Missing Trades - Request to Metatrader Developers to explain this problem - page 2

This is another topic, so contact me via email.

Do not worry!
It is impossible for the backtest is same with real trade.
It is only a reference!!

till now, in three years, I tested a lot of EA, no one can own same behaviour in backtest and real trade!!

DxdCn: Have you tried it with tick data, or just the MT4 generated tester data?

The problem is eg., that many times, large FOREX movements happen in some minutes. These are one direction moves. The generator spread the ticks though, so the tester handle it as if the price had travelled up and down, and up and down again, opening, closing many more times than it would have been possible, and blah, blah, many more unreal situations.

Using tick data, I get DIFFERENT results than with usual M1 testing. And it is very close to truth, except very volatile bars' spread changing and requotes, and swap, but you can calculate with them, at least you can see whether the logic of your strategy works or not.

It is indeed sad that MT has not responded to my queries. I have emailed them my EA but they have not made any comments.

MT PLEASE REVIEW THE EA I EMAILED YOU AND LET ME KNOW WHAT WAS THE PROBLEM!!!!!!!!!!

tradermaji please refer to bottom of first page of this discussion. I've answered.

once more: "Your EA is very sensitive to ticks data. Modelled ticks can differ from real ticks"

I have heard some said the modelling quality shows the accuracy of backtest and real trade. However, I still don't believe in this issue. Anyone can confirm this? Thanks

Since baxktest's data are different with data used by demo account in realtime !!

MT use model data from real data for backtest, demo account use real data self.

In other word, in demo account EA follow any time change with price change, i.e. tick change.

but in backtest, only have set of one HLOC data for one period, so MT use two model to generate tick change to simulate real, use only open in third model to simulate real.

so some time test results almost same with real, but most time they are different, difined by your EA code design,

as stringo said defiined by your code's sensitivity to ticks data

in theory, data have all tick and all tick change information is same as real time, but no one recored such data, anddo not need such data.

so check code and make it not so sensitive, but what is sensitive is not clear defined, only try and try to refine code.