A Guide to Creating A Successful Algorithmic Trading Strategy

Perry J. Kaufman

• Verlag: Wiley
• Jahr: 2016
• ISBN:
  9781119224754
• Auflage: 1
• Thema: Negocios y empresa
• Thema: Comercio

Turn insight into profit with guru guidance toward successful algorithmic trading

A Guide to Creating a Successful Algorithmic Trading Strategy provides the latest strategies from an industry guru to show you how to build your own system from the ground up. If you're looking to develop a successful career in algorithmic trading, this book has you covered from idea to execution as you learn to develop a trader's insight and turn it into profitable strategy. You'll discover your trading personality and use it as a jumping-off point to create the ideal algo system that works the way you work, so you can achieve your goals faster. Coverage includes learning to recognize opportunities and identify a sound premise, and detailed discussion on seasonal patterns, interest rate-based trends, volatility, weekly and monthly patterns, the 3-day cycle, and much more—with an emphasis on trading as the best teacher. By actually making trades, you concentrate your attention on the market, absorb the effects on your money, and quickly resolve problems that impact profits.

Algorithmic trading began as a "ridiculous" concept in the 1970s, then became an "unfair advantage" as it evolved into the lynchpin of a successful trading strategy. This book gives you the background you need to effectively reap the benefits of this important trading method.

• Navigate confusing markets
• Find the right trades and make them
• Build a successful algo trading system
• Turn insights into profitable strategies

Algorithmic trading strategies are everywhere, but they're not all equally valuable. It's far too easy to fall for something that worked brilliantly in the past, but with little hope of working in the future. A Guide to Creating a Successful Algorithmic Trading Strategy shows you how to choose the best, leave the rest, and make more money from your trades.

CHAPTER 1
A Brief Introduction: The Ground Rules

Everything should be made as simple as possible, but not simpler.

—Albert Einstein

If you haven’t heard it, a classic example of trading experience is the difference between a statistician and a trader. You flip a coin 99 times and it comes up heads each time. You ask the statistician, “What are the odds that it will come up heads next time?” The statistician answers, “50:50.” You ask the trader the same question and he answers, “100 percent.” Surprised, you ask the trader, “Why?” He responds, “Because it couldn’t possibly be a fair coin. The odds of getting 99 heads in a row are too high to have happened by chance.” Experience transforms theory into reality.

When I first started trading using automated systems in the early 1970s, the very idea was demeaned by professional traders as “ridiculous,” “the market just doesn’t work that way,” “you can’t make money if you don’t know the value of the stock.” Now that opinion seems to have been turned upside down. High frequency trading, the algorithmic trading system on steroids, has “an unfair advantage,” “it’s stealing money from the ordinary investor.” Times have changed, but attitudes have not.

MY OBJECTIVE

This is a no-frills book. It’s short because it just deals with the most important issues of developing a successful trading system and because you’re more likely to read it all. It is intended to be a painless lesson in reality, those critical steps that you learn over time, often by doing them wrong. For some, it will be a confirmation that you’re getting it right, and for others it may be an “Aha!” moment. It would be more responsible, and more scientific, to verify each of my conclusions yourself. But, if you’re like me, you readily accept ideas that are reasonable and seem right, and you choose to believe them. I’ve been uncertain at times, but I rarely regret a decision that comes from common sense.

Each chapter contains, in my opinion, the best way to deal with the various aspects of creating a trading system. All of the steps are important, and doing them incorrectly will show up later in your trading account, perhaps too late. It’s better to spend a little extra time up front to increase your chances of success later on.

THE GROUND RULES

Before we get immersed in the details, there are some important items to define and disclose.

First, we all have biases. They are often found in what we don’t say rather than what we actually state. I see that when I watch the political commentary on television and also in the evening news. Everyone seems to have an agenda.

My own biases are toward fully automatic trading, which include all the good and bad of it. I also like macrotrend systems, stock and various other arbitrage, and some pattern recognition, among other methods that I can’t remember right now. I don’t like systems with a lot of rules and I’m suspicious of systems that work on only one market. I’m going to try to balance my examples to show short-term and long-term systems, but the reality is that there will probably be more about long-term trend following, which I believe is used by more traders, especially at the early stages when they are dangling their collective toes in the water.

THE PROCESS

The process of developing a trading strategy involves eight well-defined steps, shown in Figure 1.1. These should be clear, except for the marks on the left that show “Change rules” and “Failed.” From the top down, the trading idea comes first, then you need to get all the data that will be used to validate your strategy. You must have a trading platform to test your idea, which could be as simple as Excel, or as sophisticated as TradeStation. You enter your rules using that platform.

FIGURE 1.1 The Development Process

You begin testing using in-sample data (more about this in Chapters 10 and 11) and evaluate the results. If they are not what you want, go back and change the rules, then retest them using the in-sample data. Continue to do this until you’re satisfied with the results. Then test the strategy on out-of-sample data, which you’ve held aside and not used. If it works, and we’ll explain what “works” means later, you’re ready to trade.

However, it doesn’t usually unfold that smoothly. When you use the out-of-sample data, the results are not as good as you expected. So you go back and change the rules. But now you no longer have the out-of-sample data that are needed to verify your work. Purists would say that, once you fail the out-of-sample test, then you need to discard the system and start with something new. But that never actually happens. We’re sure that the method still works, but we’ve overlooked something that now seems obvious, such as using volatility instead of fixed values for placing the stop-loss. Or, we should have added profit-taking.

There is a way out. When you’re all done testing and the results finally look good, you can paper trade. That way, you track the results with no money and see if they are close to your expectations. The problem is that it takes time and we’re usually impatient to make money—or in some cases, to lose money. Experience says that there are probably still changes to be made, so paper trading is your punishment for looking at the out-of-sample data more than once.

BASIC TRADING SYSTEMS

We’ll be using both trend-following and short-term systems as examples throughout this book, so a brief definition should be helpful.

The Trend System

A trend is most easily defined by a new high or a new low price, but a moving average is more common, so we’ll use a moving average as an example most often. There is some discussion of the difference between trending methods, primarily in Chapter 3. We’ll explain the most important features of a few different methods, but the differences aren’t as important as the way the trend concept works.

For a moving average system, we’ll define the buy-and-sell signals when the trend turns up and down, and not when the price penetrates the trendline. Once you have accepted the concept that the trendline gives you the trend direction (not a difficult concept to embrace), the trendline is the only important value. You can ignore the price for the purpose of generating buy-and-sell signals.

Short-Term Systems

While trend systems are mostly similar, short-term systems can be very different. Examples will focus on their common properties, such as volatility, risk, and costs, as well as frequently used rules, such as profit-taking and stops. Short-term trading systems can be more complex than trend following. We’ll discuss a number of pattern-based approaches as well as intraday breakouts.

With this brief background, let’s jump into the development of an algorithmic trading strategy.

CHAPTER 2
The Idea

Where do you get trading ideas? They’re everywhere, but they require your intuitive sense to recognize, and they must be easy to explain. You can’t find a successful trading system by combining indicators, time periods, patterns, and other techniques in a computer. You will find something that worked brilliantly in the past, but with little hope of it working in the future.

BEGIN AT THE BEGINNING

You must start with a sound premise. That could be:

• Trends based on interest rate policy set by the Fed.
• Seasonal patterns that exist in agriculture, airline stocks, vacation resorts, heating oil, and other stocks and commodities.
• Exploiting the difference between two similar stocks, such as two chip manufacturers, pharmaceutical companies, or home builders. That’s called stock arbitrage, or pairs trading.
• Buying or selling price volatility before an earnings report or after a price shock.
• Fading an upgrade announcement by a major firm (old news by the time it prints on the screen).
• Same-time-next-month patterns, when f...