We launched a new offering recently called PsyQuation EA Backtester, and are here to answer some of the common questions. Be sure to check out the product page at backtester.psyquation.com if you haven’t yet.
If you are confused about what the system is or why it’s necessary, this post is for you.
- What is backtesting anyway? Is it absolutely necessary?
- When you are developing an automated trading strategy, you need a way to test it without risk. You can trade it on a demo account, but it can take months or years to be confident in its performance. What you would do instead is test the strategy on the past – effectively running it on old prices to see how it would have performed if you started trading it years ago. This is an essential step to developing an Expert Advisor.
- Is there a wrong way to backtest?
- As a matter of fact – yes, there is. The most common and serious issue is called overfitting. It occurs when you inadvertently optimize the signal to random price fluctuations – or noise – in the past, so that the strategy looks great on paper, generating crazy profits. In reality, this noise is random and will never reoccur, which means that you will surely blow up your account if you run the strategy live. This happens when you optimize the parameters of your strategy until it performs well on past data, then take the best combination of parameters.
- Can’t I just hold out the last 6 months for testing after I am done with the optimizations?
- This is the most obvious and common way to fight overfitting, and while it is better than plain optimization, it has serious drawbacks. For one, you will rarely just do one test run on the hold-out prices. In reality, you will probably go back and change some parameters if performance on the hold out set is not up to your standards. As a result, you end up overfitting on the hold out set again. But even if you only allow yourself one run on the hold out set and discard the strategy if it doesn’t perform, the use of a hold out set is detrimental in itself, as it limits the amount of data you can work with, eliminating the most valuable part – the last few months. Research backs this up
- What is PsyQuation’s approach?
- We use state of the art techniques to compensate for the number of independent strategies generated as part of the optimization, and calculate the statistical significance of the Expert Advisor’s performance. It is based on research by leading mathematicians in the field.
- What do I get from the Analysis Report?
- Once we analyze your Expert Advisor, you will get a report that will provide you with a PsyQuation EA Score, as well as the best optimized parameters for you EA. We choose the most robust parameters that will get you the most stable results in live trading going forward, and will not be outliers, unlike regular MT4 optimization.
- Will I know how much money my EA will make?
- Our analysis only determines if your EA is statistically significant enough to be confident about its profitability, without providing exact performance values. We provide backtest performance as a reference for comparison only.
- What does the PsyQuation EA Score mean?
- It is a measure of statistical significance of the EA’s edge. Basically, if it is sufficiently high – 75 and up – your Expert Advisor has a good chance of being profitable
- How do I use it?
- The process is simple: you send us an MT4 Expert Advisor, as well as the parameters you want to optimize. We run the backtests on different combinations of symbols and timeframes with different parameters, and then analyze the results and generate the Analysis Report. You will get your Analysis Report within 2 business days.
Take two steps to get a sophisticated analysis of your EA, including the input parameters to ensure the most robust result.
- Open a new live AxiTrader MT4 account and fund it.
- Once you are done with step 1, go to backtester.psyquation.com and leave your email in the form. We will send you detailed instructions on how to get started with the analysis.
If you have any additional questions, do not hesitate to ask them in the comment section or direct them to email@example.com and our Data Science team will clarify any concerns.