL+48+o]=v
+    before={str(o):m[args[8]+o]&0xffffffff for o in FIELDS};m,e=resume(ins,284611,285141,m,SEL,handlers={278587:mapping},limit=50000,repeated_calls=True)
+    return m,e['return'],{'initial':before,'final':{str(o):m[args[8]+o]&0xffffffff for o in FIELDS},'selected':e['return'],'execution':e}
+
+def fixture(ins,raw,h,active,pattern):
+    m=mapping_seed(raw,h);m.update({RECORD+o:0x55555555 for o in range(0,76,4)});m,_,_=initialize(ins,m,{8:RECORD})
+    for o in active:m[RECORD+o]=word(0) if o==56 else 0
+    for j in range(64):m[POOL+880*j+532]=int(j in SLOTS);m[POOL+880*j+520]=CPOOL+6336*j
+    for k,key in enumerate([2,3]):m[KEYS+4*k]=key
+    candidates=[]
+    for j,id in enumerate(SLOTS):
+        for k,slot in enumerate([9,8]):
+            metrics={str(o):word(j+k+1) if o==56 else j+k+1 for o in FIELDS}
+            if pattern==1:metrics={str(o):word(5-j-k) if o==56 else 5-j-k for o in FIELDS}
+            if pattern==2:metrics={str(o):word(1) if o==56 else 1 for o in FIELDS}
+            if pattern==3:metrics['0']=0xffffffff if j%2 else metrics['0']
+            if pattern==4:metrics['56']=word(-0.5 if j%2 else 0)
+            if pattern==5:metrics['0']=5-j;metrics['8']=j+1
+            if pattern==6:metrics={str(o):word(100+j) if o==56 else 0x7fffffff-j for o in FIELDS}
+            for o in FIELDS:m[CPOOL+6336*id+856+76*slot+o]=metrics[str(o)]
+            candidates.append({'entity':id,'key':2+k,'slot':slot,'metrics':metrics})
+    return m,candidates

scripts/test_qvm_leaderboard_selector.py

@@ -0,0 +1,26 @@
+import json,unittest
+from pathlib import Path
+from test_qvm_stat_totals import signed
+from test_qvm_ground import n
+
+def expected(initial,candidates):
+    r=dict(initial);selected=-1
+    for c in candidates:
+        m=c['metrics']
+        if not all(n(m[o])>n(r[o]) if o=='56' else signed(m[o])>signed(r[o]) for o in r):continue
+        for o in r:
+            if r[o]!=(0xbf800000 if o=='56' else 0xffffffff):r[o]=m[o];selected=c['entity']
+    return selected,r
+class LeaderboardSelectorTests(unittest.TestCase):
+    @classmethod
+    def setUpClass(cls):cls.ps=json.loads(Path('docs/research/qvm-leaderboard-selector.json').read_text())['probes']
+    def test_independent_joint_threshold_and_sentinel_oracle(self):
+        for p in self.ps:self.assertEqual((p['selected'],p['final']),expected(p['initial'],p['candidates']),p['input'])
+    def test_original_mapping_call_for_every_candidate(self):
+        for p in self.ps:self.assertEqual(len(p['execution']['calls']),8);self.assertTrue(all(c['target']==278587 for c in p['execution']['calls']));self.assertEqual(p['execution']['return'],p['selected'])
+    def test_sentinel_only_never_selects_and_strict_ties(self):
+        self.assertEqual(len(self.ps),77)
+        for p in self.ps:
+            if not p['input']['active']:self.assertEqual(p['selected'],-1);self.assertEqual(p['final'],p['initial'])
+            elif p['input']['pattern']==2:self.assertEqual(p['selected'],0)
+if __name__=='__main__':unittest.main()

xtask/checks.tsv

@@ -1335,3 +1335,6 @@ group	ranking-complete	scripts/analyze_qvm_ranking_complete.py
 report	scripts/analyze_qvm_leaderboard_layout.py	docs/research/qvm-leaderboard-layout.json
 edge	docs/research/qvm-leaderboard-layout.json	scripts/test_qvm_leaderboard_layout.py
 group	leaderboard-layout	scripts/analyze_qvm_leaderboard_layout.py
+report	scripts/analyze_qvm_leaderboard_selector.py	docs/research/qvm-leaderboard-selector.json
+edge	docs/research/qvm-leaderboard-selector.json	scripts/test_qvm_leaderboard_selector.py
+group	leaderboard-selector	scripts/analyze_qvm_leaderboard_selector.py