All Articles

My best trading strategy made ₹8.9 lakh on paper. A few honest checks turned it into a loss. Here are the checks, and the data problems I had to fix before any result meant anything.

I tested day trading, swing trading, futures and options on five years of real NSE data, with every charge included. The strategies that looked great on old data stopped working on new data. A simple monthly SIP did as well as anything.

I tested Bodhan's indic-ocr and Sarvam Vision against Claude Opus 5.5 on 27 pages of Bhatkhande notation. Both read Gurmukhi well and still lost by 35 points, because the music lives in marks that aren't letters.

I had been saying "roughly 90% accurate" for months without measuring it. The eval showed that effort and model choice both move the result more than I expected, and that one config scored 59.4%, 59.4% and 90.6% on the same page.

You want one package public and the rest of the monorepo private, without copy-paste drift and without giving up the monorepo. git subtree split does it, the monorepo stays the source of truth, and the whole thing fits in a script with four guards.

Bhatkhande notation has always needed a custom font on the web, which means the text is not really text. Swarlipi takes an ASCII grammar and renders real Unicode letters in five scripts, with every mark drawn in CSS and inline SVG.