Modelling quality issue

General discussion about the Tickstory Lite software package.
Post Reply
phil09
Posts: 3
Joined: Fri Aug 21, 2015 7:13 pm

Modelling quality issue

Post by phil09 »

I'm just bought the Tickstory 1.62. My problem is I get 99.9% backtesting quality only in timeframe M5 or M15.
M1 timeframe modelling quality is 24.94%. So what is the problem?
I started MT4 via Tickstory launcher.

Thanks Phil

tickstory
Posts: 4895
Joined: Sun Jan 06, 2013 12:27 am

Re: Modelling quality issue

Post by tickstory »

Hi Phil09,

Taken from the FAQ:

Q. I am not getting 99.9% modelling quality on my back-tests or;
My MT4 back-test reports that the modeling quality is "N/A" or;
My MT4 back-test reports "TestGenerator: internal error because file not opened"

A. This problem is typically due MT4 not utilising your exported data correctly. Check the following:
  • Ensure you have launched your MT4 terminal via the Tickstory 'Launch MT4' option (press F8 when in Tickstory to do this).
  • Ensure you have exported your MT4 database to the correct location (make sure you have selected the correct server if there are multiple selections). For example, if you have a "Demo" and "Real" account, make sure you are exporting to the same one you are using for back-testing, otherwise the data will not be visible to MT4.
  • If applicable, ensure you have correctly mapped the Tickstory symbol name (eg. EURUSD) to the MT4 equivalent symbol name (eg. EURUSDm, fcmEURUSD, etc). (See the manual or search the forum for more on this topic: http://tickstory.com/forum/search.php?k ... c=1&sf=all).
Hopefully this helps resolve the issue.

Regards.

phil09
Posts: 3
Joined: Fri Aug 21, 2015 7:13 pm

Re: Modelling quality issue

Post by phil09 »

Thanks for reply, my problem solved: I was run the Metatrader from windows, not from the Tickstory, and I think the fxt file was overwritten.

But have another problem: when the backtest quality is 25% I got about 450 trades, but with 99.9% quality only 4 trades.
So what is the problem?

Tnaks, phil

tickstory
Posts: 4895
Joined: Sun Jan 06, 2013 12:27 am

Re: Modelling quality issue

Post by tickstory »

Hi Phil09,

There's a big difference between the two data sets and you can expect big differences in trade results. For one, you could have had (say) a spread of 1 pip when you initially back-tested, causing many trades to trigger in your EA. Subsequently, the spread could have been set at say 3 pips, causing your EA not to trade. Of course, this is all dependent on your EA and what it does and how it trades. My advice would just be to work forward from here because 99% modelling quality is going to be much more accurate than your previous tests.

Hope this helps.

Post Reply