About cross check of Opt.Profile/S.P.P-uniform distribution how many changes
6 replies
Jordan
3 years ago #268137
In SQ, cross check of Opt.Profile/Sys.Param.Permutation,In the third item “uniform distribution less than (5) changes from positive to negative” in filtering.
Please see the attached photos for details.
I went to sq’s website and read the paper recommended on the website. I didn’t know what the changes represented and how to set the threshold. If there are relevant articles or websites can be recommended to me, thank you.
tomas262
3 years ago #268153
Hello,
with that criteria, you do not want to see more than 5 optimizations go under “zero”…. you can see on the screenshot attached there were 3 optimization in which the results end up being less than 0
Jordan
3 years ago #268157
I’m sorry, but I still don’t understand. You mean that as long as one optimization result is less than 0, the filter value of uniform distribution will increase by 1, right? However, in the picture in your attachment, there are 3 optimization results less than 0, but the uniform distribution shows 6, greater than 5, so they are failed. Please also take a look at the two examples in my screenshot. It seems that the calculation method of the number of uniform distribution does not mean that the optimization result is less than 0.
Jordan
3 years ago #268159
In addition, the default cross standard for the third item of checks is less than AVG profile + 1 StdDev. According to the results of a large number of tests and the feedback from many other users, 1 StdDev is a little too strict, and few strategies have passed. Is there any statistical basis behind this standard?If there is, please recommend the relevant information or website, I will study it, thank you. By the way, I have read the paper “know your system turning data mining from bias to benefit through spp” recommended in the website software description. It simply mentions this standard, but there is no example or detailed description of why this standard is.
hankeys
3 years ago #268165
uniform distribution doesn mean that it must be in a loss i think
uniform is outside some normal borders – like mean, or average…i dont know exactly how its implemented in SQX
You want to be a profitable algotrader? We started using StrateQuant software in early 2014. For now we have a very big knowhow for building EAs for every possible types of markets. We share this knowhow, apps, tools and also all final strategies with real traders. If you want to join us, fill in the FORM.
Charles Umani
3 years ago #268194
tomas262
3 years ago #268373
It is not a count of “negative bars”. The value represents number of changes from positive to negative. So if you see 1 negative bar it represents 2 changes from positive to negative…
Viewing 6 replies - 1 through 6 (of 6 total)