QUANTITATIVE TRADING STRATEGIES

THE SCIENTIFIC BENT AND QUANTITATIVE TRADING
You might ask: With numerous books written about trading systems and methods
available and more coming out each month, why read this particular book? I’d answer
6 PART 1 Structural Foundations for Improving Technical Trading
that Quantitative Trading Strategies is unique because I bring quantitative analysis
into the mainstream by presenting concepts in a realistic and logical manner.
While most books promote a specific trading method, they often fail to produce
historical track records of their ideas or a background of other trading methods.
In this book, I will take old and new trading ideas and test them on a wide
portfolio of markets. While other books specifically focus on stocks or futures,
this book will apply quantitative trading strategies to all markets.
We will apply techniques to futures, stocks, and some new markets that
readers may not be familiar with. In addition, we’ll test the historical performance
of both current popular systems and some new ideas I have formulated over the
past 15 years of trading. These tests will be run on 29 commodities, 34 stocks and
stock indices, and 30 relative value markets on the past 12 years of daily price
data. Historical performance will be examined from multiple angles.
Further, I will illustrate how readers can recreate my results and create, test,
and evaluate trading systems on their own. In addition, I will outline both the benefits
and the limitations of quantitative analysis by analyzing many of the tools I
use as a trader. And, drawing on personal experience, I’ll also illustrate certain
points by drawing on anecdotes from my trading career.
While it’s important to illustrate the profitability of quantitative trading
methods, it’s equally important to discuss the method’s limitations. No traders
make money every day. Very few make money every month. Some strategies that
performed profitably in the past will break down and become unprofitable in the
future. Trading with quantitative strategies involves much risk—risk that we hope
to limit by using state of the art techniques to design, test, and trade our trading
methods.
Readers will notice that I continually refer to the process of using fixed rules
to trade markets based on previous price history as quantitative trading, rather than
the popular term, “technical analysis,” typically used in the industry. The reason for
this distinction has to do with the quality of analysis. I admit to having disdain for
technical analysts who use charts to explain past price action. An example is to
draw trendlines, or lines that connect market tops or bottoms (see Figure 1.2). The
theory is that the extension of these lines will act as support or as a resistant in the
market’s future moves. You will often hear statements such as the following from
more traditional technical analysts:
Statement 1: “The S&P 500 has been selling off due to a break of the
six month trendline at 1100.”
The preceding statement provides very little predictive value in the trading
process. Attempting to reconcile past market action using technical analysis is
nonsensical. Markets decline due to news and information. Poor corporate earnings,
worries over corporate accounting practices, excess crop supply, and lack of
end-user demand for products are just a few of the many possible reasons for a
market to decline. When explaining history, we can usually create a clear picture
CHAPTER 1 Introduction to Quantitative Trading 7
of the market factors that caused rallies and declines. History and hindsight are
always 20/20.
While I believe using technical analysis and chart reading to explain past
market behavior is foolish, technical analysis can help in predicting future market
moves. Consider the usefulness of the following statement:
Statement 2: “A break of the six month trendline may bring about extra
sellers into the market and drive prices lower.”
This statement has merit and can be used by traders. Because the market is
breaking below previous support, we are likely to see lower prices in the near term.
Therefore, we should sell long positions and establish short positions. Skillful
technical analysts will make accurate market calls based predominately on price
action and leave the explanation of historical market moves to the fundamental
analysts dissecting news and new information.
While the second statement above may be useful to traders, we can take the
process one step further by incorporating historical performance. After all, are we
sure that breaks of trendlines are a precursor to lower prices? How often in the past
has this strategy worked? Consider the following statement, which suggests that we
take action based on a particular price formation—the crossing of a moving average:
Statement 3: “Because the market crossed below its 200-day moving
average, we expect prices will continue their decline.”
8 PART 1 Structural Foundations for Improving Technical Trading
Date
S&P 500 Price Graph
700
800
900
1000
1100
1200
1300
8/1/01
9/1/01
10/1/01
11/1/01
12/1/01
1/1/02
2/1/02
3/1/02
4/1/02
5/1/02
6/1/02
7/1/02
Break of trendline signals lower prices
F I G U R E 1 . 2
S&P 500 Price Graph. Once prices broke a trendline connecting September and February lows, prices
headed much lower.
In this case, the technical analyst is predicting lower prices due to price
closing below its average of the past 200 days. The 200-day moving average is
frequently used in market timing, and the above example is commonly used in
practice. While Statement 3 does involve a forward-looking prediction, we can
add more value to the trading forecast. For example, if we followed the fixedrule-
trading strategy of buying when a market rose above its 200-day moving
average and selling when the market fell below its 200-day moving average,
would we beat a buy-and-hold strategy? How much incremental return did an
investor make by following the 200-day moving average rule over the past 5 or
10 years? The crossing of the 200-day moving average is a market prophecy that
has existed for years. But does it stand up to statistics and historical testing?
In this book, we will attempt to solve the two problems cited above. First,
unlike Statement 1, all of our trading analysis will be geared for future trades—not
to explain previous price action. Second, unlike Statement 3, when we suggest using
a trading strategy that generates buy and sell signals, we’ll test that strategy over
many differing markets, each comprising multiple years of data. These results will
be scrutinized to separate promising ideas from those fated to be unprofitable. After
all, if an idea has not been profitable in the past, why should we use it in the future?