The track record. It started at zero, in public.

{"source_table":"opentable_diners (market.db, read-only, scope=country/US/granularity=daily)","signal":"diff = yoy_pct(t) - yoy_pct(t-1)","rule":"diff > 0 -> long; else no trade","max_signal_age_days":7,"hold_trading_days":1,"one_position_at_a_time":true}

{"source_table":"nrc_reactor_status_report (live.db, read-only)","event":"unit >=1,000 MWe drops from >=95% to 0% in a non-refuel month (not Mar-May / Sep-Nov)","capacity_map":"/root/proof_nrc/build_proof.py","min_mwe":1000,"min_prev_pct":95,"hold_trading_days":5,"side":"short","one_position_at_a_time":true,"event_lookback_days":7}

{"source_table":"usda_boxed_beef_cutout (live.db, read-only)","signal":"mom5 = choice_cutout(t)/choice_cutout(t-5 obs) - 1","ic_sign":-1,"ic_provenance":"full-sample spearman(mom5, fwd 5d LE=F log-ret) = -0.024, n=5707 pairs, computed 2026-07-14 from the fleet's own usda_boxed_beef_cutout + livestock_futures tables; SIGNAL-002 grading: sign-consistent train+test k3-5, decile monotonic at k5","rule":"mom5 > 0 -> short LE=F; mom5 < 0 -> long LE=F; mom5 == 0 -> flat","max_signal_age_days":5,"hold_trading_days":5,"one_position_at_a_time":true}

{"signal":"fed_cp_cd_rates.rate_pct:mean","transform":"z20","asset":"SPY","k":5,"ic_is_sign":-1,"rule":"latest fresh forward_track prediction for the tuple; side = sign(signal_value * ic_is); hold k trading days","max_pred_age_days":10,"one_position_at_a_time":true}

Retired — pairing failed independent verification. retired 2026-07-15: pairing failed independent verification - phase-fragility; published t=-2.18 reproduces exactly as one sampling phase (phase0) but the other phase is null (-0.013, t=-0.09); full weekly sample p=0.145. See LEADLAG-PROOF-V2.md section 3. History kept, no new orders placed.

{"source_table":"box_office_daily (market.db, read-only)","window_days":7,"signal":"chg = sum(top10_gross, last 7d)/sum(top10_gross, prior 7d) - 1","rule":"chg > 0 (blockbuster surge) -> short (contrarian); else no trade","max_signal_age_days":5,"hold_trading_days":10,"one_position_at_a_time":true}

Retired — pairing failed independent verification. retired 2026-07-15: pairing failed independent verification - phase-fragile (t ranges +0.26..+1.43 across 10 offsets), zero in-sample effect (all signal in last 30% of window), and the PAYX companion leg has zero close rows on the box (untestable). See LEADLAG-PROOF-V2.md section 3. History kept, no new orders placed.

{"source_table":"indeed_postings_index (market.db, read-only, scope=national, series='total postings')","series":"total postings","lookback_days":10,"signal":"chg = index_sa(t)/index_sa(t-10 obs) - 1","rule":"chg > 0 -> long; else no trade","max_signal_age_days":25,"hold_trading_days":10,"one_position_at_a_time":true}

Retired — pairing failed independent verification. retired 2026-07-15: pairing failed independent verification - sub-significant (best |t|=1.88 < 2 threshold across non-overlapping phases; one offset ~zero). Real but modest tilt, not proof-grade. See LEADLAG-PROOF-V2.md section 3. History kept, no new orders placed.

{"source_table":"lumber_price_ppi (market.db, read-only)","series_id":"WPU081106013","signal":"mom3 = value(t)/value(t-3 monthly obs) - 1","rule":"mom3 > 0 (lumber cost inflating) -> short; else no trade","max_signal_age_days":50,"hold_trading_days":63,"side_if_positive":"short","one_position_at_a_time":true}

Retired — pairing failed independent verification. retired 2026-07-15: pairing failed independent verification - seasonality artifact (SA series WPS081 gives IC +0.04 on identical window; isolated seasonal factor alone is IC -0.66, t=-3.81, stronger than the published signal). Also n was inflated ~2x by PPI series pseudo-replication (honest n=21 vs published 40). See LEADLAG-PROOF-V2.md section 3. History kept, no new orders placed.

{"source_table":"lumber_price_ppi (market.db, read-only)","series_id":"WPU081106013","signal":"mom3 = value(t)/value(t-3 monthly obs) - 1","rule":"mom3 > 0 (lumber cost inflating) -> short; else no trade","max_signal_age_days":50,"hold_trading_days":63,"side_if_positive":"short","one_position_at_a_time":true}

Retired — pairing failed independent verification. retired 2026-07-15: pairing failed independent verification - data bug (NHTSA campno duplication, ~56x potaff overcount, confirmed and fixed); after dedup all change-based ICs |t|<1 and IS/OOS signs disagree at every horizon. See LEADLAG-PROOF-V2.md section 3. History kept, no new orders placed.

{"source_table":"nhtsa_recall_filings (market.db, read-only)","make":"FORD","lookback_months":2,"signal":"chg = potaff_sum(month t)/potaff_sum(month t-2) - 1 (current in-progress month excluded, partial-month bias)","rule":"chg > 0 (rising recall volume) -> short; else no trade","max_signal_age_days":75,"hold_trading_days":42,"one_position_at_a_time":true}

Retired — pairing failed independent verification. retired 2026-07-15: pairing failed independent verification - sign-flip on the full window (w1 -0.006 / w2 +0.060 / w3 +0.091); sign-consistency claim does not hold at any horizon on the full sample. See LEADLAG-PROOF-V2.md section 3. History kept, no new orders placed.

{"source_table":"nrc_reactor_status_report (live.db, read-only)","capacity_map":"/root/proof_nrc/build_proof.py","signal":"delta = fleet_online_mwe(t) - fleet_online_mwe(t-14)","lookback_days":14,"rule":"delta > 0 (fleet power-up) -> short; else no trade","max_signal_age_days":3,"hold_trading_days":10,"one_position_at_a_time":true}

Retired — pairing failed independent verification. retired 2026-07-15: pairing failed independent verification - short-window artifact; published stat used only a 1-year live.db window, full 105-week window gives t=-1.49 (fails). See LEADLAG-PROOF-V2.md section 3. History kept, no new orders placed.

{"source_table":"nrc_reactor_status_report (live.db, read-only)","capacity_map":"/root/proof_nrc/build_proof.py","signal":"delta = fleet_online_mwe(t) - fleet_online_mwe(t-7)","lookback_days":7,"rule":"delta > 0 (fleet power-up) -> short; else no trade","max_signal_age_days":3,"hold_trading_days":5,"one_position_at_a_time":true}

Retired — pairing failed independent verification. retired 2026-07-15: pairing failed independent verification - published n=130 impossible (max ~98 non-overlapping obs); replicated CMG t=0.72; basket IC ranges -0.112..+0.088 across weekly anchor phases, a noise signature. See LEADLAG-PROOF-V2.md section 3. History kept, no new orders placed.

{"source_table":"opentable_diners (market.db, read-only, scope=country/US/granularity=daily)","signal":"diff = yoy_pct(t) - yoy_pct(t-7)","rule":"diff > 0 -> long; else no trade","max_signal_age_days":7,"hold_trading_days":5,"one_position_at_a_time":true}

Retired — pairing failed independent verification. retired 2026-07-15: pairing failed independent verification - sign-flip (IS -0.04 vs OOS +0.64) and t=1.07 under deseasonalized non-overlapping methodology; effect concentrated in last ~3 months only. See LEADLAG-PROOF-V2.md section 3. History kept, no new orders placed.

{"source_table":"tsa_passenger_volumes (market.db, read-only)","signal":"wow = travelers(t)/travelers(t-7) - 1","rule":"wow > 0 (positive demand surprise) -> long; else no trade","max_signal_age_days":5,"hold_trading_days":3,"one_position_at_a_time":true}