Remember «EMA (5) + EMA (12) + RSI (21)» trading strategy we have backtested manually some time ago? Here you can find the results of the manual backtesting and check either it turned out to be profitable or not.
Now we are back to backtesting of this strategy: this time in the automated mode and with different parameters. Let’s check how to make this strategy bring profits and what are the best settings for this.
This EA was created with the help of our Visual Strategy Builder. The backtesting was performed with the help of Forex Tester 4.
Currency pair: EUR/USD.
Initial deposit: $100000, leverage: 1:100, the min volume of the trade: 1 lot.
The “Bars to skip” parameter defines the search range of a trading signal (in bars). All trades are opened and closed automatically.
Testing period is from 01/01/2017 to 12/31/2018.
Indicators: EMA (5; close), EMA (12; close), RSI (21; close) with level 50.Trading signals of the strategy: Buy:
Entrance at the next bar after the breakdown; the trade closes at the opposite intersection of the RSI line or due to StopLoss.
Risk: no more than 1% deposit per transaction.
Each trade is closed only when a reverse signal appears according to the rules of the strategy. TakeProfit, StopLoss, trailing stop and manual trading are not used.
According to the test results, more than 50% of the deposit was lost, the profit factor is below 1, profitable transactions are 2.5 times less than unprofitable ones, although the average profit per transaction is higher than the average loss.
All statistics results can be named as negative.
Only one trade with a volume of 1 lot was opened at a time, dangerous methods of money management were not used. The deposit was in the drawdown most of the trading time, the maximum drawdown was almost 90%.
An active loss of profitable entry points on a strong trend has been recorded, although the strategy is positioned precisely as a trend one. The signal skipping is to some extent related to the “Bars to skip” parameter, but the test showed that many small illogical deals appear in a narrow flat. Most of these transactions were losing ones.
The main reason for the negative result was the trading technique. Even if the entry was made correctly, then in order to close the deal, the market had to turn quite strongly against the open position so that the indicators could recalculate the data and generate a return signal.
Trailing Stop and TakeProfit were not used, as a result, profitable positions were closed too early on false feedback signals. As a result, a profit shortfall is observed in almost every profitable transaction, and sometimes, in anticipation of a return signal, the transaction could generally lose.
Each transaction opens with a fixed TakeProfit of 75 points, StopLoss of 25 points. It closes only by stop orders. Trailing and manual trading are not used.
If there is a signal, several trades with the same parameters can be opened, but no more than 5. Strategy parameters in the Forex Tester system are similar.
Profit amounted to almost 100% of the initial deposit. The number of profitable trades is also 3 times less than unprofitable ones, but the average profit is 3 times higher than the loss (as laid down in the TakeProfit / StopLoss conditions).
The profit factor is slightly lower than the recommended one, there is almost no margin for recovery. In general, statistics can be estimated as average.
The load on the deposit did not exceed 30%, that is, no more than 3 transactions of 1 lot were opened at the same time, dangerous methods of money management were not used. Most of the test time, the trading account was in confident profit, the maximum drawdown in the short-term was more than 50%.
The fixed profit/loss system has proven to be much more effective, although the strategy still does not get part of the profit on a stable trend. The TakeProfit parameter could well be chosen more. A skipping of strong signals is observed.
On the flat market, the strategy also opens up many small transactions, but due to the fixed StopLoss, they turned out to be uncritical for the deposit. In a flatter market, the strategy will give more negative trades and the result may not be so optimistic.
The purpose of this test: to shorten the search period for trading signals. Each transaction opens with a fixed TakeProfit of 75 points, StopLoss of 25 points. It closes only by stop orders.
Trailing and manual trading are not used. If there is a signal, several trades with the same parameters can be opened, but no more than 3. “Bars to skip” parameter: 20.
In addition to a sharp increase in profit (170% of the initial deposit), an overall improvement in indicators, in particular, the profit factor and recovery factor, can be noted. The overall ratio of trades improved towards profitable ones.
According to the strategy, no more than 3 transactions with a volume of 1 lot worked simultaneously, dangerous methods of money management were not used. The deposit was in profit most of the time, the drawdown zone worked out almost synchronously with the previous test, the short-term maximum drawdown was still more than 50%.
The test showed an increase in the number of profitable trades, and, apparently, each entry was more successful. The drawdown remained at the same level, but in general, a decrease in the “Bars to skip” parameter gave an excellent result, not only for profit but also for system reliability.
In general, the strategy looks balanced and profitable, especially since it has long been known and shows stable results on almost all assets. But there are a few comments.
The system is very sensitive to non-standard volatility and poorly filters any impulses (news, gaps, the speculative thrust of volumes), the overbought/oversold state is not controlled, hence a lot of false signals and inputs are at the end of the trend. This means that just breaking the balance line on the RSI to confirm the signal is not enough.
Flat areas, regardless of their duration, the system performs poorly. In a wide range, the results may be better, but it is still recommended to trade only trend assets on it. It is advisable to add volume control to the system and optimize the oscillator parameters for a specific asset.
Warning: tests were carried out without taking into account spreads/swaps/ commissions. Before using on a real account, it is necessary to conduct tests on the history of at least 2-3 years.
Share your personal experience of effective use of Forex Tester software - on real quotes of any asset and in various trading conditions. Was this article helpful to you? It is important for us to know your opinion - we are waiting for you on the forum!