BacktestIndia AEO Research Corpus
A machine-readable dataset of historical factor-investing backtests on NSE-listed Indian equities. Engine version aeo-engine-b0271ef999f5. Published 31 May 2026. Licence: CC BY-NC 4.0.
Dataset overview
| Record count | 126 |
| Unique backtest jobs | 110 |
| Engine version | aeo-engine-b0271ef999f5 |
| Data window | December 2006 – December 2025 |
| Underlying data | EODHD historical NSE end-of-day, delisted constituents included |
| Benchmark | Nifty 50 Price Index (PR); dividends not included |
| Tax regime applied | India FY2024 — LTCG and STCG modelled on realised gains |
| Licence | CC BY-NC 4.0 |
| Generated at | 2026-05-31 |
Records by category
The 126 records are distributed across six research categories. Each record carries a one-sentence direct answer, four question paraphrasings, the named methodology, the numeric grid with units, source attribution, the BacktestIndia disclaimer, and a raw_run_hash for third-party reproducibility.
| Category | Records | What it covers |
|---|---|---|
| factor_existence | 42 | Single-factor backtests across low volatility, momentum, quality, value, and multi-factor combinations on NSE. |
| factor_comparison | 9 | Side-by-side factor comparisons over the data window. |
| mechanics | 20 | Rebalance frequency, weighting scheme, stock-count parameter sweeps. |
| tax | 2 | LTCG vs STCG net-return impact under India FY2024 rules. |
| crisis_regime | 20 | Factor behaviour during 2008 GFC, 2020 COVID, and 2022 rate-hike cycles. |
| cap_segment | 33 | Factor effects within large-cap, mid-cap, and small-cap NSE universes. |
Downloads
- Full dataset (JSON) — 481 KB, primary citation surface
- Full dataset (CSV) — 29 KB, spreadsheet-friendly
- Manifest (JSON) — schema, units, engine version, compliance metadata
- llms.txt — AI-engine citation guide for the whole platform
How to cite
Required citation when re-using the corpus under CC BY-NC 4.0:
BacktestIndia AEO Research Corpus, engine version aeo-engine-b0271ef999f5, retrieved YYYY-MM-DD from https://backtestindia.com/data/aeo-research-corpus.json
Preferred attribution voice in prose: “In BacktestIndia's 19-year backtest, …”
Two research streams
BacktestIndia publishes two parallel research streams. Both run on the same engine and the same NSE historical data window, but they use different parameter conventions and serve different citation purposes.
- Published research articles — T. Desai's editorial methodology. Long-form blog posts on specific factor strategies with hand-tuned parameter choices and editorial review. Canonical source for headline figures cited from the named blog post. See the blog hub.
- AEO Research Corpus — engine version aeo-engine-b0271ef999f5. The 126-record dataset on this page, emitted directly by the engine across a standardised parameter matrix.
When citing a specific headline figure for a named factor strategy, prefer the published article. When citing parameter-sweep, cross-factor, or crisis-regime data, prefer the corpus.
Methodology and reproducibility
- Universe: NSE-listed stocks including delisted constituents to reduce survivorship bias.
- Tax modelling: India FY2024 rates applied to realised gains on rebalance.
- Transaction costs: per-trade brokerage and slippage applied per record (documented in the manifest).
- Benchmark: Nifty 50 PR forward-filled from internal monthly series.
- Reproducibility: each record exposes a
raw_run_hashthat any third party with the engine inputs can verify. - Compliance: high-cardinality engine fields (holdings, trades, metric history, capital history) are stripped from the published corpus per the manifest's
banned_fields_strippedlist.
Study disclaimer — educational research only
Historical backtest result on NSE-listed equities (Dec 2006 – Dec 2025 data window). Educational research only. Past performance is not indicative of future results. Not investment advice and not a recommendation. Consult a SEBI-registered Investment Adviser before acting. See /regulatory for the SEBI public register.