> However, Monte Carlo keeps failing. I am using the default Monte Carlo options.

Monte Carlo tests the robustness of the strategy by modifying historical data, market variables, or strategy parameters. If a strategy fails, it means it was overoptimised.

Overoptimisation (or curve-fitting) is the major problem when developing and testing strategies on historical data.

How can we fight that?
- using Monte Carlo - it tests the strategy behaviour in a changed environment
- using Multi Market - it tests the strategy on different data sources (usually compatible symbols). Testing in the same market (symbol/period) but with a different broker also helps.
- Out-of-sample testing - when we backtest the strategy on data not used during the building stage. usually newer data the backteser never seen.

How can you make it easier?
- Use the Reactor to automatically generate strategies and pass through different validation tools
- Use Validator to test already created collections.

Here is an example of the Reactor with Monte Carlo enabled. It generates strategies and automatically validates them. Only the strategies that pass all of the enabled validation tools are pushed to the Collection.