Keep Records

243 trading strategies from two dozen books and courses. Tested 1,927 times, with real costs.

Only 2 survived. This book gives you both, with every rule, every cost and every bad year.

I didn't start as a tester. For ten years I was the trader the courses are sold to: a good salary, a phone full of apps, ₹5 lakh worth of courses, and later Python bots running on a server while I slept. One night, one of them lost ₹1.4 lakh before I woke up.

My father had fixed deposits, a squeaky bicycle and a diary where he wrote down every rupee. Guess which one of us is debt-free and sleeps well.

Keep Records is the true story, with real numbers from my own broker statements, of how I lost about ₹10 lakh. It's also about what I built when I finally stopped believing and started testing.

Inside you'll find:

• The 2 strategies that survived, in plain words, including their worst years
• Why all 414 intraday setups and all 171 machine-learning predictors I tested failed
• The backtest that showed 20,000% and became −19% with realistic slippage
• What leverage does to a strategy that works at 1x
• The checklist I now run before a single rupee goes in, and 20 safety rules for trading bots

Funny, honest and written in simple Indian English, for anyone with a salary, a trading app and a feeling that the next trade will fix everything.

Not investment advice. Backtests are not promises, and this book shows you why.