个人 GitHub 账号下公开仓库的所有者可以登录并验证所有权,自行管理介绍,无需管理员批准。
登录并生成验证文件,在 30 分钟内将其提交到此仓库的默认分支。其他分支或分叉仓库中的文件无效。然后在下方检查文件,即可无需管理员批准直接管理项目。请保留文件以维持管理权限。
如需登录,将前往登录页面。登录后请回到此处,再次点击所选的验证按钮。
在项目文件夹中运行 /nomorevibe verify,即可验证所有权并自行管理介绍。
一分钟完成验证正在加载完整 README。 以下为已保存的摘录。
Clean, long-horizon-only rebuild of the NiftyQuant long-horizon strategy: sma200_slope_63 top-15 cross-sectional trend-momentum on a PIT-clean, large+mid, solvent Nifty-500 universe.查看完整 README ↗
Clean, long-horizon-only rebuild of the NiftyQuant long-horizon strategy:
sma200_slope_63 top-15 cross-sectional trend-momentum on a PIT-clean, large+mid,
solvent Nifty-500 universe.
nq/ package; transplant only the data corrections, empirical history, and golden
master that can't be regenerated from a spec.skills/ — pre-registration, n_trials/DSR, the
promotion bar, the baseline_v0 anchor. Binding.baseline_v1 — the pinned baseline of record (ADR-0006, dataset-pin-20260701).
sma200_slope_63 cross-sectional momentum, 63-day hold. Gross 15.46% CAGR /
Sharpe 0.667 / MaxDD −46% / ~12% after-tax. Byte-reproducible from the pinned
OHLCV. In-sample; no real capital has traded.weekly-swing 0094 — the weekly-swing book surfaced on the marketing site
("this week's scan"). NET of costs, corrected universe 2017–2026: 8.1× (₹10L → ₹81L,
+711%) / Sharpe 1.13 / CAGR 24.7% / MaxDD −42% / 59% win over 255 trades.
Reproduce: python scripts/run_bhanushali_weekly_rank.py. DSR 0.894 → below the
deflated-Sharpe gate → NOT certified. It is a registered proposal for the 2026-10-01
review + the forward wall — paper-tracked, not proven, before 20% STCG, no real capital.frontend/ — the marketing landing (src/pages/LandingV2.jsx) + the React
dashboard. The public landing serves aggregate, non-actionable stats only (no tickers,
no prices) via GET /api/landing-stats, with the "not certified · paper-tracked"
framing intact.dashboard/backend/ — FastAPI: signals, positions, trades, Kite (Zerodha) OAuth,
and the public landing_stats router (the frozen track-record / equity / weekly-scan
constants the landing renders).Numbers discipline: every figure on the landing is reproducible from the committed pipeline — never a chat transcript. Honesty framing (in-sample, deep drawdown, uncertified, no real capital) is load-bearing, not decoration.
nq/data OHLCV + PIT membership mask + features (sma200_slope_63, atr, ADV, D/E)
nq/engine panel + exits (shared) + portfolio.simulate (the backtest/live engine)
nq/strategy the long-horizon scan
nq/validation cpcv / bootstrap / dsr / metrics
nq/runner live scanner
research/ pre-registered findings (NNNN-*.md), overlay registry, exports
dashboard/ FastAPI backend (signals, Kite, landing-stats)
frontend/ React dashboard + marketing landing (LandingV2)