Origin Story
Why we built this
In 2023, Vaasudev was paying ₹4,500/month to a popular Telegram stock advisory channel. After 6 months, he tallied the actual results: 52% accuracy on their calls — barely better than a coin flip, at an annual cost of ₹54,000.
The frustrating part wasn't the accuracy. It was that they never published their accuracy. They only shared the wins on social media. The losses were quietly forgotten.
SIGNAL was built to answer one question: What if a trading signal service published every single call — wins and losses — and let the data speak for itself?
2023 — The frustration
₹54,000 spent on Telegram calls
6 months · 52% accuracy · zero published track record. Started building own ML model on weekends.
Jan 2024 — First model
Random Forest v1 — 61% accuracy
Trained on 3Y NSE data · 8 indicators · First signal posted publicly on Twitter. Small following started.
Jul 2024 — v2 model
Added Delivery % · FII/DII flow
Accuracy improved to 68%. Twitter community growing. First DMs asking "how do I get your signals?"
Mar 2025 — v3 model
71.4% accuracy · 15 parameters
Added Twitter sentiment, sector momentum. Backtested 5 years. Decided to build a proper product.
Jun 2026 — SIGNAL launches
Website + App · Free to start
Full platform live — signals, algo builder, paper trading, broker sync. ₹199/month. Fully public track record.