Hi

Want to take the opportunity to present one of my systems, the WealthBuilder, and how it is created and is evolving.

It is a 100% automated system that runs on a VPS, and the basis for the system is;
*The main trading principle is mean reversion, so it will enter a trade if there is a quick and rather major pullback from the existing longer trend.
*SL and TP are put when a trade is entered, so it is “safe” from that perspective if the VPS gets disconnected.
*NO MARTINGALE OR GRID.
-*It consists of approx 3 -4 EA (have evolved during the time), each as uncorrelated as possible to the other.
* 1 - 15min time frame for EURUSD and GBPUSD , the trading rules will allow other pairs to be added if (and only if) there EA trading are uncorrelated to the existing ones. I am doing evaluation of a new EA in my test account, will if verified with the model be included in December.
* I have used Amibroker for the development and evaluation, and everything is back tested on data 2000-01-01 -> 2020.01.01, 1min tick.
* The leverage is decided by running a MonteCarlo simulation on each EA, where I decided that the leverage for each EA shall correspond to the level where there is a 95% probability that the worst Drawdown not is bigger than 40%. 40% is of course quite a lot for larger amounts of money, but is on the other side needed if you want higher return. We have to remember that the risk for a given system always corresponds to the growth potential.
* On system level, whenall EA are combined, the maximum historical drawdown since 2000-01-01 is 42%, and I have a system stop loss (that is, the system will stop and the system will be discontinued) on 50% drawdown. I have some thoughts to increase this, by calculating a MonteCarlo simulation on system level, but that is on my to do list.

I track the system's progress and compare it to the theoretical performance each month, so I use the real 1min data from IC markets and import to Amibroker, where I do forward testing on the system. That is compared to MT4 trading results, to keep track of ;
Cost for swap
Commision
Slippage
System performance and ability to trade “correctly”

MT4 and Amibroker should have exactly the same performance if the system is perfect, but real life is not perfect… There are always minor things that are not the same, slippage for example. So what I do is to update the model in Amibroker with real average slippage / swap etc, to get a model that better reflects real-life trading, and by this also verify the system or make needed updates.

I have made one major modification after 6 months. I notice that some EA really underperformed in comparison with the model during nighttime, most likely due to the spread that exists during nighttime.

The system runs on IC markets ECN account (and unfortunately not possible to connect it to paybackfx or similar, so I pay way too much commission…. All new accounts have a lower commission as I use paybackfx), and the spreads are nearly always during daytime 0.0 pip (tracked each trade), and I have now updated it to prevent trading 00.00 - 02.59 IC Server time. I have further removed one EA that created a lot of trades but rather low growth, due to the commission it created.

New evaluation is planned at the end of November to see if the changes made some difference.

I know that the system has performed really well since 2000, and I used K-ratio as a parameter target to get a straight EQ curve during the 20 years the backtesting has been tested on. The longest drawdown period has been 3 years, so I am not so stressed over a period of bad performance. The most important factor for me now is to;
* Verify the model
* Updated the model so it better reflects real-life data.