Run-it: widened discovery cycle
21 of 64 tasks doneDecision batch, un-hold 4,490, DSR calibration, data completion, sector onboarding.
InboxINB-139AlphaFactory Build-3: grammar extension (donchian_bo + squeeze_bo) + parked-row remap + bank refill
BacklogINB-186Strategy Bench reopening: universe widening + base families + dual admission + cockpit (P0-P6)
Needs MohitINB-242v11 vbt-clone engine: parallel port of the parity v11 strategy, parity kept as referee
DoneINB-246Gapless whole-history intraday: finish AV seam drip + stitched build + venue flip decision
DoneINB-249Onboard thematic sectors to the symbol universe: nuclear/uranium, quantum, robotics, AI, biotech (+ full data load)
Needs MohitSBN-11Bench cockpit: parameter-distribution dashboard (10+ interactive plots with family/TF filters)
Needs MohitSBN-13Bench cockpit: Cycles & Seasonality tab (monthly/halloween/election/halving/regime analysis per sector)
BacklogSBN-15Bench housekeeping: BOOKS registry vtt-gold twin, TON mapping, export-failure alert, top-200 refresh, weekly venv
BacklogSBN-16Seasonality tab v-next: surrogate nulls, robust seasonality stats, OOS column, PS/LT regimes, vol-seasonality (per cyclicity brief)
DoneSBN-17Nightly status.json export times out (600s) since 2026-08-30 — Operations tab stale since 2026-08-27
BacklogSBN-18Cockpit candidate table: CAGR/DD/Sortino/Sharpe/trades/TF + B&H return/Sortino/DD columns, header sort + per-column filters, always-visible horizontal scrollbar
BacklogSBN-19Standing evidence backfill: re-run backtests for candidates missing B&H / window / trade-count / core metrics (metrics only, decisions untouched)
DoneSBN-2Codex Mode A review fallout: 6 filed defects in 88bc84c..9f3effc (variant exporter, IBKR clamp, stitched builder)
BacklogSBN-21Candidate generation rethink: Mohit lacks confidence in AlphaFactory's generator — design how new candidates should be created before any re-seed
BacklogSBN-22Gauntlet tab v2: per-gate sub-tabs (G-B→G-A→G-C→G-D), cumulative survivors highlighted, died-here + untestable collapsibles, chain-diagnostic verdicts for historical rows
Needs MohitSBN-23Gauntlet restructure to six gates: A return>0, B Sortino>B&H, C return>=50% of B&H return, D trade history, E PBO, F DSR — engine + lens + UI
Needs MohitSBN-24Verify first six-gate nightly export (2026-09-04) + drain #3 outcome + DSR=0 DB spot-check
DoneSBN-26Seed 4 FMZ gap concepts (VWAP trend, VWAP revert, Heikin-Ashi trend, Kalman trend) via manual intake
BacklogSBN-27Phase Oscillator (Saty PO) morning momentum family on hourly QQQ: candidate seed + attended measure
BacklogSBN-4Metric parity: everything Screener and the tournament compute must also be computed and shown in the bench
DoneSBN-41SBN bookmark batch 2026-09-13: 7 X-bookmark strategies into the bench (3 asterisk intraday concepts on >=1h)
Needs MohitSBN-44Telegram strategy-idea intake: Strategy Ideas topic -> claude -p clarity check -> vault note -> INB-222 intake -> live VPS bank candidates
DoneSBN-49Proper standalone backtest of vidya_atr_flat on BTC/ETH/SOL 4h (RS-8-style pre-registered)
DoneSBN-5Decide the bench's strategy ranking rule: Sharpe-first (tournament) vs blended composite (Screener CRS)
Needs MohitSBN-50Equal backtest coverage: every strategy x every asset group x every timeframe the data allows
DoneSBN-51Seed the "10 classic quant systems" gaps (Dual Thrust, R-Breaker, Fairy Four Price, Gap, Oliver Kell EMA, Dolphin, Escalator, Martingale) as bench candidates
BacklogSBN-56Idea intake: re-harvested idea marked exact_duplicate of itself, hiding its real outcome
Needs MohitSBN-6Pre-"run it" blockers: G-B for all candidates, exogenous vol data, de-risk overlay, generator cost realism
DoneSBN-60Instant idea test: share a trading idea, get a verdict vs VIDYA+ATR in minutes (design + plan)
DoneSBN-61Build the missing bench strategy types so parked Telegram ideas can be tested (8/21 EMA, CCI first)
Needs MohitSBN-64Idea test STRONG: volume-confirmation filter on VIDYA v10 (crypto) - confirm before any live use
Needs MohitSBN-66Instant idea test: finish-out (label guard, baseline refresh decision, remaining proofs, review + merge)
BacklogSBN-67Idea test: BigBeluga "Secret BTC Strategy" look-alike scored STRONG in 3 slices, but the wins are hollow
Needs MohitSBN-68Idea follow-up: is the GLD 1h result of the Triple HTF Bollinger test real or luck? (PAXG slow-band check)
Needs MohitSBN-69Bench: add 6 strategies from the ArunKBhaskar Pine Script repo (3 mechanical + 3 ICT asterisk)
DoneSBN-7Deploy the live nightly checkout onto master (62 commits behind; SPAC-shell fix not running)
Needs MohitSBN-71Idea test: EMA200 vs SMA200 cloud flip on crypto 4h beat v10 in 10-11 of 20 coins (Popeye chart) - decide next check
BacklogSBN-72Idea test: Jungle Rock Turtle (Hurst gate + Donchian/SuperTrend/MA ensemble) - gate adds nothing, ungated ensemble beats v10 but not buy-and-hold
BacklogSBN-73Idea test: QuantifiedStrat put-call ratio above 3 - no signal since 2016, not reproducible, not worth it
BacklogSBN-75Bench funding downloader hangs on newer coins; bench server has funding for only 14 coins
DoneSBN-76Idea test: bitcoinjack "big trend up, buy the dip or the quiet range, never chase" - patient entries do NOT beat chasing