Mladen, thank you for sharing your genius with us. It's a great gift.
[1]
This is WITHOUT optimisation of the MA used as a crossover filter fore the trend following strategy. selecting a fast and slow (in this case, 10 and 50 MA).
2 year test result:
NORMAL MOVING AVERAGE USED TO FILTER TRADES (SIMPLE CROSSOVER TREND, PLUS 1 ADDITIONAL REQUIREMENT TO FILTER OUT RANGEBOUND)
MLADEN NEW (MCGINLEY) MOVING AVERAGE USED TO FILTER TRADES (SIMPLE CROSSOVER, PLUS 1 ADDITIONAL REQUIREMENT TO FILTER OUT RANGEBOUND).
At the first glance, it appears that the profit is the same - but look at the losses and other metrics. So I consider this as a potentially promising result in the first instance, even though we are looking at 150 to 189 trades over 2 years.
Below [to be edited in later as processing], I use a higher timeframe for the filter, and I also perform the test over around 6 years of 100% quality data, with 6 optimisations on the entry and exit strategy. I do not optimise the filter still (because I will save that for finding the best setting for live trading as a final step). I model both spreads and commission conditions for a more realistic simulation.
Without Mladen's new MA
This is using a single instrument (only one currency actually). I feel that the results are very promising, especially because the optimisation is minimal (not 100s of parameters or options).
Finally, I will apply this strategy and show you a portfolio result which looks very promising for future trading with optimisations say every 3-6 months.
TBC
We keep pushing the limits! 