The strategy tester built into MT 4 allows you to evaluate on a demo account how much a tactic could have been had it been launched earlier. The possibility of testing at different time intervals allows you to see how the strategy behaves in “stressful” situations, how viable it is, what the level of risk is, and also to choose the optimal timeframe and pair. And if the results of the strategy in the test mode are positive, is it possible to run it on a real account?
How to test strategies in MT4
The built-in MT4 tester provides three types of situation modeling:
- At opening prices. The fastest option, but at the same time the most inaccurate. It happens that it makes no sense to use it.
- By checkpoints. The “golden mean” of testing – takes a relatively short time, but the accuracy gives an average. That is, it is not a fact that on a real account, the strategy will “shoot”.
- For all ticks. The most accurate method, but it is also the longest (not everyone will have enough patience). It is noteworthy that a loss-making strategy on the previous method (falling equity) on this method can turn into a profitable one.
It would seem that if the strategy tester in MT4 on the third method shows the best results, then it only makes sense to use it and? We take a strategy and check it by the second two methods:
This is a control point method. Equity is ascending and relatively flat. The strategy seems compelling.
But after running the test on all ticks, everything turns out to be not so rosy: deep drawdowns appear, the profit decreases. When testing by checkpoints, the strategy tester in MT4 does not analyze every price movement, but only takes price values after a certain period of time, due to which order execution occurs. As practice shows, testing by ticks improves the results of trading strategies according to the trend, and against counter-trend strategies (as in the figures above), the result is the worst.
Suppose a satisfactory result is obtained on the exact test method. We change the timeframe and it turns out that the strategy is not so good. The question arises: if, for example, a tester shows a profit on D1, and a loss on M5, then what result to believe? Or you can’t believe any result at all?
And one more thing: suppose that you decide to test the same strategy on Friday and Saturday, without changing anything in the settings. You will be surprised, but the results may be completely opposite. In the “Character Properties” you can see which spread is taken into account when testing. Before closing the week, the spread is increased several times and the tester on Saturday will take into account the value of the spread for closing Friday. A significant increase in costs will be the reason why the strategy was profitable on Friday, and on Saturday the tester draws a loss.
Conclusion. Can you trust the strategy tester in MT4? The answer is unequivocal – no. The effectiveness of the strategy in the past does not guarantee that there will be success in the future. The tester does not take into account slippage (another weak spot), accuracy of quotes is important (the results will be the most accurate for the last 1-2 years), the type of spread is important, etc. In other words, there are countless variables that can affect the result, and which combination of these variables will be correct will be known only after launching the strategy on a real account.
What should investors do:
- Use a tester to run trading advisors, but do not rely on the results obtained. The strategy tester in MT4 gives only a general idea. If the results are different with different settings, the output is unequivocal.
- Determine exact test criteria (timeframe, method, etc.).
- Learn to optimize strategies.
And how do you determine the efficiency of strategies? Share your thoughts in the comments!