We offer the following services:
- Testing strategies on real tick history from various brokers (Dukascopy,
- Alpari, FXCM, TrueFX…) with floating spreads, slips, etc.
- stability tests
- analysis of trading results
The main differences in modeling methods in the strategy Tester MetaTrader 4 (MT4)
1. Using the “regular” mode of the strategy tester. Modeling accuracy (according to the strategy tester) – 90%.
We consider only the option of working with quotes (historical data) obtained from a REAL account and in the mode “All ticks (the most accurate method, based on all available lower timeframes for generating each tick)”. Data from demo accounts are not required to be at least somewhat adequate. In addition, it should be noted that due to differences in quotes, the tests of the same “expert Advisor” on the data obtained from different brokers are likely to be different for the same tool (with identical settings and all other things being equal).
So – the “All ticks” mode in this case uses the data of not only the nearest smaller timeframe, but also all available smaller timeframes to generate a “tick flow”. Therefore, it is highly desirable to have minute data covering the entire range of testing.
Spread – you can select “Current” or set any FIXED value.
This mode of testing will be suitable for advisors in the periods from H1 and above. For expert scalpers working on smaller periods and more dependent on the smallest nuances of price behavior, this method of testing is suitable conditionally, purely for primary culling.
2. Use of tick history uploaded “manually” or with the help of special programs (for example Tickstory or StrategyQuant Tick Data Downloader). Modeling accuracy (according to the strategy tester) – 99%.
In this case, the flow of ticks for modeling is taken directly from the file. The reliability is higher, but … Ask is still calculated as Bid plus fixed spread, which is far from the harsh reality, not to mention the existing real-life slippage swaps and so on.
Anyway – it’s better, but I would lower the confidence level to 97 percent.
3. The best option is to use Tick Data Suite. The accuracy of modeling (according to the strategy tester) is 99.90%. There are doubts with all due respect to the program.
It allows you to upload the tick history of any of the dozen brokers (Dukascopy, Alpari, etc.) with the preservation of prices and Bid and Ask. Thus, the simulation uses a REAL spread, which during important news can reach several tens of pips, which is radically different from the fixed spreads used in previous methods.
The control button for modeling parameters is integrated directly into the strategy tester interface, which is very convenient.
Here you can select the modes:
- using floating/fixed(absurd :)) spread with additional settings.
- the use of the slip, again with additional settings.
- set up calculation of swaps and commissions