The first step of my optimizing project with training on a top 40 global Indices (MJ40) instrument list with data from 1999 and forward and using instruments with long history have the first round of results.
As with my previous and similar test with the Forex30 instrument list the long history from the training data make it possible to do a correct conducted Out-of-Sample test, the cut off was between 1998 and 1999.
Notably with the test periods was that any period close to the cut off had better results with a significant drop in test values after 2008. The target calculation preferred the shortest period of time available. Regarding the features it was clear that no one of the top 20 was using any kind of date related information.
The best setup from each of the 16 three years periods will later make up the base for each sub classifier for the IQ19o_FX30 automated trading system.