The problem
The client traded on technical setups – price stretching outside the Bollinger Bands, returning inside, and crossing long moving averages – and checked them by hand on charts. With hundreds of symbols and several time frames, good setups were missed and the rules were applied inconsistently. They wanted the exact same rules run automatically on every symbol, every candle.
What we built
- Indicator engine: Bollinger Bands, EMA 73/146/220 and DEMA 80/160, calculated candle by candle.
- Setup rules: outside/inside run counters, band-touch tracking, EMA crossover segments and growth measures, exactly as the trader defined them.
- Multi-timeframe scans: each symbol checked on several intervals with one call, reusing one processor for speed.
- State persistence: the latest state of every symbol stored in SQL Server, so a scan resumes instantly instead of recalculating history.
- Validation: every rule checked against real market data and the trader’s own manual results before release.
Technology
- C# and .NET with a clean, single-call API for the screening engine.
- SQL Server for candle data and per-symbol state, with migration scripts for existing databases.
- Market data from exchange APIs, synchronised in the background.
Result
The screener replaces hours of manual chart checking with a scan that runs in the background and lists only the symbols that match. Because the rules are code, they are applied the same way every time, and new rules can be added and tested quickly. See our trading software development service if you need a screener, alert system or back-testing tool.
