qvm: recover ranked tie predicate

docs/FINDINGS.md

@@ -1499,3 +1499,6 @@ RE-813–822 replaces the scoreboard decimal assumption with actual37914 executi
 
 
 **RE-833:** Rust now composes the full recovered7805 priority and matches438 continuous source cases, including pending clears and mode9 reset/event intentions. [Evidence](research/RUST_EXIT_COMPLETE.md).
+
+
+**RE-834:** Actual 7662 and child 33715 execute across 168 cases. Playing count and the enabled global gate ties; team-classified modes compare global scores, mode9 uses ranked field308, mode11 field696, and other modes field248. Raw numeric names remain conservative; ranked pointer legality and native execution remain open. [Evidence](research/QVM_EXIT_PREDICATE.md).

docs/PROVISIONAL_TODO.md

@@ -1515,3 +1515,6 @@ RE-833: Rust now composes the full recovered7805 priority and matches438 continu
 
 
 - [ ] RE-834: Recover supplied exit predicate/transition/native effects, compose ranking→exit→scoreboard state, and validate actual runtime traces and macOS/Linux rendering. [Batch scope](research/QVM_FLOAT_EXIT_RULES_BATCH.md).
+
+
+RE-834: Actual 7662 and child 33715 execute across 168 cases. Playing count and the enabled global gate ties; team-classified modes compare global scores, mode9 uses ranked field308, mode11 field696, and other modes field248. Raw numeric names remain conservative; ranked pointer legality and native execution remain open. Full runtime completion remains open. [Evidence](research/QVM_EXIT_PREDICATE.md).

docs/RESEARCH_LEDGER.md

@@ -1839,3 +1839,8 @@ Rust now composes the full recovered7805 priority and matches438 continuous sour
 RE-833 final review: added five explicit source-to-Rust dependency edges and registry regression assertions; the targeted registry test and repeated `rust-exit-complete` focused group pass. Added the ten-task batch summary and RE-834 follow-up scope.
 
 Batch RE-824–833 boundary: the single `cargo xtask test-full --since 1ed43d40` run passed after task ten: 1,628 Python tests, 127 Rust tests (assets 10, gameplay 89, xtask 28), 43 deterministic report producers, storage policy, 392 documentation files and whitespace checks. No generated report drift; validation notes amended into task ten.
+
+
+## 2026-10-27 — RE-834 Original exit tie predicate and mode classifier
+
+Actual 7662 and child 33715 execute across 168 cases. Playing count and the enabled global gate ties; team-classified modes compare global scores, mode9 uses ranked field308, mode11 field696, and other modes field248. Raw numeric names remain conservative; ranked pointer legality and native execution remain open. Focused `cargo xtask test-group exit-predicate` passes (5 Python tests); selected reports regenerate twice deterministically, with storage, documentation links and whitespace guards. Full suite deferred to task ten. Original assets unchanged; source fixtures do not establish complete native/runtime parity.

docs/REVERSE_ENGINEERING_TODO.md

@@ -1552,3 +1552,6 @@ Purpose: turn the supplied installation into an evidence-backed specification of
 
 
 - [ ] RE-834: Recover supplied exit predicate/transition/native effects, compose ranking→exit→scoreboard state, and validate actual runtime traces and macOS/Linux rendering. [Batch scope](research/QVM_FLOAT_EXIT_RULES_BATCH.md).
+
+
+- [x] RE-834 Original exit tie predicate and mode classifier ([evidence](research/QVM_EXIT_PREDICATE.md)).

docs/research/QVM_EXIT_PREDICATE.md

@@ -0,0 +1,7 @@
+# Original exit tie predicate and mode classifier
+
+RE-834. Reproduce with `python3 scripts/analyze_qvm_exit_predicate.py`; focused validation: `cargo xtask test-group exit-predicate`. [Machine report](qvm-exit-predicate.json).
+
+Actual 7662 and child 33715 execute across 168 cases. Playing count and the enabled global gate ties; team-classified modes compare global scores, mode9 uses ranked field308, mode11 field696, and other modes field248. Raw numeric names remain conservative; ranked pointer legality and native execution remain open.
+
+Original executable/native runtime and complete Linux/macOS Rust/Bevy/Solari acceptance remain unverified.

docs/research/README.md

@@ -1091,3 +1091,6 @@ The extension collector adds five focused tests for native tail calls, command-r
 [RE-833: Rust complete recovered exit-rule decision and mutation composition](RUST_EXIT_COMPLETE.md).
 
 - [RE-824–833 float formatter and exit-rule batch](QVM_FLOAT_EXIT_RULES_BATCH.md).
+
+
+[RE-834: Original exit tie predicate and mode classifier](QVM_EXIT_PREDICATE.md).

docs/research/qvm-resumer-audit.json

@@ -8830,6 +8830,31 @@
       "repeated_calls_opt_in": false,
       "repeated_calls_setting": "default_false"
     },
+    {
+      "backward_conditional_branches": [],
+      "calls": [],
+      "explicit_result_offsets": null,
+      "helper": "qvm_trace_support.resume",
+      "line": 18,
+      "literal_range": [
+        33716,
+        33740
+      ],
+      "path": "scripts/qvm_intermission_support.py",
+      "repeated_calls_opt_in": true,
+      "repeated_calls_setting": "literal_true"
+    },
+    {
+      "backward_conditional_branches": [],
+      "calls": [],
+      "explicit_result_offsets": null,
+      "helper": "qvm_trace_support.resume",
+      "line": 25,
+      "literal_range": null,
+      "path": "scripts/qvm_intermission_support.py",
+      "repeated_calls_opt_in": true,
+      "repeated_calls_setting": "literal_true"
+    },
     {
       "backward_conditional_branches": [
         285126,
@@ -9989,10 +10014,10 @@
     }
   },
   "summary": {
-    "call_expressions": 199,
-    "files": 152,
+    "call_expressions": 201,
+    "files": 153,
     "legacy_expressions": 158,
-    "nonliteral_ranges": 25,
+    "nonliteral_ranges": 26,
     "unresolved_mode_expressions": 3
   }
 }

scripts/analyze_qvm_exit_predicate.py

@@ -0,0 +1,18 @@
+#!/usr/bin/env python3
+import itertools
+from pathlib import Path
+from qvm_research_support import source
+from qvm_intermission_support import seed,execute
+from analyze_qvm import write_json
+
+def cases():
+    for playing,enabled,mode,pair,order in itertools.product([1,2],[0,1],[-1,0,3,8,9,11,12],[(0,0),(1,2),(-1,-1)],[[0,1],[1,0]]):
+        clients=[{'248':pair[0],'308':pair[1],'696':7},{'248':pair[1],'308':pair[0],'696':7 if pair[0]==pair[1] else 8}]
+        yield dict(playing=playing,enabled=enabled,mode=mode,first_score=pair[0],second_score=pair[1],order=order),clients
+
+def main():
+    q,raw,h,ins=source();rows=[]
+    for a,c in cases():
+        m,p=execute(ins,seed(c,**a),7662,7805,28);rows.append({'input':a,'clients':c,**p})
+    write_json(Path('docs/research/qvm-exit-predicate.json'),{'source':q['source'],'sha256':q['sha256'],'probes':rows,'limits':'Actual7662 with actual33715 mode classifier, signed gates and ranked-client field selection.168 finite fixtures. Original native runtime and raw invalid rank pointers remain unverified.'})
+if __name__=='__main__':main()

scripts/qvm_intermission_support.py

@@ -0,0 +1,31 @@
+"""Original predicate and intermission helpers; external children remain captured."""
+from qvm_exit_rules_support import seed as exit_seed,E,CPOOL,POOL
+from qvm_trace_support import resume,captured
+EXTRA=[452,4036,4324,6156]
+
+def seed(clients=(),**a):
+    m=exit_seed(clients,**a)
+    m.update({98884:a.get('enabled',1),34992592:a.get('first',0),34992596:a.get('second',1),34999604:a.get('waiting',0),34999608:a.get('wait_start',0),34992580:a.get('rank_count',len(clients))})
+    for j,c in enumerate(clients):
+        for o in EXTRA:m[CPOOL+6336*j+o]=c.get(str(o),0)
+        m[POOL+880*j+424]=c.get('entity_flags',0)
+        m[POOL+880*j+520]=CPOOL+6336*j
+        m[34992592+4*j]=a.get('order',list(range(len(clients))))[j]
+    return m
+
+def classify(ins,m,a):
+    m[E+0x1000+20]=a[8]
+    m,e=resume(ins,33716,33740,m,E+0x1000,repeated_calls=True,sync_child_words=True)
+    return m,e['return'],{'source_child':33715,'return':e['return']}
+
+def execute(ins,m,start,stop,frame,args=None,targets=(),handlers=None):
+    for o,v in (args or {}).items():m[E+frame+o]=v
+    initial=dict(m);requests=[]
+    hs={t:captured(t,requests) for t in targets};hs[33715]=classify;hs.update(handlers or {})
+    m,e=resume(ins,start+1,stop,m,E,handlers=hs,repeated_calls=True,sync_child_words=True,segment_limit=4096,limit=32768)
+    changes=[{'address':a,'before':v&0xffffffff,'after':m[a]&0xffffffff} for a,v in initial.items() if m[a]!=v and not E<=a<E+0x2000]
+    return m,{'return':e['return'],'requests':requests,'changes':changes,'execution':{k:e[k] for k in ['stop','trace_length','trace_sha256']}}
+
+def predicate(ins,m,args=None):
+    out,p=execute(ins,m,7662,7805,28)
+    return out,p['return'],p

scripts/test_qvm_exit_predicate.py

@@ -0,0 +1,16 @@
+import json,unittest
+from pathlib import Path
+
+def expected(a,c):
+    if a['playing']<2 or not a['enabled']:return 0
+    mode=a['mode']
+    if mode>=3 and mode not in [9,11]:return int(a['first_score']==a['second_score'])
+    key='308' if mode==9 else '696' if mode==11 else '248'
+    return int(c[a['order'][0]][key]==c[a['order'][1]][key])
+class PredicateTests(unittest.TestCase):
+    def test_source_result_against_independent_policy(self):
+        r=json.loads(Path('docs/research/qvm-exit-predicate.json').read_text());self.assertEqual(len(r['probes']),168)
+        for p in r['probes']:self.assertEqual(p['return'],expected(p['input'],p['clients']))
+    def test_no_external_writes_or_captured_effects(self):
+        for p in json.loads(Path('docs/research/qvm-exit-predicate.json').read_text())['probes']:self.assertEqual(p['changes'],[]);self.assertEqual(p['requests'],[])
+if __name__=='__main__':unittest.main()

xtask/checks.tsv

@@ -2154,3 +2154,8 @@ edge	docs/research/qvm-exit-class.json	scripts/analyze_rust_exit_class.py
 edge	docs/research/qvm-exit-limits.json	scripts/analyze_rust_exit_limits.py
 edge	docs/research/qvm-exit-timers.json	scripts/analyze_rust_exit_timers.py
 edge	docs/research/qvm-exit-complete.json	scripts/analyze_rust_exit_complete.py
+report	scripts/analyze_qvm_exit_predicate.py	docs/research/qvm-exit-predicate.json
+edge	docs/research/qvm-exit-predicate.json	scripts/test_qvm_exit_predicate.py
+group	exit-predicate	scripts/analyze_qvm_exit_predicate.py
+edge	scripts/analyze_qvm_exit_predicate.py	scripts/analyze_qvm_resumer_audit.py
+edge	scripts/qvm_intermission_support.py	scripts/analyze_qvm_resumer_audit.py