Most traditional trading models use historical price data series as the only data inputs for the rules that decide whether to buy or sell. However, you can build better systems by looking beyond this singular way of thinking. By designing and collecting your own measures of market breadth, you can integrate these data into a trading system and get a much better picture of market dynamics, particularly with markets having many components. Here, we will use an example trading system to demonstrate the usefulness of designed breadth data. Looking at the E-mini S&P 500 and its related component stocks, we can write rules that take advantage of this new information. First, though, we need to collect the right data.
MEASURING MARKET BREADTH
Market breadth analysis in general refers to technical indicators dedicated to the study of the overall behavior of the stock market based on statistics that are drawn from all the traded stocks within the stock market. We have all heard about advance/decline issues, advance/decline volume, 52-week new highs and new lows. Those are market breadth data of the stock market as a whole. The good news is these breadth data are available through the stock exchanges in real-time. You can easily access them so you can gauge the overall strength of the stock market.
However, if you are a trader whose primary interest is trading a specific stock market index, then the broad based market breadth data may not provide the specific information you need for the best analysis. A stock market index has only a small set of symbols compared to the thousands of symbols that are traded within the stock market as a whole. The breadth data for the entire equity market may not accurately reflect what the index components are doing. The solution is to create your own custom designed breadth data indicators.
ROLLING YOUR OWN
The idea is to design our own breadth measurements as opposed to looking at the classic ones that have been handed to us decades ago. As long as we can define the measurement and apply it across all the components, we will get our own designed breadth data. Because our breadth data will measure very specific aspects of the components, we should be able to uncover new information about the market we are trading.
One measure of market breadth is the count of three-day advance issues. The three-day advance issues number is simply the number of stocks that have a current price higher than their close from three days ago.
By counting all 500 component stocks of S&P 500 that meet this criterion, we get a customized three-day advance issues data series. This information can be collected in real-time and can be used alongside any other data you received in real-time. “Who’s rising?” (below) shows an example of the three-day advance issues line for the S&P 500.
Due to how the custom three-day advance issues indicator is designed, there is essentially no need of a three-day decline issues counterpart. That’s because unlike in the broader stock market, there are exactly 500 symbols that we care about. For example, ignoring the minor if inconsequential effect of stocks that are flat, if the value of the three-day advance issues is 300 today, then the corresponding value of the three-day decline issues must be 200.
SYSTEM UTILIZATION
Because the three-day advance issues data series is range bound (from zero to 500), we can assume that it behaves like an oscillator. One approach to investigate the usefulness of this market breadth data is to implement a simple overbought/oversold system using the data series to see if it has any significance in terms of price prediction on the E-mini S&P 500 index futures.
An overbought condition can be defined as high values. In other words, there are a large number of individual stocks trading above their close from three days ago. For an oversold condition, there are a large number of individual stocks trading below their close from three days ago. The assumption in both cases is that the extreme nature of either condition will not be sustained; and when it breaks, the market will tend to reverse direction. Of course, defining what’s extreme can be a moving target. This is an acknowledged problem with simple overbought/oversold systems, but not one that should subtract from its usefulness.
Here are the rules and specifications of the system:
Always in the market
Regular trading hours only
130-minute bars are used (this removes the last 15 minutes of E-mini S&P trading)
SPD3A = three-day advance issues
SMA = eight-period simple moving average on SPD3A
Long when SPD3A > SMA and SPD3A crosses above 220
Short when SPD3A < SMA and SPD3A crosses below 360
In plain English, the system always stays in the market with either a long or short position. The system goes long when the three-day advance issues line is rising from an oversold condition while sitting above its trend. The system goes short when the three-day advance issues line is dropping from an overbought condition while staying below its trend.
The reason why the last 15 minutes of regular trading hours in the E-mini S&P 500 must be discarded is because there are no breadth data for that time period. Breadth data only can be collected while all the components (the individual stocks that make up the S&P 500) are actively trading. This is a limitation of breadth data
in general.
“Rules for trading” (above) shows the rules for the system written as NeoTicker formulas (these simple rules could easily be ported into any reasonably flexible trading system software that allows custom systems). “Playing with breadth” (below) graphically depicts how the system goes long and short with relation to the three-day advance issues line and its average.
PERFORMANCE
The performance of the system is interesting in many ways (see “Profits without price,”).
First, the S&P 500 has moved up about 400 points during the testing period, and that translates to about $20,000 profit for a buy-and-hold strategy executed in the E-mini S&P 500. The system presented here has made more than twice that, demonstrating that the approach has some merit and is worth further investigation.
Second, the short side of the system is making about the same amount of money as buy-and-hold strategy. In such a strong bull market, this implies that the system does not overly rely on Beta (simple exposure to the market) for its profits.
However, there are always issues before a basic model can be deployed in real life. For example, the system does not employ any money management rules to protect itself from market shocks; thus, large drawdowns are expected. If you are going to build a trading system around this concept — or any concept, for that matter—you must work out the money-management and risk-control details first.
Finally, the decision making of the system is independent from the price series itself, making it possible to be used as a timing tool for other instruments that are highly correlated to the S&P 500 index.
DATA COLLECTION
There are several ways to collect your own market breadth data in real time. Some real-time trading platforms have the ability to construct synthetic symbols based on a combination of multiple symbols and their indicator values. If your platform has this feature, you can define your custom breadth data directly. Some platforms even allow you to reconstruct historical data for these synthetic symbols, making it easy to conduct research on your own customized breadth design.
For those platforms that do not support such functionality, you can accomplish something similar through writing a custom indicator that works on multiple data series within a chart. The disadvantage with this approach is that you cannot save up the historical data when there is a need to change components.
While collecting your designed breadth data in real-time, you must remember to maintain an updated component lists. With only a few issues switched in and out of an index within a month, there is no big impact on the usefulness of the data. Thus, there is no need to update the component list every day. In general, if you can keep up with the component changes once a month, the data you have collected will remain accurate.
However you go about collecting your data, this simple three-day advance issues system is just an example of how you can design your own breadth measurements to improve how you trade a stock index. It demonstrates the usefulness of simply having an improved version of the classic breadth measurements. By taking time to create more sophisticated breadth data yourself, you will be able to uncover more useful information about the indexes you trade.
Lawrence Chan is a trader and principal designer of NeoTicker. Lawrence@tickquest.com. Code for “Rules for trading,” will be available at futuresmag.com, under downloads.