Advice on live MT4 data source

So I have been backtesting on the StrategyQuant Dukascopy historical data (shifted so there are no Sunday candles) which seems to be decent quality.

I opened a Dukascopy MT4 account and compared the daily candle data expecting it to match. It does not! There are garbage Sunday candles from 4th March 3028 to 12 August 2018, the open and close prices differ etc.

This means any EAs used on these charts will not work. I find it strange that Dukascopy would offer free data that does not match their MT4 server data.

Has anyone successfully gotten around this problem – should I use a different MT4 broker? Any suggestions? Thanks.

Hello, yes, some weeks on DC MT4 chart have Sundays candles while others do not. In those weeks Open & Close prices are not corresponding. Not sure why this is the case. Need to check with Dukascopy support

Some time ago I got the following answer from Dukascopy support.

The main problem with the historical data before May 2018 is that it was taken from the MetaQuotes official servers and it could differ from Dukascopy history for that period. In order to get full Dukascopy data for 2017, I can suggest You to download our main JForex platform and use Historical Data manager. This is a direct path to all stored prices and it fully compatible with the MT4 (it can be exported to MT4 historical tester).
There is no other way at the moment Unfortunately.

 

Hi Tomas

Based on this topic posts can you please confirm:

1. Are all the robustness tests in SQX supposed to eliminate the dependency for profitable strategies on the Data used in their testing?

2. Can we use Dukascopy MT4 Data for all SQX testing to produce profitable strategies (independent of Data)?

3. If the answers to queires 1 & 2 above are “Yes”, then the only issues from Coensio Posts are that his is unable to directly compare profitable strategies produced from SQX on Dukascopy MT4 Data to a Demo/Live Dukascopy MT4 account/backtesting. This is correct?

4. Are there any other recommendations/solutions SQ Support can offer/advise regarding what data and how to use this data to test for potential profitable strategies.

Thanks

Kev

 

 

 

Hello,

1) yes, that’s the purpose of RT test. Especially tests like “Randomize history data” or “randomize starting bar”

2) yes, it can be done to certain extent. No strategy is bulletproof. Impact on strategy performance will differ based on what timeframe, strategy logic, or testing precision is used

3) DC data is just a default solution. We plan to add more data providers into later SQX builds. If you serious about trading you should purchase high quality data. For education and basic strategies sampling DC data is just fine (as mentioned before mainly depends on strategies themselves)

4) there are many providers around, just google some and compare prices and offers