Hello, Matthew. Slowly but stably your Expert Advisor comes close to leading positions. How did you estimate you EA’s possible success before the Championship? Did you expect to see it in the Top 10? Or, perhaps, you are sure it will become a leader?
I went through several optimizations and re-optimized it once before I sent it in using about a year’s worth of data. I wouldn't say I expected it to be in the Top 10, but I am not very surprised that it is. I will be surprised if it finishes in the top 3, because there are so many competitive EAs in the competition.

Could you, please, describe the trading strategy you are using?
It’s a type of crossing moving averages. When one crosses above the other, you enter a buy, when it crosses below, you enter a sell. The code is set up to be able to use TakeProfit and StopLoss levels, but the code running on the competition has very high TP and SL, so that it will likely never hit them. It simply waits for a signal in the opposite direction. So it is always in a trade, switching back and forth between a buy and a sell.

Very attractive is one of the features of your strategy: the EA does not allow losses to grow. How is it achieved? Describe the used mechanism, please.
Again, the EA will close one position and open another when the MAs cross. This means that if the market starts trending against the EA’s current position, the MAs will soon catch up and enter into an opposite trade. This effectively limits the size of the losses. So the only way it really loses a lot is if the market enters an oscillation that has a frequency equal to the delay of the MAs.

Could you, please, explain the usage of such large Stop levels?
The EA takes care of itself: When the trade starts to go the wrong direction, the EA switches position, so there is really no need for stops. If the market suddenly and drastically changes, then it could have a bad result, but that’s money management that needs to be in place for live trading. For this competition, I assumed a bit more risk by expecting that there would be no very drastic and sudden changes in value. Even if there are some, there’s a 50% chance that the market will work for me and not against me!

None of your trades was closed by a stop order. Could you tell us how positions are closed?
Again, the EA will close one position and open another when the Moving Averages cross. The same signal that is used to open a position is also used to close the previous position.

Your orders have large StopLoss and TakeProfit levels – 1000 points (10 figures). Actually, we can say they do not exist, because at the moment a position in 3 lots with a StopLoss of 1000 points means losing about $26 100. Can the EA close a position, in case of danger? Or will it experience a Stop Out (a forced closing if the account is smaller than 50% of margin)?
There is no added safety to keep the account from a forced closing; you just have to trust that the EA will get the signal to enter an opposite trade and close the position in favor of a better one. A more realistic StopLoss is something I do use in live trading.

Your Expert Advisor uses the Stop & Reverse system and is always in the market. Visually it is clear that not all the potential profit is taken. Haven’t you tried to use a trailing stop?
I ran an optimization adjusting the trailing stop, but there was no benefit. It is possible that if I did a full optimization, re-optimizing all the variables including a trailing stop, it would be more profitable. There are so many variables already that if I add more, it will take an unacceptable amount of time to optimize the system. It’s a trade off; do I want to spend more time optimizing other pairs and other ideas, or try to improve a system that is already very good?

You have written about using moving average signals. You also use your own two indicators. Could you, please, tell us about them?
They are both types of Moving Averages, custom indicators that react better to the market. Better than a standard moving average.

Why have you chosen this particular currency pair?
This pair gave me the highest profit in back testing. A friend of mine and his roommate uploaded a similar system on different currency pairs, but they were banned from the competition on suspicion that it was one person uploading two different EAs. It’s really unfortunate, too. They would easily be in the top 5. The system works on several different pairs, you just need to tweak the variables a bit based on back testing and optimization.

At the current moment the Z-account of your EA is 1.85, which quite precisely points at the negative dependence between trades (though only 15 trades are closed and it is hard to speak about dependences yet). It means that after taking profit the next trade is more likely to be a losing one. Didn’t you think about using this fact somehow? Perhaps, the dependence actually exists? What did testing before the Championship show?
Testing before the Championship did not show any relation between consecutive trades. It is just the way the EA is reacting to the market right now. I will probably look into it for a future addition to the code if this negative dependency persists.