Draw-Go taps out for Teferi after identifying the need to hold interaction #11

New Issue

bstein · 2026-07-06T17:51:05Z

bstein commented

2026-07-06 17:51:05 +00:00

What happened

In deployed Draw-Go vs Storm g04, Draw-Go first identified that passing with Counterspell and March of Otherworldly Light covered the danger from two reducers, then tapped out for Teferi, Time Raveler and declined to bounce a reducer.

What should have happened

The evaluator should keep the pilot consistent with the recognized combo-risk plan unless Teferi's static or bounce mode materially improves the same turn cycle. Against two reducers, tapping out should require strong justification.

Context

correction target: ["pilot_policy", "atomic_evaluator", "card_note"]
evidence: [{"frames": "79/84-87", "game id": "edge_1v1_storm__1_g1_2026070604", "level": 4, "local log": "artifacts/deployed_batteries/edge_case_0521_live_20260706T124234Z/logs/dashboard/edge-case-smoke-deployed-proof-20260706t101310z-edge-1v1-draw-go-vs-storm-g04-level-4/edge_1v1_storm__1_g1_2026070604/game.replay_events.ndjson", "run group": "edge-case-smoke-deployed-proof-20260706t101310z-edge-1v1-draw-go-vs-storm-g04-level-4", "summary": "Draw-Go explicitly chose to hold Counterspell plus March, then reversed into a tapped-out Teferi against two reducers and declined to bounce a reducer."}]
feedback issue key: explicit:teferi_tapout_plan_reversed_against_live_combo_reducers
recommended source files: ["src/veles/strategy_policy_runtime/policy_ranking.py", "src/veles/playtest_decisions/decision_caches.py", "src/veles/codex_decision_worker/legal_action_context.py", "../veles-content/strategies/modern/edge_1v1_draw_go_control/1/guide.md"]
root cause: teferi_tapout_plan_reversed_against_live_combo_reducers
source findings file: reports/edge_case_atomic_eval/0.5.21_live/draw_go_storm_unique_feedback_findings_20260706.json

## What happened In deployed Draw-Go vs Storm g04, Draw-Go first identified that passing with Counterspell and March of Otherworldly Light covered the danger from two reducers, then tapped out for Teferi, Time Raveler and declined to bounce a reducer. ## What should have happened The evaluator should keep the pilot consistent with the recognized combo-risk plan unless Teferi's static or bounce mode materially improves the same turn cycle. Against two reducers, tapping out should require strong justification. ## Category - kind: misplay - severity: debatable_to_obvious ## Context - correction target: ["pilot_policy", "atomic_evaluator", "card_note"] - evidence: [{"frames": "79/84-87", "game id": "edge_1v1_storm__1_g1_2026070604", "level": 4, "local log": "artifacts/deployed_batteries/edge_case_0521_live_20260706T124234Z/logs/dashboard/edge-case-smoke-deployed-proof-20260706t101310z-edge-1v1-draw-go-vs-storm-g04-level-4/edge_1v1_storm__1_g1_2026070604/game.replay_events.ndjson", "run group": "edge-case-smoke-deployed-proof-20260706t101310z-edge-1v1-draw-go-vs-storm-g04-level-4", "summary": "Draw-Go explicitly chose to hold Counterspell plus March, then reversed into a tapped-out Teferi against two reducers and declined to bounce a reducer."}] - feedback issue key: explicit:teferi_tapout_plan_reversed_against_live_combo_reducers - recommended source files: ["src/veles/strategy_policy_runtime/policy_ranking.py", "src/veles/playtest_decisions/decision_caches.py", "src/veles/codex_decision_worker/legal_action_context.py", "../veles-content/strategies/modern/edge_1v1_draw_go_control/1/guide.md"] - root cause: teferi_tapout_plan_reversed_against_live_combo_reducers - source findings file: reports/edge_case_atomic_eval/0.5.21_live/draw_go_storm_unique_feedback_findings_20260706.json

bstein added the

labels 2026-07-06 17:51:05 +00:00

bstein commented

2026-07-09 06:32:59 +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: Draw-Go tap-out reversal coverage now preserves the identified interaction plan around reducer/combo states unless the tap-out has a valid tactical justification. 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: Draw-Go tap-out reversal coverage now preserves the identified interaction plan around reducer/combo states unless the tap-out has a valid tactical justification. Closing.

bstein closed this issue

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