OmniTrader Forum OmniTrader Forum
forums calendars search
today this week
 
register logon control panel Forum Rules
You are currently browsing as a guest.
You should logon to access more features
A Self-Moderated Community - ALL MEMBERS, PLEASE READ!
Vote for Members who contribute the most to your trading, and help us moderate content within the Forums.


This message cannot be deleted. It is the first message of the thread.
Administrators or moderators may choose to delete the entire thread


 
mholstius

Veteran
100252525
Posts: 175

Joined: 1/13/2017

User Profile
 
Subject : Interpreting Expectancy Strategy Wizard Data
Posted : 1/22/2017 6:16 PM
Post #38720

I’ve had a few questions on how I analyze / interpret the output from Strategy Wizard runs in OT, and I thought it'd be simpler to put together a few snags to try to explain my process. I know a number of you have a lot more experience with this than I, so I hope you’ll pass on any suggestions you might have…

The following examples use a run I made by varying these 2 values on the SP500;





1) MinAvgPPT from 1-3 using steps of 0.25
2) Lookback # Trades from 6-11 in steps of 1

To be consistent in my 2D charts I set them up so that;

1) MinAVg PPT will be the X axis
2) Lookback #Trades will be the Diamonds on the chart “Z Axis”

The Y axis will be the effect on various statistics - easily selected using the drop down.





This chart shows me that the FT # Trades (Y Axis) goes down rather smoothly as the AvgPPT (X Axis) increases. Also, the groupings of Lookback # Trades (6-11) represented by the Z Axis Diamonds on the chart stay in tight groups as AvgPPT increases.

If I put the cursor on any Diamond it will tell me the value of the Lookback # Trades for that point, and the intersection of 2% AvgPPT and the Diamond for #Trades = 9 (“A”) gives a FT NT of 1320 on the Y axis. I’ve added a box showing that the original # Trades without the filter was 15,951 – my feeling is that as fewer trades are taken, the results are probably less robust and more in the category of curve fitting.

Here’s the same data on a 3D chart;




Note that the X Axis (MinAvgPPT) and Z Axis (Lookback # Trades) are fixed and the Y axis (chosen with the dropdown) adds visual depth to the results. It’s the same data, but the depth can make it easier to find the effect of the combinations – and hopefully a robust solution. I’ve got the cursor on the same point (Intersection of MinAvgPPT 2% and Lookback # Trades 9) in both charts. On the 3D chart, the data for the point under the cursor shows up under the chart.

I’ve added labels to the chart to show how the 3D scales / data correspond to the 2D scales / data. I’ve also added Purple labels to identify how to rotate the chart, choose the colors, and choose the type (Bar, Scatter, & Surface) in case you're not familiar with those.


In the 2D chart below, all I’ve done is change the dropdown to FT PPT, which changes what’s displayed on the Y Axis. The X Axis is the same and the Z axis still has Diamonds for each of the Lookback # Trades.




I’ve located the cursor to select the same point (Intersection of MinAvgPPT 2% and Lookback # Trades 9) - which gives a FT PPT of 0.7 on the Y axis. I’ve added a box showing that the original FT PPT without the filter was 0.48.

The 2D chart helps me see that as the MinAvgPPT increases past 2%, the results are more scattered and the FTPPT is largely influenced by the Lookback # Trades chosen. I interpret this to mean that combinations using MinAVgPPT over 2% are probably less robust, and the pairs giving high FTPPT are probably just curve fitting.(?)


A 3D chart using the same data adds the benefit of depth;




Once again, I’ve located the cursor to select the same point (Intersection of MinAvgPPT 2% and Lookback # Trades 9) - which gives a FT PPT of 0.7 on the Y axis.

I find it easier to locate the areas where the input pairs yield output values (FTPPT in this case) that don’t vary widely if the input values are changed slightly. I’d rather choose a pair that has stable results in the surrounding pairs in the hope that the choice is more robust.

I’ve also attached a file named “AvgAPR Expectancy Using 435 ETFs.png” that contains 9 side-by-side 3D and 2D charts from a Strategy Wizard run I did using VBX3 against a list of 435 ETFs. The run varied the Min Avg APR from 0-200% in steps of 20% and the Lookback # Trades from 1-20 and took about 14 hours to complete.

The file is too large to insert in this post - but please feel free to download it (rt click / save link as), zoom in a bit, and scroll down through the charts. I hope it’ll help illustrate the information I’ve tried to share above.

As always, feedback and suggestions are appreciated – and good luck in your trading,
Mark


[Edited by mholstius on 1/23/2017 1:25 PM]

Attached file : 01-Port Sim Values.png (7KB - 857 downloads)
Attached file : 02-2D chart 1.png (150KB - 954 downloads)
Attached file : 03-3D chart 1.png (181KB - 974 downloads)
Attached file : 04-2D chart 2.png (129KB - 1092 downloads)
Attached file : 05-3D chart 2.png (283KB - 954 downloads)
Attached file : AvgAPR Expectancy Using 435 ETFs.png (4248KB - 11 downloads)

Deleting message 38720 : Interpreting Expectancy Strategy Wizard Data


Nirvana Systems
For any problems or issues please contact our Webmaster at webmaster@nirvsys.com.