Discussion of article "R-squared as an estimation of quality of the strategy balance curve" - page 8

New comment
 

Here we have a picture:

Depending on how the graph is stretched, we see that for the same segment we have a different angle. Suppose we take the second graph and a set of these points as a reference:

Vasiliy Sokolov:
PointValueTime
№112018.03.21
№2103482018.08.13

Then for a set like this:

PointValueTime
№112018.03.21
№2103482018.08.21

This line will already be like this roughly:


The angle of the blue line is obviously smaller than the angle of the red line. Here it is a success! Pinocchios rejoice! However, the problem is that the originally chosen angle of the red line could have been anything really, like 45% or 20 or 1%. In that case, the angle of the blue line would have changed too. I.e. there is no criterion that objectively tells us that the originally taken coordinate system really has this or that angle. So one user who took picture #2 as a basis for the blue trend angle would have 45%, and another user who took picture #3 as a basis for the same blue trend angle would have only 10%.

 
Vasiliy Sokolov:

Here we have a picture:

Depending on how the graph is stretched, we see that for the same segment we have a different angle. Suppose we took the second graph and a set of these points as a reference:

Then for this set:

PointValueTime
№112018.03.21
№2103482018.08.21

This line will already be like this approximately:


The angle of the blue line is obviously smaller than the angle of the red line. It's a success! The Pinocchios rejoice! But the problem is that the originally chosen angle of the red line could have been anything really, like 45% or 20 or 1%. In that case, the angle of the blue line would have changed too. I.e. there is no criterion that objectively tells us that the originally taken coordinate system really has this or that angle. So one user, who took picture number 2 as a basis, will have 45% angle of the blue trend, and another user, who took picture number 3 as a basis, will have only 10% angle of the same blue trend.

That's what I'm telling you, you need to make a benchmark, you know? A person does not need to share with others (if the benchmark is common, you can share), but to give an estimate of the result in a certain sample. If we took option 2, then we use it.

 
I'm not saying it's a good criterion for evaluation and so on, I'm talking about the ability to measure the angle of the picture.
 

Haven't read all the comments. Maybe someone has already written it. There is such an error in the CustomR2Equity(double& equity[], ENUM_CORR_TYPE corr_type = CORR_PEARSON) function:

for(int x = 0; x < total; x++)
      estimate[x] = x*a+b;

and it should be like this:

for(int x = 0; x < total; x++)
      estimate[x] = a+b*x;
 
Once again a high level article from the author.
Thank you
 

Thanks for the great article. 

In the article it is said that one of the drawbacks of R2 method is that it is "Applicable exclusively for estimation of linear processes or systems trading with a fixed lot". I'm trying optimize my strategy for the stock market and I use dynamic position volume based on the ATR value, so I guess I could not use this method in my case. But my question is how one could optimize a non linear system?

 
Normalize the volume : Take the profit and divide by the lot size
 
That is genius. Thanks alot for the great article! I wonder how R2 measuring quality compares to measuring standard deviation from an account balance moving average.
 

Hi,

if I try to use R²:

#include <Rsquare.mqh>
.
.
.
double OnTester()
{
   return CustomR2Balance();
}

Is okay.

but with

#include <Rsquare.mqh>
.
.
.
double OnTester()
{
   Manager.SetCustomOptimizeR2Balance(CORR_PEARSON);
   return Manager.OnTester();
}

I get:

'Manager' - undeclared identifier

How can I declare it in my EA?


How should I declare Manager in my EA

 
Does the archive attached to the article work for anyone? I have 37 errors in the author's classes 
CMessage::Init(ENUM_MESSAGE_TYPE type,string source,string text)
and so on... Does the author's Expert Advisor work for anyone?
123456789
New comment