For example if I test, 7/1 to 7/28 with an optimization window of 5 and a test window of 1 the walk forward results will be...
7/6/2016 12:00am - 7/6/2016 12:00am
7/7/2016 12:00am - 7/7/2016 12:00am
7/8/2016 12:00am - 7/8/2016 12:00am
and on...
And even weirder, the trades on "7/6/2016 12:00am - 7/6/2016 12:00am" (keep in mind this is on a day long test period) are...
7/5/2016 10:30am - 7/5/2016 4:00pm
7/7/2016 10:30am - 7/7/2016 10:30am (This trade's open and close happened within the same second but moved like 70 ticks?)
So both were trades on days outside of the test period.
What is going on here?? Is walk forward testing broken? Is the test fine and the report broken?
I'm using a CQG/Continuum demo for data. Could this be messed up data?
Comment