So this result is rather baffling: running the full backtest from January 2003 until now on an algorithm that trades only once per month - in the first trading day of the month respectively and right when the market opens - should theoretically produce the same result, no matter if I run it in daily backtesting or minute backtesting, since nothing in the algorithm is done differently.
Yet the results are vastly different. Over 15 years, once backtest shows a return of 900%, the other a return of 450%...