I have been building automated trading systems in ProRealTime for quite some time. They tend to be profitable for a period, and then—like all systems—they eventually suffer from alpha decay and have to be stopped. Recently, I’ve been watching presentations and interviews with traders who have studied the survival rate of automated systems using tools like StrategyQuant.

This led me to wonder whether anyone has managed to analyze the survival rate of trading systems using ProRealTime itself. That’s why I’d like to open this forum thread: to discuss how to analyze and improve the predictive power and robustness of automated systems built with ProRealTime.

To keep the discussion practical, I’ve uploaded a recently built strategy to the strategy library. It passes robustness tests such as Monte Carlo simulation and Walk-Forward analysis, and you can find it here:
https://www.prorealcode.com/prorealtime-trading-strategies/nasdaq-misilh1longs/

I’ve also attached a few screenshots so you can see the starting point for analysis and potential improvements.

Questions for experienced users:

How would you evaluate the future survival rate of a strategy like this one?

I’m especially interested in:

🔸 Opinions on using ProRealTime to estimate the survival rate of automated strategies and to improve their predictive capability.
🔸 Beyond a simple forward test, what other methods do you use to assess whether a strategy is likely to keep working in new market conditions?
🔸 Which metrics do you consider the most reliable (PF, CAGR, DD, Expectancy, Walk-Forward robustness)? If possible, could you share concrete examples?

I’m open to creating a working group and/or collaborating with existing ones focused on analyzing the predictive power and robustness of automated systems built with ProRealTime.

Thank you all very much.

Merry Christmas and Happy New Year all. 
Alfonso

Portada-Post-PRC-NQ-M5-MisilH1TSvelasExit1-91.jpeg WF-NQ-M5-MisilH1TSvelasExit1-91.jpeg NQ-M5-MisilH1-TSvelasExit1-91.jpg