Forex robots (expert advisors) for algorithmic trading in Metatrader

Probably if you are new to the forex market you will ask this question. Maybe you think that the human brain can process the chart information better and make better forex analysis and decisions about the possible market directions. Well... you are wrong. The forex market is unpredictable as the most things in our lives. If the market becomes predictable, it will stop existing. Then what can we do? We can monitor and find specific market situations after which something is happening. Of course, you will never find such case in where after A follows B in 100% probability. But you can find a 60% one. And this will lead to developing a forex trading strategy and when the strategy is crystal clear, it can be coded into expert advisor. If we need to make an expert advisor definition, it will be - trading robot based on statistics.

No matter how many stats you watch, no matter how good the equity curve looks, you never know how it will perform in the future. It is possible that an expert advisor that has medium results, to be much better than the best forex ea now. Let's say that a good looking expert advisor has a 60% chance to be also good in the future. So if we choose 1, from which it will be - from the 40 loosing or the 60 wining? Again... we don't know. But the bigger portfolio of experts we make, the closer we get to that 60 -40%. What if we monitor closely the things and shut down the bad performers and substitute them with better expert advisors? Then we can reach 70-30 or even higher.

As everything around us, the markets change too. The swaps change. Economic results change. Market volumes change. Trends are becoming bigger or smaller, ranges are becoming wider or thinner. The forex expert advisor is designed to work within some specific market parameters. When they change beyond a certain limit, the ea will stop working as expected. Usually, this doesn't happen right away but like a process of constantly getting less and fewer profits. We must be prepared that all the expert advisors will fail at a certain point and the questions are not if, but when.

There are many types of trading systems and many lifecycle models but the one we choose has tested, out of sample test and trade. The test is the largest period where we test our rules. It must be big enough to get enough trades for good stats and not so big, because as we said in the previous paragraph -it can be close to its end or have bad backtest results just because we are including different market conditions. For example, 10-15 years ago we had 5-8 points spread on EURUSD, now we have 0.8. If we test on such old data we will have good results wich are not real. We have chosen that backtest period to be close to 2 years. The next important period is the out of sample test. In this test, we verify that the system still have the same performance parameters as in the test period. If the system can pass that period it is ready for trading. In the most common case, we have 2 years backtest, 6 months out of sample test and 1-year trading.

Yes. The trading robots logic is restricting them to have more than one position at a time. This means that they may have many entry signals, but they can use them only when there are no other open positions. This means that if you download an expert advisor and start it at a different time/price, your positions can be different from the same system which is started at a different time. Usually, after several local trends, the systems will start making the same positions, but there is a chance that this will not happen for some specific trade logics.