Our opinion about the Expert Advisor
Information on sales strategy
The seller uses “scarcity” as a psychologically very effective marketing strategy. There is only a certain number of a product for the current price. After that, the price is raised sharply or the product is no longer sold. This strategy builds up an inner pressure in the potential buyer and prevents rational action, after all, one does not want to miss anything. Of course, this sales strategy is not forbidden and is often used, but one should be aware of this strategy and not be rushed.
Backtest not possible before 2011
The backtest is only possible from 2011, backtests before 2011 are blocked by the seller. The seller states that Aura Gold EA has stable results in the years 2011-2020. This indicates that the seller has optimized the parameters for this time period and that results outside this time period are unsatisfactory. In this way, the seller wants to prevent backtesting until 2011. If that would be the case, then there is a deliberate deception of the buyer. (This is only our personal opinion and assumption and not an assertion.) We cannot endorse preventing backtests by the developer, obviously he wants to hide certain results. The buyer must have the possibility to evaluate all results independently.
Signs of possible overfitting
The backtest results look very good at first glance. However, you have to keep in mind that you cannot see the results before 2011. Additionally, it is noticeable that at the time of the Expert Advisor’s release, the results slightly deteriorate. The version 2.3 we tested apparently uses the same algorithm as the version 2.1 from November 10, 2020. The backtest results deteriorate abruptly from November 4, just before the release of this version. You can see the figure of the portfolio based on the three official set files Aura Gold EA v2.3 below. It is impossible to ignore the fact that the date from which the results start to deteriorate coincides with the date of the Expert Advisor’s publication. Both the timing and the way the results deteriorate indicate overfitting. (This is only our personal guess and not an assertion.) Even if it looks like the Expert Advisor is still profitable after November 4, it should be remembered that we have adapted Tick Data Suite to IC Markets. IC Markets has very low spreads, commission and swaps, with a broker with slightly higher costs, the curve could fall. In addition, one should keep in mind that a backtest is only an approximation of reality.
Furthermore, we also tested version 3.1 (released on January 16, 2021). In this version the parameters have been updated compared to version 2.3. Nevertheless, a possible effect of overfitting can be seen here as well.
Comparison of version 2.3 with 3.1
The raw backtests of both versions can be viewed below. It can be clearly seen that the developer has changed the algorithm between the “old” version 2.3 and the “new” version 3.1 (referring to the default settings). In fact, the correlation between the results is only 31%, which means that the backtest results of the two versions are practically uncorrelated. While in the old version many parameters could still be adjusted by the buyer, this is no longer the case in the new version. This means that the seller has used a “different” Expert Advisor in the new version without announcing it in advance. After updating to the new version, the buyer can no longer repeat the backtest results of the old version. This, of course, deprives the buyer of the possibility to detect the effect of “overfitting”. We would like to emphasize that this practice is unprofessional, because it means that you can not rely on the seller.
In the comments and reviews about the Expert Advisor on MQL5 Market, this practice is criticized by the buyers.
Comparison between backtested and live results
We kindly received live results of the Aura Gold EA v3.1 from a trader. We compared his live results with our backtest results. The figure below shows the equity curves of the backtested and live trades. Even though we could only compare 13 trades, the comparison still shows that there is a very high correlation between backtests and real results. This large correlation shows that the backtest results we produce are trustworthy.
Our final assessment
Firstly, the backtest period after the Expert Advisor’s release is too short to make a reliable statement regarding overfitting. To judge whether there is an overfitting, a much longer period after the release has to be tested. Furthermore, the seller does not show a real signal, so one has to “believe” the seller that the Expert Advisor is profitable. In addition, no backtests can be performed before the year 2011.
Unfortunately, we cannot give a definite conclusion. The profitability of the Expert Advisor is unclear from our point of view. Whether the Expert Advisor is worth it for 1,999 USD (February 2021) is something everyone has to decide for themselves.
The seller has not published any live signal (neither demo nor real signal), so there is no proof that the seller’s strategy works.
Raw backtest results
For version 2.3 we used default settings and two set files provided by the vendor. For version 3.1 we used default settings, there were no set files provided by the vendor. As always, we tested with lot size = 0.01.
The backtest results are best viewed on high-resolution screens.
Click on one of the tabs below to open the corresponding backtest.