Backtesting - Real Tick Data / Tross - page 10

New comment
 

I'm searching for a EA with 99% modelling quality

Hi,i'm searching for Expert Adivisors with 99% modelling quality that with work with metal 4.

Thanks and Regards

 
tunji123:
Hi,i'm searching for Expert Adivisors with 99% modelling quality that with work with metal 4. Thanks and Regards

To get 99%, you would need to use third-party tick data (or collect your own). Unfortunately, since Metaquotes have removed the 'Recalculate' option, you can no longer do either of these - you are stuck with Backtesting using the sh*tty data they supply! What people should be asking themselves is why Metaquotes have done this...

Yep, Metaquotes, the traders friend...

 
 
codybear:
I found this in my travels and have never seen it discussed on these forums.

the attachment is the same as this post but in a zip file

Will you be so nice and can you try to make some tick data conv. for us, for example GBPJPY and EURGBPfor few monts back? I was trying, but seems i am too unhandy in it

Please and thank you very much

 

Indicators not accurate with backtest

I think good data and 99% moddeling quality is a lot better than 90% moddeling quality. The only problem with backtesting is that the indicator of the current bar change with the price - this doesn't happen with backtesting. Put for instance a 3 period moving average on a hour chart and see how the value changes. If you use the current bar and not a close (previous bar) or open price for current bar I believe that the backtest is worthless even with 100% moddeling quality. Trade a system for a month and then backtest it. You'll see that your backtest entry price is different from the account history price although it is the same bar. During news you also get the wider spreads which you won't get on demo trading etc. I wish the indicators also updates for backtesting because this will improve backtest results.

 

hello,

i have no experience in tradestation, or any other "expensive" platforms, is there any platform that let us make a perfect backtesting? (without care about wider spreads)

it looks like the only way to make a perfect backtest is having a system that works only in closed bar prices, wich would be a retarded system... then no fake signals at all

i have found this page guys:

BeckTest EA Perfect, Real different. - MQL4 forum

regards!

had

 

Doing this 99% backtesting modeling is getting really difficult. The instructions are somewhat incomplete. Could some maybe clear some things, so I don't have to constantly bother Tross.

1) Copy Downloaded *.csv files to /experts/files(Naming it eg. EURUSD_ticks)

2) Copy FXTHeader.mqh to \experts\include

3) Copy simple_csv2fxt.mq4 to \experts\scripts

4) Copy GainTicks2fxt.mq4 to \experts\scripts

5) Restart Terminal

6) Run terminal without a connection*

Now this is what is confusing me, if I have say M1 data, the backtesting is going to use the M1 data located in history folder. Also if I run the script on the chart, I get all the appropriate files.

Ok so say I copied *.fxt to tester\history

Then the next instruction says not to recalculate, where the heck is that?

I dont understand how your going from copying files to recalculating, if someone doesn't mind clearing that up, that would be nice.

Thanx...

 

Recalculate exist in older version of MT4, like before 200 or so. If you can't find it, it means your MT4 is the newer version. So don't bother.

But, to tell you the truth, I don't know if this will affect your backtesting or not.

 

So that means there is no tick by tick way of backtesting?

 
TheLegace:
So that means there is no tick by tick way of backtesting?

Read what I wrote just a few posts ago...

1...34567891011
New comment
Reason: