MT4 with Accurate Historical Data

Hey guys,

Just a question re MT4 and not even sure if it is doable or will make any sense at all, but here goes.

I am a fan of MT4 and have designed and had built a few EAs for this platform. Now I appreciate that backtesting is not entirely accurate, but it does me a pretty good idea is something is working or not. I do use several different platforms for this and my results tend to be all over the place with various degrees of modelling quality, but never more than 90% (sometimes).

I have read that 99% can be obtained, and I've downloaded some info on how to do this from various places like Dukscopy etc, but never seem to have had any luck in getting anywhere near 99%. All my testing ends up with either crap data or gaps in the data. So obviously doing something wrong.

Now to my question. Would it be possible for a third party to be given access to a demo account of mine, where they could then take control of it, then add the required historical data, and then return control of the platform back to me. Where I could then run my tests and consistently obtain 99% modelling quality??

This would be easier (and cheaper) than me paying someone to run tests on my EAs and then supply me with the reports.

Thanks, and hopefully it is not too dumb a question. Just don't know enough about the tech side of things when it comes to data and MT4. Cheers. Jim

- I don't think you are doing anything wrong. Your results seem to be typical, or better than most including mine.
- Even if you achieve 99% quality, I doubt if the results are meaningful. The MT4 historical data is 1M and ticks are interpolated which means they are not accurate. Back testing is useful to practice and get familiar with the EA but not for any meaningful results. You need real tick data for that.
- The solution you are proposing is sort of possible. Somebody could run tests on your account, but they would need total access to your machine. Would you want to allow someone to do that? This is more effort for the tester and would probably cost more.

Thanks CodeMeister for your quick response. I appreciate that testing is somewhat unreliable, but I guess I am just looking for a bit of consistency when testing. And at the end of the day, backtests do give me a pretty good idea if something is worth further investigation or not. More of a first stage filter than anything else.

I have no problem giving my EAs to a third party, then give them my testing parameters, and then supplying me with the results. As long as they aren't fudged or photo-shopped etc.

Or if a third party opened a demo account for me (any broker that has big leverge and permitted hedging), downloaded the required historical data into that platform, and then supplied me with the login details. No harm done here as it is only a demo account, but I will have what I need. If that is viable?? Even if I opened the demo account, then gave the third party the login details where they could download the correct data, and then pass control of the platform back to me. Again, no risk, as it is only a demo. Hope that makes sense. Cheers. Jim

That is a much better plan than the first one. There were too many things that could go wrong with that one.

The latest one should be viable. There is still is the access and security problem. I am thinking perhaps a VPS might be the answer. I don't know very much about them, but that would eliminate some of the security issues (maybe all?). Maybe someone else can jump in if they have experience in that area and give you an idea of the challenges. There are a few VPS gurus who post occasionally, but I can't remember their usernames.