Forex Blog

First-hand Forex trading experience and information about foreign exchange market that will be useful to traders

Archives

Unmatched Data Error

May 1, 2011 by

When back-testing an expert advisor in MT4 Strategy Tester it is sometimes possible to encounter an error that significantly spoils the quality of testing:

TestGenerator: unmatched data error (aaa value xxx.xxx at some.date and price yyy.yyy mismatched)

where aaa is either “high” or “low”, xxx.xxx — some price rate and yyy.yyy is some other price rate.

In general, this error can be called “unmatched data error” and informs the trader that some data in the MT4 chart history doesn’t match between different timeframes. For example, if some M1 bar high is 123.45 and H1 bar that contains this M1 bar has high at 123.44, the unmatched data error will be displayed.

So how to treat this error? There are 3 simple ways to do it:

  1. Delete previous chart history and download it again. All chart history in MT4 is stored inside “/history/” subfolder of the main installation folder. New history files can be downloaded via Tools->History Center menu of MetaTrader platform.
  2. If for some reason, previous method didn’t help, you can try reinstalling your MT4 platform. Just uninstall it and install it anew (preferably downloading it again from your broker).
  3. Lastly, you can always switch brokers because it may happen so that the broker’s history data is flawed at their MetaTrader server. In this case, the only solution to the unmatched data error is to install MT4 from some other Forex broker or directly from MetaQuotes.


If you have your own ideas for dealing with MT4 unmatched data error or have some questions about it, feel free to reply to this post using the form below.

2 Responses to “Unmatched Data Error”

  1. Samm

    Well how significantly is the modelling quality reduced by this test data error. I particularly do not see tester testing as a do or die affair given only forward tests (live/demo) define the feedback needed to retain confidence in an EA. So if for instance it is no worse than using the Open prices only option then personally I can live with whatever quality generates (on the grounds of course that one does not necessarily test to assure profitability but also to observe visually how EA trades e.g. responds to highs and lows, navigates through congestion etc). Thanks

    Reply

    Andriy Moraru Reply:

    The difference can be quite significant if the EA analyzes currency rates by ticks and/or uses stop-loss and take-profit levels. There is even a noticeable difference between backtesting in MT5 on “Every tick” and “Every tick based on real ticks.” Of course, forward testing is always the best option and I would even recommend doing it on a small live account rather than a demo account because order execution differs strongly between real and virtual accounts.

    Reply

Leave a Reply

required
required (will not be published)
optional