AMD or Intel as well as the memory brand - page 56
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
That's how I would like to see their results in the second version of the optimization, I'm afraid there wouldn't be anything super-excellent there.
The scripts, as their author has even agreed, are spherical horses in a vacuum.
Since properties of such horses are very diverse, I am not going to comment on them since they do not give any meaningful information.
I sorted the table of optimization results of the second Expert Advisor by time*frequency. Almost everything is clear there :) And most importantly, it is quite logical!
And you put the candidates in the table and link to it. Let it be partly empty and remind you of yourself.
That's how I would like to see their results in the second version of the optimization, I'm afraid there wouldn't be anything super-excellent there.
The scripts, as their author has even agreed, are spherical horses in a vacuum.
Since properties of such horses are very diverse, I am not going to comment on them since they do not give any meaningful information.
I sorted the table of optimization results of the second Expert Advisor by time*frequency. Almost everything is clear there :) And most importantly - quite logical!
I agreed? Why?
joo wrote >>
Hmmm, maybe, and most likely you are right..........
The proportion between the script test results and the EA test results remains constant within the margin of error (usually assumed to be 5%).
Thus, testing on an EA is no better than testing with a script.
Did I say yes? Why would you do that?
The proportion between the test results on the script and the test results on the expert remains constant within the margin of error (usually assumed to be 5%).
Thus, testing on an Expert Advisor is no better than testing with a script.
I didn't see any significant difference between testing the script and the EA either. Apart from the fact that there may be some execution ambiguity with the EA.
Testing a simple EA is not a very interesting task. It would be more interesting to test an EA that calls a serious indicator with a large amount of calculations.
The logic of opening and closing positions is rough and unambiguous, so that the number of deals and balance would be the same for all. And this must be checked to assess the validity of the results.
Unfortunately, I don't have anything ready to offer.
Did I say yes? Why would you do that?
The proportion between the test results on the script and the test results on the expert remains constant within the margin of error (usually assumed to be 5%).
Thus, testing on an expert is no better than testing with a script.
Your post of 28.09.09 (22:30) on page 50. Quote "you may be right", in the same post you quote my post about the script being a spherical horse.
And no 5%. One of the slowest on the script, the Pentium 4 670 shows just phenomenal results on EA. The same goes for four2one.
Testing in the Expert Advisor is better if only because we use a real, albeit primitive indicator, real quotes and real calls of trading functions. This is what we need in the real optimization. That is why we need this forum thread. Your scripts are good for answering the question "what processor is better at assigning a variable value in a loop". Raise your hands those who are interested in it!
And begemot61's other processors also show very high performance on the EA while being quite predictable on the script.
I didn't see any significant difference between testing the script and the EA. Apart from the fact that there may be some execution ambiguity with the EA.
Testing a simple EA is not a very interesting task. It would be more interesting to test an EA that calls a serious indicator with a lot of calculations.
And the logic of opening and closing positions is rough and unambiguous, so that the number of deals and balance would be the same for all. And this should be checked to assess the validity of the results.
Unfortunately, I have nothing ready to offer.
You are the one with the ambiguity - see post above.
By the way I suggested a solution to the problem, the link is on page 50. Check it on the attached Expert Advisor with the offered settings.
The difference in the number of trades even with slightly different quotes will be minimal. Just check it and show us.
I haven't seen any significant difference between the script and EA tests.
Belford's Phenom II behaves quite strangely: in the script it shows much worse results than the "blue" Core 2 Duo series, while in the optimization it slightly outperforms them. This is all for the first variants, in my table on page 49.
In short, it looks like Docent's optimization is more realistic. But we can try to do what begemot61 suggests:
It would be more interesting to test an EA that calls a serious indy with a lot of calculations.
I have already suggested testing an Expert Advisor that does not open any trades on page 41. Such an Expert Advisor could really be used to test the work of the MT optimizer. In this case there would not be any misunderstandings.
But this proposal did not interest anyone.
I have already suggested testing an Expert Advisor that does not open any trades on page 41. Such an Expert Advisor could really be used to test the work of the MT optimizer. In this case there would not be any misunderstandings.
But nobody was interested in this proposal.
Perhaps from the fact that the Expert Advisor does not open any trades. There is a certain flaw in it for testing, don't you think?
Perhaps, the Expert Advisor does not open any deals. That is a disadvantage for testing, don´t you think?
I'm interested in how fast the tester works on a particular hardware. The simpler the code, the less influence it has on the results. The tester is wobbling the code compiled by the standard compiler. Therefore, it does not matter what it is - a script or an Expert Advisor. However, there will always be some ambiguity with the Expert Advisor since we will never get the ideal conditions. From my viewpoint, begemot61's suggestion is interesting. But even in this case it is not necessary to run the indicator from an Expert Advisor. The time can be measured inside the indicator.
This is the purpose of testing hardware with various synthetic benchmarks, minimising the impact on what in a particular case should not affect the result. Testing the memory subsystem, processor, bus, etc., then deriving an overall performance index.
The real estimation of a trader's machine would be the creation of several such benchmarks - scripts, non-trading Expert Advisors, indicators, and the subsequent calculation of the final rating.