quantooo.com is a free quant research lab for learning how quantitative investing really works. It scores US (S&P 100) and India (Nifty 50) stocks from 0 to 100 with every calculation step shown, tests machine learning models with walk-forward backtests and a sealed holdout, and keeps a live track record that is fingerprinted and timestamped on Bitcoin before the market opens.
- Discussions: comments on the site's Learn articles appear here (one thread per article), plus questions, ideas and critique.
- Track record: every day's scores are fingerprinted (SHA-256) and timestamped on the Bitcoin blockchain (OpenTimestamps) before the market opens, so they can't be changed afterwards. Download and verify them on the track record page.
- 12 models, from simple rules to LSTM and GRU networks, tested walk-forward: none ranked stocks reliably by next-month return (model leaderboard).
- The score's sealed final test year came out significantly negative, and we published it (research).
- Forecasting risk works better than forecasting direction (forecasting volatility).
- A one-year live challenge runs until October 2027: can rule-based portfolios beat their benchmarks? (model portfolios).
track_record/ holds every published score file, copied here nightly:
| Folder / file | What it is |
|---|---|
scores/ |
US (S&P 100) scores, one CSV per trading day, with its OpenTimestamps proof (.ots) |
scores-in/ |
India (Nifty 50) scores, same format |
preregistration/ |
Rules of the one-year challenge, registered before the first trade |
ledger.csv |
Date, file name, SHA-256 fingerprint and publication time of every file |
CORRECTIONS.md |
Every correction we have made, and why |
Check a file yourself: compute its SHA-256 (sha256sum track_record/scores/2026-10-20.csv on Linux or macOS,
certutil -hashfile <file> SHA256 on Windows) and compare it with the row in ledger.csv. To check the Bitcoin
timestamp, open the .ots file at opentimestamps.org together with the CSV.
Model-specific files are kept private; the ledger still lists their fingerprints.
Every formula is published on The math. The code that runs the site is private.
Plain-language articles, each backed by tests published on the site:
| Topic | Article |
|---|---|
| Getting started in quant | What is quant research? A beginner's guide · Quant glossary: 30 terms · Free quant tools and data |
| Quant backtesting | Quant backtesting checklist: 12 checks |
| Quant investing in India | Testing a stock score on the Nifty 50 |
| Machine learning for stock prediction | Can machine learning predict stock returns? We tested 12 models |
| Backtesting pitfalls | Look-ahead bias · Overfitting · Survivorship bias · Trading costs |
| Backtest methodology | Walk-forward testing · Sealed holdout |
| Factor investing | Does momentum investing still work? |
| Evaluating signals | Information coefficient (IC) |
| Volatility forecasting | HAR, GARCH, machine learning and the VIX compared |
New articles every week, including a weekly summary of the biggest score moves.
- Can machine learning predict stock returns? We tested 12 models
- What is look-ahead bias in backtesting?
- Overfitting in backtesting, explained with a real example
- Beginner's guide to the site
Free and educational. Not investment advice, not a recommendation to buy or sell anything. Results are hypothetical, not a claim of returns. Contact: contact@quantooo.com