You still need to add beaten margin and some sort of date selection with your tester.

I have a copy of a system tester that was posted on here a long time ago where you used results from 4 different sets of 150 races to see if the system would hold up.

The main thing was to use different times of the year and different weeks and such to create a more random test.

With your tester we have no idea of the time frame, i have a copy of the software that is sold via this website that is much more advanced and i have created systems from it that win but can go through some years at a loss but you can test this type of thing because it has a date component.