In part III of our ArrowHunterEa video course, we will perform a robustness test on our optimized settings. In previous two videos we’ve optimized our system and proved that the resulting setting is also profitable in ‘out-of-sample’ period. But even this setting can still be curve-fitted! We need to perform a parameter stability test. To do so we will manually ‘jiggle’ all parameters around the optimized values. This is similar to ‘monte-carlo’ analysis, but we will do it manually, since MT4 platform does not provide this option. For each parameter we will select a range, located ‘slightly below’ and ‘slightly above’ the previously optimized setting. When we will run the optimization, the genetic optimizer will ‘jiggle’ with our initial settings. We will ‘jiggle’ all optimized parameters. Now we want to switch to ‘optimization by profit’, to see the influence of each parameter change on the resulting profit. Let’s start our manual ‘monte-carlo’ stability test. The Optimization Graph now displays the resulting profit using randomly changed strategy settings. Ideally all results should be profitable! This means that the strategy parameters are robust and stable. And that we did not select, this ‘only one’, special, profitable (curve-fitted) setting. There are no negative results so far… We need to wait and finish the optimization. As you can see we are using the whole data set for this test (including the out-of-sample period). Waiting… Now we can skip to 8:20min of this video…. The optimization run is finished. We do not see any negative results! This is also presented on the Optimization Graph. All ‘dots’ are positive. This parameter set has passed the robustness test! That’s it!