Ciao, You have in sample results. Basically you have an high over-fitting risk few things to do in order to get more realistic results are: - Introduce WFO, and test the script with out of sample data - Oversampling - the stop larger than tp sound very suspect...
