Earik Beann – Mechanical Trading Systems

Discusses the pros and cons of system trading when compared to discretionary trading. Systems have some distinct advantages. Specifically, they avoid a lot of issues of stress and psychology that plague discretionary traders, they can implement techniques too complicated for humans to master, and they can easily be leveraged up to turn small accounts into incredibly massive ones. Best of all, they can do all of that without needing human interaction, which makes them ideal for people who have better things to do during the day than stare at price charts.

Good and bad trading systems both have very distinct ways of telling you whether or not they will continue to work when moving forward into the future. After working through this chapter, you will understand the basics of how to go about evaluating a system, and will gain experience by using these methods as we take apart and discuss good (and bad!) systems working forward through the book.

This system was developed all the way back in 1997, and was actively used not only in Earik’s personal accounts, but also in the accounts of his partners when he worked at the Board of Trade. In the years when it was actively traded, this system crushed the Bond market with accuracy approaching 90%. Although it is now an old system, and Bond futures have changed significantly in the last 16 years, the core ideas behind this method continue to be valid, and if accuracy is your cup of tea, you will be hard pressed to find anything that can hold a candle to this.

To give you a taste, here’s the system report from just one of the nineteen individual patterns that go into this method:

Notice the accuracy number. Over 91% of the trades were winners, with the worst losing streak ever being a set of two losses in a row. This system report has been run from 1990 through 2013, and this pattern continues to crank away as it has been for the last 16 years running on live data.

All systems can be made to be extremely accurate, and after reading through the William Tell chapter, you’ll know exactly how to build 70%, 80%, and even 90% accurate systems. The methods that result in this level of accuracy are universal, so if you already have a system you like, odds are you can transform it into a hyper-accurate system as well, simply by adding a few changes to the way you enter and exit your trades.

In this chapter, we’ll take a look at parameter migration, where an optimized parameter set works one year, but then fails the next. This phenomena of markets is the main reason why so many systems found for sale in the classified ads in the back of trading magazines fail to ever earn a dollar of profits when traded live. If you’ve ever purchased one of those methods, you know what I’m talking about!

Not only will we cover what the issues are that cause parameter migration, but we’ll also discuss the solution. To prove that our solution works, we’ll actually build a system that makes simple moving average crossovers profitable! Yes, moving average crossovers can be made to make money when traded in out-of-sample data, but only if you can solve the issue of parameter migration. Understanding this one core concept will save you from countless hours of unprofitable development work, and will allow you to build extremely robust systems that will not fall apart as they move forward into unseen data.

Mean reversion systems are a class of systems that, when implemented properly, are extremely robust and profitable. Once you understand the core concept, they are also very easy to design. Check out this one from the book, which can be run with only two lines of programming:

Sure, it’s a little wild, but it also made almost $750,000 hypothetically trading one contract of the SP! Think of this as the raw edge, which can be taken and developed further. That gets done in later chapters, and if you are looking for some ideas to base your own systems on, this is a great place to start. The system shown in the chart above is actually over 12 years old, and works just as well today as the day it was first built. These techniques don’t lose their effectiveness over time, which is why we like them.

This chapter busts some myths. One of the absolute worst things you can do to a mechanical system is to add dollar-based stops and targets to it. But for whatever reason, this mistake is made every single day by novice system developers and traders. There are better ways to protect your account against losses, and better techniques to use in your own trading. This also holds true for the discretionary traders. You may not realize it, but the very tools you are using to protect yourself against losses may in fact also be the reason why your trading isn’t generating any profits. Learn the truth about stops and targets, and start using the techniques that actually help you generate profits rather than losses.

Many in the Wave59 community may have some familiarity with Earik’s Lunar System, first presented at the 2011 PowerUser Conference. This chapter reviews Spherical Astrology, a new way to define planetary aspects, and discusses the functions within Wave59 that allow this technology to be implemented in trading systems. Unlike typical astro, we can backtest Spherical Astrology, and prove that it actually does provide a timing edge. What kind of edge? Check out the chart below:

If you’ve ever thought that the Moon and Sun might have some impact on markets, that equity curve proves that you were correct. This chapter walks through building an astro system from the ground up, all the way from the theoretical ideas behind Spherical Astrology, through implementing and fine tuning the approach on the Dow.

For those already familiar with this system (or already trading a version of it), you’ll be interested in some of the work done on Venus, which provides a starting point for developing something even more powerful than the core Lunar system.

This chapter discusses the machine learning algorithms available in Wave59, both in the current PRO2 platform, as well as the upcoming PRO3 platform. In addition to an overview of the technologies, we will also evaluate actual systems built using both Hives and the new Genetic Algorithm toolkit.

After Earik released the Unified Theory of Markets over a year ago, he went into a period of very intense development work which resulted in the creation of an artificial research assistant. Give your assistant various systems and system ideas, and using the power of artificial intelligence and machine learning, you can make extremely effective leaps in the creation of mechanical trading systems. How effective are these leaps? Check out the chart below, which details an ES trading system developed using this new technology:

In testing, this approach made almost $200k trading one contract of the ES since 2000. It wons over 60% of its trades and has never had a losing year. What does it use? UltraSmooth Momentum, one of the core technical indicators in Wave59. I’m sure many people are familiar with this indicator, but I doubt many would have been able to use it as skillfully as our machine learning program did. No need to perfect trying to read this tool anymore, just let this system take it and run.

This chapter presents four systems based on machine learning algorithms, two developed with Hive Technology, and two developed with the new tools available in PRO3. All four of these systems can be used in the most current editions of Wave59.

Some exit approaches, like most dollar-based stops, will take a good system and turn it into a consistent loser. Others can take a mediocre system and turn it into a rock star. This chapter discusses one of the most most amazing techniques in the course, an exit approach so powerful that it actually works when using completely random entry signals.

Check out the equity curve below:

This graph comes right out of the book, and shows what would have happened to a theoretical account had it bought and sold every single bar on the ES and implemented this exit system in every one of those trades. In other words, this is the result of taking every single possible trade available in the ES since 2000 using this exit approach. Since we are tracking both buying and selling every bar, trend and timing cancel each other out, and all we are left with is the result of how we manage those trades.

The practical result of doing this test is that we now have an exit method that can actually be used as a trading indicator all on its own, regardless of the entry signal used to enter the trade. Because of that, you can create an attractive system by FLIPPING A COIN for your entries. Seriously. Combine this with an entry signal that actually has a real timing edge, and you’ll have something special.

This chapter discusses a technique to create an adaptive approach based on using many systems together. By allowing the strength of one system to counteract the weakness of another, two systems working together can generate more profits with less risk than each was able to accomplish on its own. Done properly, this approach makes the overall result extremely drawdown resistant, and robust enough to be almost immune from issues of parameter migration, curve-fitting, and a host of other problems that plague lesser systems.

By pulling together everything done in previous chapters, we are able to arrive at a base system that has an incredibly smooth equity curve as shown above. This system generates $13k per ES contract per year, on average, with over 60% accuracy, very low drawdown and very high reliability. If you do nothing else but read this chapter and implement this system, you’ll have an awesome ES method.

In order to complete a trade, we must know three things: when to enter, when to exit, and how much to trade. The majority of retail traders focus their energy on the least important of these three elements, the entry. Our random exit proved we can make money without even needing an entry signal, and this chapter shows how to take a profitable system and leverage it up into a fortune. There are various position sizing approaches floating around the industry, and in this chapter Earik shares the one he uses himself, which is a modified formula that results in extremely fast account growth while keeping drawdowns reasonable.

This equity curve is from the exact same merged system as shown before, with one exception – we have implemented position sizing. Give the first system one contract and 13 years, and it generated almost $200 thousand in profits. Give the second system one contract and 13 years, and it cranked out $200 million. Same signals, same amount of effort, same starting account size. If you are serious about trading, then you must implement position sizing appropriately.

This chapter follows in the steps of the last, detailing a way to use options rather than outright stock purchases in order to gain massive leverage on a stock-based trading system. Futures trading can be done with huge leverage, but the initial required account size in order to trade futures is oftentimes quite large, especially for new traders. On the other hand, options contracts can be bought for only a couple hundred dollars, and provide a way to achieve similar leverage with only a fraction of the required account size.

Unfortunately, there are many ways to lose money trading, and even more if you are trading options. If you have a good stock system, and simply purchase as many options as you can for a fixed percentage of your account, you will see your account crumble, even though your system itself remains theoretically profitable. This chapter presents a position sizing formula for options trading that will be extremely useful when trading small accounts, based on a technique developed here at Wave59. Simply input your account size, various options pricing numbers, and your trading signal, and it will spit out exactly how many options contracts to purchase at what strike price in order to maximize your results.

The chart above shows the results of implementing this approach over the last two years. The blue line is a stock trading system (for the SPY), and shows the result of what would have happened using maximum margin trading a $20k initial account. You can see that doing this earned 150% in two years time, a great result. The red line shows what would have happened had we traded this approach using our options formula instead. Not only did we earn more than twice as much, we did so with less risk. Harnessing volatility is key, and when done right, options provide more leverage than any other investment vehicle on the planet. This chapter details the formula, and steps through how to implement it in your own trading.

Last, but not least, we have the appendix: almost 30 pages of Wave59 QScript source code, which will allow you to implement every system discussed in the book on your own. There are no secret black-box methods here. The logic to everything is completely revealed, and pre-programmed for you. Just enter the scripts into Wave59, apply them to a chart, and you’ll have a full working version of whatever system you are interested in. An auto-trading script is also provided, which will allow you to set Wave59 up to trade hands-free on its own, without needing any human input aside from making sure the Internet is on and the broker is connected.