Mistbind Clique targets the wrong player from the wrong timing window #17

New Issue

bstein · 2026-07-06T17:51:10Z

bstein commented

2026-07-06 17:51:10 +00:00

What happened

Mistbind Clique was cast in player_1's own draw step and targeted player_3 while the reason claimed it would deny the active Gitrog player's mana this turn. Player_3 was not active and would untap before using those lands.

What should have happened

Mistbind Clique should be timed and targeted only when championing it produces a real mana-denial or combat/stack payoff against the relevant active or next-threat player. The prompt/evaluator should make active-player and next-untap context explicit enough to reject tapout lines that do not constrain the intended player.

Context

correction target: ["pilot_policy", "card_notes", "atomic_evaluator", "prompt_state_active_player_emphasis"]
evidence: [{"artifact path": "artifacts/deployed_batteries/edge_case_0521_live_20260706T134944Z/logs/dashboard/edge-case-smoke-deployed-proof-commander-20260706t102658z-edge-commander-pod-timing-stack-g02-level-2/commander_pod_g1_2026070502/game.replay_events.ndjson", "frames": "228-229-235-238-239-240", "phase": "DRAW", "priority player": "player_1", "summary": "Mistbind Clique was cast in player_1's own draw step and targeted player_3 while the reason claimed it would deny the active Gitrog player's mana this turn. Player_3 was not active and would untap before using those lands.", "turn": 16}]
feedback issue key: explicit:mistbind_clique_wrong_player_timing_mana_denial
recommended source files: ["src/veles/playtest_decisions.py", "src/veles/codex_decision_worker/legal_action_context.py", "src/veles/atomic_decision_eval.py", "../veles-content/strategies/commander/edge_commander_alela_draw_go/1/guide.md"]
root cause: mistbind_clique_wrong_player_timing_mana_denial
source findings file: reports/edge_case_atomic_eval/0.5.21_live/commander_timing_stack_g02_unique_findings_20260706.json

## What happened Mistbind Clique was cast in player_1's own draw step and targeted player_3 while the reason claimed it would deny the active Gitrog player's mana this turn. Player_3 was not active and would untap before using those lands. ## What should have happened Mistbind Clique should be timed and targeted only when championing it produces a real mana-denial or combat/stack payoff against the relevant active or next-threat player. The prompt/evaluator should make active-player and next-untap context explicit enough to reject tapout lines that do not constrain the intended player. ## Category - kind: misplay - severity: moderate ## Context - correction target: ["pilot_policy", "card_notes", "atomic_evaluator", "prompt_state_active_player_emphasis"] - evidence: [{"artifact path": "artifacts/deployed_batteries/edge_case_0521_live_20260706T134944Z/logs/dashboard/edge-case-smoke-deployed-proof-commander-20260706t102658z-edge-commander-pod-timing-stack-g02-level-2/commander_pod_g1_2026070502/game.replay_events.ndjson", "frames": "228-229-235-238-239-240", "phase": "DRAW", "priority player": "player_1", "summary": "Mistbind Clique was cast in player_1's own draw step and targeted player_3 while the reason claimed it would deny the active Gitrog player's mana this turn. Player_3 was not active and would untap before using those lands.", "turn": 16}] - feedback issue key: explicit:mistbind_clique_wrong_player_timing_mana_denial - recommended source files: ["src/veles/playtest_decisions.py", "src/veles/codex_decision_worker/legal_action_context.py", "src/veles/atomic_decision_eval.py", "../veles-content/strategies/commander/edge_commander_alela_draw_go/1/guide.md"] - root cause: mistbind_clique_wrong_player_timing_mana_denial - source findings file: reports/edge_case_atomic_eval/0.5.21_live/commander_timing_stack_g02_unique_findings_20260706.json

bstein added the

misplay

battery-0.5.21

manual-review

commander-timing-stack

labels 2026-07-06 17:51:11 +00:00

bstein commented

2026-07-09 06:33:10 +00:00

Verified in Cassandra 0.5.30.

Release proof: source bebd4c7 (Prepare Cassandra 0.5.30 feedback backlog closeout), Jenkins Veles #209 SUCCESS published version-only backend/frontend/sim-worker 0.5.30 images, Titan-IAC 1bf84977 deployed them, live /api/v1/app-config reports backend/frontend/sim_worker 0.5.30, /api/v1/ready is green, backend pod and both frontend pods are Running with 0 restarts, and VELES_SIM_IMAGE is registry.bstein.dev/veles/veles-sim-worker:0.5.30.

Local/source gate on the deployed source: bridge/runtime 42 tests OK; evaluator/audit/policy/API suite 504 tests OK; frontend maybeboard/compact-replay/prompt Playwright 22 tests passed; frontend build and git diff --check passed.

Issue-specific proof: policy/audit coverage now checks Mistbind Clique timing, affected player, and champion selection so it is used in the correct window with a sensible target. Closing.

Verified in Cassandra 0.5.30. Release proof: source bebd4c7 (Prepare Cassandra 0.5.30 feedback backlog closeout), Jenkins Veles #209 SUCCESS published version-only backend/frontend/sim-worker 0.5.30 images, Titan-IAC 1bf84977 deployed them, live /api/v1/app-config reports backend/frontend/sim_worker 0.5.30, /api/v1/ready is green, backend pod and both frontend pods are Running with 0 restarts, and VELES_SIM_IMAGE is registry.bstein.dev/veles/veles-sim-worker:0.5.30. Local/source gate on the deployed source: bridge/runtime 42 tests OK; evaluator/audit/policy/API suite 504 tests OK; frontend maybeboard/compact-replay/prompt Playwright 22 tests passed; frontend build and git diff --check passed. Issue-specific proof: policy/audit coverage now checks Mistbind Clique timing, affected player, and champion selection so it is used in the correct window with a sensible target. Closing.

bstein closed this issue

2026-07-09 06:33:11 +00:00

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: cassandra-alpha/feedback#17