IHSG AlgoTrader

The problem

Overnight and swing trading on the Indonesian Stock Exchange means acting on the same rules every day, at the open and the close. Done by hand, entries are late, stops are skipped and there is no honest record of whether a strategy works.

The approach

IHSG Pulse turns the rules into code. A pure-Python engine computes the signals (momentum, reversal, volume, structure and candle patterns) and runs two strategies, overnight swing and mean reversion. Django apps wrap it with the market data, a signal log, a portfolio and a dashboard, and APScheduler runs the daily jobs. Market data comes from Yahoo Finance and indicators from the ta library.

• Risk controls in code: a position sizer, stop-loss rules and filters.
• A backtester with metrics, an optimizer and walk-forward validation, so a strategy is tested out of sample.
• A paper trader with orders, positions and trades, and a Telegram notifier.
• LQ45 and IDX80 universes, an equity curve dashboard, and 197 tests across six files.

The result

It is still in development and runs on paper only; nothing here places live orders. The repository is private.