August 20 / Backtesting on Signal Sigma [Measure-Trace Update 1]
Backtesting is Live on Signal Sigma!!
We’re unspeakably excited to introduce our first backtesting feature called “Measure-Trace”.
This powerful tool transforms our Stock Screener by enabling time series analysis and signal generation, allowing you to track metrics historically and uncover actionable insights.
Below are detailed instructions, use cases, and current limitations for Measure-Trace. Updates to this feature will be very regular, so keep an eye on the Research Hub and the Community on X for the latest developments!
Overview
The Measure-Trace backtest allows you to generate a time series chart sourced from stock screener metrics (e.g., GEX to volume for SPY in the video) and derive signals from them. Unlike the screener’s single-point-in-time view, this feature automates historical data tracking, eliminating tedious manual copying. Backtests can be output as Strategies or two Market Study types: Measure-Trace (covered here) and Signal-Trace.
Instructions
Follow these steps to use Measure-Trace in conjunction with the Stock Screener:
Create a Screener:
Navigate to the stock screener.
Click "Create New Screener," name it, and confirm to load all tickers.
Select Tickers and Metric:
Manually select tickers (e.g., SPY, QQQ) from the ticker column. Applying filters is not yet supported.
Choose one metric to track (e.g., Total GEX to Volume, “Dark Pools BUY Ratio, Gross Profit Margin, EPS, Z-Score) and apply it to view the current value.
Launch Backtest:
Click "Launch Backtest" to open the Measure-Trace configuration form.
Choose "Static" mode (dynamic mode, for historical ticker selection, is planned for future - won’t be different).
If the metric doesn’t appear in the dropdown, refresh the screener.
Select the metric from the dropdown menu.
Choose an aggregation method (e.g., average, sum, median; for single values, it’s irrelevant).
Name the backtest.
Optional: Secondary Study for Signals:
Click "Next."
Enable secondary study to generate signals (or skip for basic measure trace).
Select a study (e.g., "Copy of Primary Metric").
Define signal conditions (e.g., "Above Threshold").
Name the signal.
Finalize and Run:
Enter a backtest name and optional description.
Set a simulation start date (e.g., January 1, 2025).
Review estimated trading days and resource usage.
Opt for email notification on completion.
Click "Confirm and Launch" to create a Market-Study (status: "Initializing").
Wait for the email or refresh the page to track progress.
View Results:
Results display a time series chart (e.g., Your Tracked Metric over time).
Add benchmarks (e.g., SPY) to the chart, adjust scales or move to panels for comparison.
For signals: View overlaid signals, and adjust opacity if too bright.
Use the Signal Analysis tab to evaluate performance at various intervals.
Use Cases
Measure-Trace unlocks dynamic analysis for stock market research:
Historical Tracking: Create time series for any screener metric and visualize trends without manual data collection & entry.
Correlation Analysis: Overlay metrics with benchmarks to study relationships, like how a certain metric impacts price movements.
Signal Generation: Generate signals to identify regimes like positive gamma (example in the video).
Performance Insights: Analyze signal effectiveness at various time intervals.
Future Development: Planned enhancements include dynamic filtering, more secondary studies, scheduled backtests, advanced signal generation (using screener tabs as logic gates), signal filtering tools, and strategy integration.
Known Limitations and Bugs
The initial version has some limitations, with fixes in progress:
Manual Ticker Selection: You must manually select tickers; filtering or other columns aren’t supported.
Chart Scale Reset: Adding benchmarks resets the chart to the first timeline date; click the return icon next to the zoom button to reset.
Dynamic Screener Mode Inactive: Static mode is the only option currently.
Simulation Start Date: The earliest available historical date is 2024-01-02 (this will be updated as the database renders)
Roadmap
We’re working on:
Dynamic filtering in screeners for automated ticker selection.
Additional secondary studies for richer analysis.
Scheduled backtests with email alerts for signals.
Advanced signal generation using screener tabs as logic gates (Signal-Trace type backtest)
Signal filtering tools in the Signal Analysis tab.
Integration of signals into Strategies.