(Back)Testing Popular Trading Strategies

Welcome to my thread on backtesting popular mechanical trading strategies.
 
Even though I am doing this to test our new MT4 platform software, fxRulesEngine, I hope the results will be useful more generally for all traders. All strategies are can be traded manually with no dependence on our software. Testing your trading ideas is absolutely essential before putting your money on the line. Forex trading is a risky business and being unprepared is the surest way to lose your bank balance.
 
I shall discuss some of the underlying indicators used in strategies; post results for EURUSD over Jan to Oct 2016 (plus sometimes GBPUSD and others as comparisons) with different settings; and Excel compatible text files of buying and selling decisions made over the test period. Finally I  shall upload the fxRulesEngine configuration file for those who want to follow along with the software. Therefore I am attaching a prerelease copy of fxRulesEngine (for MT4 platform only) with documentation. The software can used for forward testing on any demo account without restrictions. It is limited until Feb 2017, but by then there will an unrestricted demo version available.
 
All my comments will be of a general educational nature only, and in no way are any sort of trading advice. Please be aware of the high risks associated with Forex trading and seek professional financial advice if need.
 
If you want to install fxRulesEngine, then please refer to the Quick Start Guide attached. Sorry, all documentation is still in development at this stage.
 
From the next posting I shall discuss and test some moving average strategies. See you soon. Alex
 

Moving averages (MA) are the most popular way of analyzing past Forex pricing. They can be calculated different ways in most Forex trading platforms. Simple MAs, my interest here, are calculated by "averaging" prices, often closing prices, over specific number of periods. Traders use MAs in different ways, such as to visualize trends or to show support and resistance areas.
 
So a simple trading plan, reacting to the underlying trend, could be (1) buy when the closing price of the last period is above the SMA calculation (indicating an upward trend) and (2) sell when the price closes below the SMA calculation (downward trend). So let's find how well it performs, even though the plan seems too simple to be profitable.
 
My first backtesting covered SMAs for 20 and 50 periods for one day and for 4 hour timeframes. In one instance I let orders close from stop losses and take profit settings and then in another instance with forced closures from opposite signals. As we might have predicted the losses were extensive with the various combinations. I included results as examples of using market price above or below SMA calculation (EURUSD SMA50 1 day and GBPUSD20 4 hour). My late testing returned a profit when I ran fxRulesEngine on a daily period for the GBPUSD testing. You can do that in the configuration file, but I did not want to complicate the uploaded file.
 
Next postings I shall look at three different ways of enhancing the SMA usefulness: multiple SMAs including multiple timeframes; price cross-overs; and SMA cross-overs. Also I later want to test popular moving average strategies like 3 Little Pigs, 3 Ducks Trading System and other MA strategies at ForexFactory. So SMAs and EMAs are commonly part of many successful trading plans. Getting the right combination does require some persistence.
 
The attached results are:

1. EURUSD: SMA50 1 day with 5 minute updates (loss $241.21)
2. GBPUSD: SMA20 4H (forced order closes on reversal) with daily updates (profit $555.28)

See you soon. Alex

Trade with 90% winrate / ghostbiker's system
Donald Trump wants to buy this system.

Previously I looked at the simple price above or below an SMA to trigger a buy (if above) or a sell (if below). In many settings this simple strategy lost money. I did post one profitable combination. In this post let's look at using two SMA comparisons. This is the basis for popular strategies like 3 Little Pigs. Traders often refer to fast and slow SMAs. Fast SMAs are typically 5 or 10 periods and slow 50 periods or more. By checking the price is above or below both your plan trades with a trend in a trending market. Let's see this works in practice. EURUSD returned good results whereas I returned losses with GBPUSD over this year.
Â
EURUSD 4 Jan-29 Oct 2016 checked daily - profit $846.49
Â

1. If Price closed above SMA 10 and 50 4 hour timeframe then buy
2. If Price closed below SMA 10 and 50 4 hour timeframe then sell

Â
I then added a trailing stop based on ATR to increase profit $993.90
Â
GBPUSD 4 Jan-29 Oct 2016 : all situations I checked returned losses.
Â
In my next post I'll test the well-known trend trading strategy 3 Little Pigs (http://forums.babypips.com/free-fore...ng-system.html) Alex

interesting idea... following. Hope you find the holy grail!

HA HA that's the problem.
Â
I see many traders, often losing lots of money (unfortunately that's the statistics) or ultimately never trading live, obsessed with finding THE trading method. Forums are filled with desperate questions about very specific indicator settings (as if that makes a real difference any way) or wanting links to the latest magic indicator or EA. Which they don't understand by the way. I know. i went through a lot of this. I see "hey this a great system I made two successful trades or I tested this over the last week...". "I did make some losses but that was my fault." You have got to be joking.
Â
Can you make money? Yes but it's hard work - slow and steady. You need to be cold and calculated (with some thick skin to withstand the inevitable losses). Thorough testing is an important part of any business strategy. That's why I wrote fxRulesEngine to help with that mindset.
Â
I'm a professional software developer with many years of building small, medium, and large systems for all sorts of business clients. So instead of creating the magic EA to work silently making money (doomed idea I believe), I wrote software to make me a better trader. It has and now I'll be releasing V2 commercially.
Â
Here are some ideas I've learnt so far...

1. Have a very solid understanding of risk management and use it to control every trade. it's all about capital preservation. I restrict order risk through stop losses to 1% or less of my available margin.
2. Restrict yourself to a limited number of currency pairs and get to know the fundamentals behind them
3. Work with simple analytical methods to build a limited selection of simple trading plans that suit your pairs for different trading trading conditions (volatile/non-volatile, trending/ranging and so on). If you can't explain the indicator and how it works to a non-trader then don't use it until you can.
4. Test ideas on past data before committing to live trading. At a minimum, question advice of the experts (may be even ignore in many cases).
5. Develop plans that work for your lifestyle - day (real) job, family, socialising. For many that means once a day or weekly reviewing of your plan performances.

There are more. You know the funny thing. I started testing fxRulesEngine with single indicator trading plans (combined with good risk built into the software) and many made money over the longer term. For example simply using CCI as a price reversal strategy would have made money EURUSD since Jan 2016. Unfortunately very good months were whittled away by poor months. That's why I built an option to stop the plan on major reversals after reaching preset profit targets.
Â
Sorry, I'll get off the soap box...€‹

3 Little Pigs Trend Trading Strategy
 
Popularized by BabyPips (http://forums.babypips.com/free-fore...ng-system.html), 3LP uses three simple moving averages (4 hourly, 1 day, and 1 week) to identify a trend to trade. My take on the buying rules are open a buy order when:

    1. Price is above the 55 SMA on the Weekly timeframe
    2. Price is above the 21 SMA on the Daily timeframe
    3. Price crosses above the 34 SMA on the 4 Hour timeframe

The selling rules are opposite. In my testing I set stop losses and take profits with ATR calculations. I tested three currency pairs from Jan to Oct 2016 - EURUSD, GBPUSD, and AUDUSD.
The first two image shows the equivalent in fxRulesEngine. The first simply checks if price is above and the second checks for cross-over. There are two rules. However, as you can see, a rule can be split over many lines if preceded by a + symbol. The other images are my testing results.
And, frankly, they were disappointing. I was unable to return a profit on any combination I tried. Some months were profitable but most were not. Had I enabled a special feature in fxRulesEngine I would have gained around $400 profit after two weeks in January on one currency pair; however that was lost in later trading.
GBPUSD had the best result, which not surprising shows the strongest trend (downward). THe last image shows as specific selling decision that showed a loss. The order had a stop loss of 141 pips.
 
I'm happy to look at a different currency pair, if someone gets better results than this. I've included on fxRulesEngine trading plan for those who want to follow along.

Mentioned in another thread...I recommend reading:  http://www.zerohedge.com/news/2016-11-12/trading-living

i see you guys are doing manual back tests.... is there a way to que up lots of ea's and have them back tested on different currencies etc? so essentially i write hundreds of strategies and variants and the software back tests and saves results for later analysis?