At the moment this is just an experiment to see if it is possible to build a self optimizing strategy that can adjust itself as markets change to try to adapt the new market environment. Whether this is more profitable than just picking a fixed set of values is yet to be seen.

The values 40 and 30 might not have been the most profitable from 1995 onwards – maybe they just went into the lead at that point and were never overtaken. Perhaps 50 and 60 had made a loss up until 1995 and then suddenly became very profitable but because they started their recovery from such a low starting point in 1995 they never took the lead even though they made more money from 1995 onwards than 40 and 30 did. Hence why I am now modifying the idea with a third variable for a look back period so that we just look at recent data and see what has been most profitable recently. The idea is to also have this look back period as self optimizing too.

The strategy that I am basing it on appears to be profitable across a very broad range of values for x and y but some of the equity curves are more up and down than others. The hope is that a self optimizing strategy could adjust itself if it finds that it is entering a dip and switch to values that have been recently the most profitable.