One of the major goals of any system testing effort is the determination of the optimal set of parameters to trade. The Parameter Stepping Graphs allow traders to see how the values of a given summary measure varies as the values for a particular parameter changes.
For a single parameter test run, the above Parameter Stepping Graph shows a line indicating the values of the measure for each step in the test. For example, the above graph for a stepping test of the Donchian System's Entry Breakout parameter shows an almost 45% CAGR% value at step value of 28 days for the Entry Breakout dropping to 35% as the Entry Breakout moves higher to 31.
For tests with more than one parameter, the Parameter Stepping Graphs shows a band of values for each parameter step.
The above graph shows a dual stepped test using the Entry and Exit Breakout value creates channel results in each parameter's graph. Entry Breakout parameter days stepped from 14 to 60 days, and the Exit Breakout parameter stepped from 4 to 28.
The above graph shows another test that also stepped the Exit Breakout from 4 to 28.
This graph channels means that for each of the steps on the Entry Breakout for the Exit Breakout combined created graph plots that show the average, minimum, maximum and a band that is one standard deviation above and below the average for each value.
A narrow band around the average indicates that the other parameters affected the results in a relatively minor way. A wide band indicates that the other parameters affected the test in a large way.The wider the band, the greater the influence the other parameters had on the test measure.
The preferences item Measure of Goodness Index from the Reporting General preference page controls the measure used for the Parameter Stepping Graphs. In the above examples, this value was set to CAGR%.
Edit Time: 6/25/2017 11:43:12 AM
Topic ID#: 184