I am trying to do view results of optimization over a longer time-period (~ 6 to 10 years) and the results often look much more noisier towards the end of the period - and also exponential if the strategy works. The log scale can probably highlight the true trend of the strategy.
For instance, attached below is the sample graph (of a 3 parameter system) run over optimized data-series (effectively 4 parameter) and naturally for a 4 parameter system there will some results that will look too good.
Comment