Multi-market swing screener

A screening engine that scans an equity universe and only lets a setup through once its stop, targets and risk to reward are defined.

Measures: Risk against reward, before entry
Reading: 500 → 59, scanned to candidates
Stack: Python, FastAPI, AWS, Quantitative modelling
Links: Open the live app Source

Swing screener results for the Nifty 500: 500 scanned, 59 candidates, each row with score, entry, stop loss, target and risk to reward. — One Nifty 500 scan: 500 in, 427 filtered out, 59 candidates with entry, stop and target.1200 × 750 px

Context

Most screeners stop at "this looks interesting". The hard part of a swing trade is the exit: where it is wrong, where it is right, and whether the distance between them is worth it.

The idea

Make risk part of the screen, not an afterthought. Candidates are scored by a multi-factor model, then sized with ATR-anchored stops and tiered targets, and anything below a minimum risk to reward never reaches the list.

Details

Volatility-adjusted pricing and momentum indicators feed a probabilistic scoring matrix.
It looks for two kinds of setup: mean reversion and trend continuation.
Stops come from ATR, not fixed percentages, so a volatile stock gets a wider stop and a smaller position.
A journal tab keeps the trades taken next to the scans that produced them.

What shipped

Runs as a FastAPI service on AWS (ap-south-1). A full Nifty 500 scan takes about five minutes. Try it.