Dear Traders, Look at the following two graphs, and check which Optimization graph represents better trading Robot?
Back Testing with a specific setting values, will not prove the Expert is working. Therefor the optimization process is need to be done in full with the below Instructions.
A

B

As we know for a fact that optimization process with genetic algorithms, shuffle the external variables of the robot similar to a binary search, in order to find the highest results of the developing population of results as the evolutions of the cells performing the search for the optimal settings. (ref : OptimizingWithGA.PDF - https://www.msi.umn.edu/sites/default/files/OptimizingWithGA.pdf )
In Order to select the most probable and profitable settings for the external variables, by examining the Y axis of a desired Profit, you will need to check how many of those in the X axis have succeed in producing the same profit. As an example lets take the two graphs above, and check visually, which of the two have more population in relation to a selected Y axis profit. If we look carefully, for example in Graph B the Profit of 180 USD has the least amount of population as the X is increased. One should also take into account the number of produced results in the X axis, as the number of results increase the probability of the expert increases for it's success. Which Means that, the probability of having profit with external variables selected in the area of 180 USD, is low relative to the 180 USD value in the Y Axis of graph A.
In short, from the above two graphs as an example you can test and compare any different expert advisors with the two graphs produced by the optimization process, keeping in mind that these graphs change for different time frames and different conditions of the Market.
Back Testing with a specific setting values, will not prove the Expert is working. Therefor the optimization process is need to be done in full with the below Instructions.
A

B

As we know for a fact that optimization process with genetic algorithms, shuffle the external variables of the robot similar to a binary search, in order to find the highest results of the developing population of results as the evolutions of the cells performing the search for the optimal settings. (ref : OptimizingWithGA.PDF - https://www.msi.umn.edu/sites/default/files/OptimizingWithGA.pdf )
In Order to select the most probable and profitable settings for the external variables, by examining the Y axis of a desired Profit, you will need to check how many of those in the X axis have succeed in producing the same profit. As an example lets take the two graphs above, and check visually, which of the two have more population in relation to a selected Y axis profit. If we look carefully, for example in Graph B the Profit of 180 USD has the least amount of population as the X is increased. One should also take into account the number of produced results in the X axis, as the number of results increase the probability of the expert increases for it's success. Which Means that, the probability of having profit with external variables selected in the area of 180 USD, is low relative to the 180 USD value in the Y Axis of graph A.
In short, from the above two graphs as an example you can test and compare any different expert advisors with the two graphs produced by the optimization process, keeping in mind that these graphs change for different time frames and different conditions of the Market.
Attachments
Last edited: