jgrusewski 079a06e485 test(sp17): asymmetric A → deterministic argmax under MIN_TEMP=0.5 ...

Verifies that when one direction's centered advantage dominates
strongly enough, action selection picks it deterministically across
all N=1000 Philox-keyed runs.

Plan specifies "Thompson temp = 0.0 → pure argmax E[Q]", but the
kernel floors thompson_temp at MIN_TEMP=0.5 per
pearl_blend_formulas_must_have_permanent_floor — pure τ=0 isn't
ISV-accessible. With τ=0.5 the blend is q_eff = 0.5·E[Q] +
0.5·q_sample; deterministic argmax across all τ=0.5 draws requires
the E[Q] gap to exceed atom_span/2.

Construction: NA=3 atoms [-0.1, 0, +0.1] (atom_span=0.2). A_dir
puts +300 at z=2 for Long, -100 at z=2 for others (per-atom mean
zero ⇒ centering preserves shape). Long centered E[Q] ≈ +0.1, others
≈ -0.05 (gap 0.15 > 0.1). q_sample[Long] = +0.1 with prob ≈ 1
(softmax(300) ≈ delta at z=2); q_sample[other] ∈ {-0.1, 0} (zero
prob for +0.1). q_eff[Long] = 0.1; q_eff[other] ≤ -0.025. Long
strictly wins all draws.

If the centering breaks (Long no longer dominant under centered E[Q])
or τ blends a non-Long sample over Long, this test fires.

Plan: docs/superpowers/plans/2026-05-08-sp17-dueling-q-network.md

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

2026-05-08 22:25:51 +02:00

..

behavioral

test(sp15-p2b): 17 behavioral tests (Group 1 + Group 2) as #[ignore] contracts

2026-05-06 15:05:34 +02:00

common

…

edge_cases

…

fixtures

…

integration

…

regression

…

unit

…

validation

…

ab_testing_integration.rs

…

action_loader_real_csv_test.rs

…

action_masking_smoke_test.rs

…

activation_tests.rs

…

adx_features_test.rs

…

aux_trunk_oracle_tests.rs

test(sp14-c9): synthetic smoke for aux trunk gradient chain + C.8/C.9 audit close-out

2026-05-08 03:30:44 +02:00

barrier_optimization_test.rs

…

bayesian_changepoint_test.rs

…

bessel_correction_test.rs

…

bug16_portfolio_features_test.rs

…

bug18_cash_reserve_solvency_test.rs

…

bug20_portfolio_normalization_test.rs

…

bug21_bug22_bug23_compilation_fixes_test.rs

…

bug24_bug25_compilation_fixes_test.rs

…

bug28_unused_import_test.rs

…

cash_accounting_fix_test.rs

…

cash_reserve_requirement_test.rs

…

checkpoint_test.rs

…

configurable_capital_test.rs

…

contract_multiplier_test.rs

…

cusum_test.proptest-regressions

…

debug_position_delta_test.rs

…

diffusion_integration.rs

…

dqn_accumulation_convergence_test.rs

perf(checkpoint): async best-ckpt serialize via spawn_blocking + mapped-pinned param snapshot

2026-04-28 18:54:00 +02:00

dqn_action_collapse_fix_test.rs

refactor(dqn): strip dead use_* flags + tighten count_bonus API

2026-04-27 17:52:23 +02:00

dqn_action_dependent_reward_test.rs

…

dqn_action_masking_integration_test.rs

…

dqn_action_position_sign_convention_test.rs

…

dqn_activity_penalty_fix_test.rs

…

dqn_adapter_paths_test.rs

…

dqn_advanced_metrics_integration_test.rs

…

dqn_barrier_integration_debug_test.rs

…

dqn_c51_shape_validation_test.rs

…

dqn_diagnostic_logging_test.rs

…

dqn_diversity_penalty_test.rs

…

dqn_dueling_batched_wave62_test.rs

…

dqn_early_stopping_termination_test.rs

refactor(dqn): f64 → f32 for kernel-facing hyperparams

2026-04-24 01:55:14 +02:00

dqn_episode_boundaries_test.rs

…

dqn_epsilon_decay_validation_test.rs

…

dqn_feature_cache_test.rs

…

dqn_feature_defaults_test.rs

…

dqn_gradient_accumulation_test.rs

perf(checkpoint): async best-ckpt serialize via spawn_blocking + mapped-pinned param snapshot

2026-04-28 18:54:00 +02:00

dqn_gradient_collapse_root_cause_test.rs

…

dqn_gradient_dtype_simple_test.rs

…

dqn_hft_barriers_test.rs

…

dqn_hyperopt_json_export_test.rs

…

dqn_hyperopt_test.rs

…

dqn_hyperparams_kelly_fields_test.rs

refactor(dqn): f64 → f32 for kernel-facing hyperparams

2026-04-24 01:55:14 +02:00

dqn_inference_test.rs

perf(checkpoint): async best-ckpt serialize via spawn_blocking + mapped-pinned param snapshot

2026-04-28 18:54:00 +02:00

dqn_kelly_regime_integration_test.rs

…

dqn_logit_clipping_bug12_test.rs

…

dqn_long_training_test.rs

perf(checkpoint): async best-ckpt serialize via spawn_blocking + mapped-pinned param snapshot

2026-04-28 18:54:00 +02:00

dqn_pnl_calculation_tests.rs

…

dqn_rainbow_config_test.rs

…

dqn_rainbow_test.rs

…

dqn_regime_conditional_integration_test.rs

…

dqn_reward_calculation_test.rs

…

dqn_tensor_shape_validation.rs

…

dqn_trainer_integration_tests.rs

fix: update all test/example files for VarStore removal + 7-level exposure

2026-04-11 15:21:21 +02:00

dqn_trainer_p1_tests.rs

…

dqn_training_pipeline_test.rs

perf(checkpoint): async best-ckpt serialize via spawn_blocking + mapped-pinned param snapshot

2026-04-28 18:54:00 +02:00

dqn_transaction_cost_bug2_test.rs

…

dqn_transaction_costs_test.rs

…

dqn_use_double_dqn_test.rs

…

dropout_scheduler_tests.rs

…

ensemble_4_models_integration.rs

…

ensemble_hot_swap_test.rs

…

ensemble_hyperopt_tests.rs

…

ensemble_inference_integration_test.rs

refactor: remove configurable state_dim — use STATE_DIM constant everywhere

2026-04-20 15:29:05 +02:00

evaluation_net_pnl_bug3_test.rs

…

feature_extraction_46_test.rs

…

feature_normalization_test.rs

…

fxcache_roundtrip_test.rs

feat(fxcache): OFI_DIM 20→32, persist 12 real-math microstructure signals,

2026-04-23 11:36:34 +02:00

gpu_backtest_validation.rs

feat(sp15-wave3b): host-side wire-up — eliminates 5 orphan launchers via GpuBacktestEvaluator constructor signature change

2026-05-06 23:39:45 +02:00

gpu_per_integration_test.rs

fix(sp13): B0 cascade gap — 5 unaudited insert_batch call sites

2026-05-05 10:46:20 +02:00

gradient_accumulation_tests.rs

…

huber_loss_test.rs

…

hyperopt_action_masking_test.rs

…

hyperopt_early_stopping_infrastructure_test.rs

…

hyperopt_kelly_params_test.rs

…

hyperopt_paths_test.rs

…

hyperopt_price_extraction_test.rs

…

hyperopt_tft_early_stopping_test.rs

…

imbalance_bars_test.rs

…

inference_engine_test.rs

…

integration_ppo_ensemble.rs

…

kan_integration.rs

…

kelly_criterion_integration_test.rs

…

liquid_cfc_training_test.rs

…

liquid_networks_test.rs

…

liquid_nn_training_tests.rs

…

log_size_test.rs

…

lr_scheduler_tests.rs

…

mamba2_accuracy_fix_test.rs

…

mamba2_early_stopping_test.rs

…

mamba2_gradient_extraction_test.rs

…

mamba2_hyperopt_edge_cases.rs

…

mamba2_hyperopt_p0_p1_fixes.rs

…

mamba2_weight_update_test.rs

…

mbp10_parsing_test.rs

…

microstructure_features_test.rs

…

model_registry_checkpoint_test.rs

…

model_validation_comprehensive.rs

…

moe_kernels_test.rs

feat(moe): adaptive load-balance λ via gate-entropy ISV controller

2026-04-28 09:10:31 +02:00

multi_cusum_test.rs

…

ofi_features_test.rs

fix(ofi): align OFI window to bar t's formation interval — kill 31/32-slot lookahead

2026-04-28 09:23:32 +02:00

ood_input_handling_tests.rs

…

pages_test_test.rs

…

paper_trading_integration_test.rs

refactor: remove configurable state_dim — use STATE_DIM constant everywhere

2026-04-20 15:29:05 +02:00

partial_reversal_support_test.rs

…

partial_reversal_validation.rs

…

pnl_realism_tests.rs

…

portfolio_integration_tests.rs

…

portfolio_tracker_reset_test.rs

…

portfolio_value_normalization_test.rs

…

ppo_45_action_network_tests.rs

…

ppo_action_masking_tests.rs

…

ppo_checkpoint_roundtrip_test.rs

…

ppo_circuit_breaker_tests.rs

…

ppo_continuous_bounds_test.rs

…

ppo_continuous_transaction_costs_tests.rs

…

ppo_dual_phase_backtest_tests.rs

…

ppo_entropy_regularization_tests.rs

…

ppo_factored_action_tests.rs

fix: update all test/example files for VarStore removal + 7-level exposure

2026-04-11 15:21:21 +02:00

ppo_hidden_state_tests.rs

…

ppo_huber_loss_validation.rs

…

ppo_hyperopt_validation_test.rs

…

ppo_long_training_test.rs

…

ppo_lstm_architecture_tests.rs

…

ppo_lstm_training_loop_tests.rs

cleanup: declarative rewrites for ml-tests and trading-engine TODOs

2026-04-23 08:44:05 +02:00

ppo_portfolio_tracker_tests.rs

…

ppo_position_limits_tests.rs

…

ppo_recurrent_integration_tests.rs

cleanup: declarative rewrites for ml-tests and trading-engine TODOs

2026-04-23 08:44:05 +02:00

ppo_recurrent_performance_tests.rs

…

ppo_reward_normalizer_tests.rs

…

ppo_sequence_batching_multi_episode_tests.rs

…

ppo_step_counter_fix_test.rs

…

ppo_stress_testing_tests.rs

…

ppo_transaction_costs_tests.rs

…

ppo_validation_real_data_test.rs

…

preprocessing_bessel_integration.rs

…

preprocessing_test.rs

…

preprocessing_validation_tests.rs

…

price_validity_tests.rs

…

production_trainer_adaptive_features_integration_test.rs

…

production_training_smoke_test.rs

perf(checkpoint): async best-ckpt serialize via spawn_blocking + mapped-pinned param snapshot

2026-04-28 18:54:00 +02:00

pso_budget_calculation_test.rs

…

q_value_constraint_test.rs

…

rainbow_loss_shape_test.rs

…

real_data_helpers.rs

…

real_data_pipeline_test.rs

…

recovery_tests.rs

…

regime_transition_features_test.rs

…

risk_action_masking_test.rs

…

risk_position_limit_integration_test.rs

…

run_bars_test.rs

…

safety_comprehensive_test.rs

…

sample_weights_test.rs

…

scatter_add_gradient_test.rs

…

smoke_test_real_data.rs

test(integration): migrate CudaSlice→MappedF32Buffer in tests blocked since Bug-1 close-out

2026-05-03 09:52:33 +02:00

softmax_exploration_test.rs

…

sp4_producer_unit_tests.rs

feat(sp13): B1.0 — ISV[117] retirement + scale-free MSE bridge

2026-05-05 10:52:13 +02:00

sp5_producer_unit_tests.rs

sp7(controller): activation flag + Welford-α to escape bootstrap trap

2026-05-03 11:34:21 +02:00

sp11_producer_unit_tests.rs

fix(sp11): B1b smoke-recovery — z-score normalization for mag-ratio canary

2026-05-04 11:51:30 +02:00

sp12_reward_math_tests.rs

test(sp12): GPU oracle tests for the 3 reward math changes

2026-05-04 19:32:18 +02:00

sp13_layer_b_oracle_tests.rs

feat(sp13): B1.1b — producer kernel + replay direct path + experience collector

2026-05-05 14:45:47 +02:00

sp13_phase0_oracle_tests.rs

feat(sp13): B1.1a — K=1→2 + softmax CE kernel rewrites + struct flips

2026-05-05 14:09:48 +02:00

sp14_oracle_tests.rs

fix(sp16): floor Wiener-α at 0.4 for non-stationary control loops

2026-05-08 20:14:18 +02:00

sp15_phase1_oracle_tests.rs

fix(sp15-cubin-preload): pre-load 6 SP15 launchers' CudaFunction handles in experience collector

2026-05-07 23:36:07 +02:00

sp17_dueling_oracle_tests.rs

test(sp17): asymmetric A → deterministic argmax under MIN_TEMP=0.5

2026-05-08 22:25:51 +02:00

streaming_pipeline_edge_cases.rs

…

supervised_gpu_smoke_test.rs

…

target_update_tests.rs

…

test_dbn_sequence_256_features.rs

…

test_dqn_cuda_device.rs

…

test_extract_256_dim_features.rs

…

test_grn_weight_initialization.rs

…

test_scatter_source_gradients.rs

…

test_sell_bug.rs

…

test_streaming_loader.rs

…

test_var_gradient_flow.rs

…

tft_adapter_paths_test.rs

…

tft_causal_masking_validation.rs

…

tft_hyperopt_real_metrics_test.rs

…

tft_inference_latency_benchmark.rs

…

tft_lru_cache_test.rs

…

tft_quantile_loss_validation.rs

…

tft_real_dbn_data_test.rs

…

tft_test.rs

…

training_chaos_tests.rs

…

training_edge_cases.rs

fix: update all test/example files for VarStore removal + 7-level exposure

2026-04-11 15:21:21 +02:00

transaction_cost_application_test.rs

…

transition_matrix_test.rs

…

triple_barrier_test.rs

…

volatility_epsilon_test.rs

…

wave16_checkpoint_regression_test.rs

…

wave16q_ohlcv_mutation_test.rs

…

wave16r_position_limits_test.rs

…

wave16s_price_validation_test.rs

…

wave_d_e2e_nq_fut_225_features_enhanced_test.rs

…

wave_d_e2e_nq_fut_225_features_test.rs

…

wave_d_edge_cases_test.rs

…

xlstm_integration.rs

…