9.2 Billion profit in 5 month as back tested in Meta Trader 4 - page 6

New comment
 

could you please send me the EA too

Thanks

Netsu

puturta@yahoo.co.id

 
Zap wrote >>
Backtesting works with real historical tick data only.
Mt4 platform simulate ticks from bar data, where M1 bars are the smallest.
Bars are nothing more than lossy-compressed price data: by making a M1 bar, you keep Open, High, Low, and Close and throw the rest (around 70-80%) of the data away. When simulating ticks, the platform tries to "guess" this missing part, but as you can see, not very effectively: if you compare a period made of simulated ticks, and real ticks, you will see, that simulated ticks often contain unrealistic movements.
These movements lead to unpredictable results: this can lead to the false results of such EAs, as these EAs accidentaly exploit the effect of these unrealistic, false movements.
By collecting ticks, or getting historical ticks, you can backtest on the 100% of the data, and you can accurate results, AND your results of running the backtest on the _same period_, with the _same EA_ will be exactly the _same_.
If you need more info on acquiring ticks, see me at forexzap<at>gmail<dot>com.

I have read alot of discussion about importing good tick data into Mt4 for back testing but one thing I dont quite understand about this, if MT4 simulates ticks from the M1 bars how does importing tick data into MT4 solve this problem ? Will not MT4 just do the same thing with the imported tick data, create M1 bars from it and then simulate the ticks ?
 
SDC:

I have read alot of discussion about importing good tick data into Mt4 for back testing but one thing I dont quite understand about this, if MT4 simulates ticks from the M1 bars how does importing tick data into MT4 solve this problem ? Will not MT4 just do the same thing with the imported tick data, create M1 bars from it and then simulate the ticks ?
You can find info on how to backtest on tick data here -> http://eareview.net/tick-data.
 

I did already read Birts review and I have downloaded all the tick data I need using his scripts, but I must have missed something in his explanation as I never really understood why MT4 would not do the same thing with the new tick data as it did with the tick data it recieved from the broker, i.e. create M1 bars and then simulate the ticks when backtesting

 
SDC:

I did already read Birts review and I have downloaded all the tick data I need using his scripts, but I must have missed something in his explanation as I never really understood why MT4 would not do the same thing with the new tick data as it did with the tick data it recieved from the broker, i.e. create M1 bars and then simulate the ticks when backtesting

MT4 was never designed to save tick data (and it seems like MT5 will not do that as well). It did use to have a "recalculate" option, which if unticked meant that FXT files were not recreated every time 'start' was pressed in the Tester. This meant that u can put your own FXT files and the Tester would use them. Since FXT files structure is known, this can be used to put FXT files with real tick data. Unfortunately, the 'recalculate' option was removed a long long time ago (I think version 210 was the last one that had it). Birt has a workaround for that in his site...
 

oh i see !! It all makes sense now thanks for the explanation Gordon

 



Thread start date - 2007.04.29
123456
New comment
Reason: