NEURAL NETWORKS IS GAINING THE ATTENTION OF CHICAGO TRADERS

Trading Chicago Style – Chapter Thirteen

No Market’s an Island;
Mendelsohn’s “Method Behind the Madness”

Louis B. Mendelsohn, President and CEO of Market Technologies, work in neural networks is gaining the attention of Chicago Traders.

Neal:
How do you think financial markets are different from the last decade?

Louis:
During the past decade, the financial markets have undergone an irreversible global integration. Whereas previously isolated from one another, they now function within a highly interconnected context. Now it is necessary for serious futures traders to factor intermarket dynamics into their trading decision-making. Traders who continue to focus internally on only one market at a time (single-market analysis), oblivious to the intermarket effects that related markets have on that market are in my opinion are putting themselves at undue risk.

It’s no longer adequate to simply follow buy/sell signals generated from single-market trading systems. Such approaches, widely advocated in the 1980s and still popular today, fail to allow for individual differences in risk propensity, account capitalization, trading styles and objectives. I now believe that such approaches are too inflexible and rigid. Most importantly, they fail to account for decision-making inputs such as insight into market psychology, trading expertise, and rational judgment that only a human being can possess. Technical analysis results generated by today’s trading software should be viewed as decision support information and cannot substitute for sound decision-making itself. It is foolhardy to think that an intelligent trader would ever want to turn over to his computer software the responsibility for making trading decisions.

Neal:
That is controversial in light of system trading.

Louis:
Traders assume needless risk when they restrict their analysis to a single market’s past price history, no matter how much back-testing is performed or how many single-market indicators are examined. Now, I believe, the surest way to be successful on a consistent basis and to protect trading capital against large losses is to incorporate intermarket analysis into the trading decision-making process and for the trader himself to assume the active role of decision-maker.

Neal:
Traders could use moving averages, though, right?

Louis:
I believe most traders would agree that moving averages are an excellent tool for smoothing out short-term and random fluctuations in prices. However, traditional, single-market moving averages are a “lagging” indicator because they are slow to react at turning points. They typically get in and out of trades after a change in market direction has occurred, often by several days, giving back profits and often turning winning trades into losers. In addition, there are serious pitfalls involving “curve-fitting” when re-optimizing the sizes of the moving averages. Another problem involves the occurrence of false signals during sideways or non-trending markets.

Neal:
Tell me a little about your VantagePoint software.

Louis:
Comprised of five neural networks, VantagePoint overcomes these limitations. One network predicts tomorrow’s high, a second network predicts tomorrow’s low, and a third network predicts a “neural index” based on a three-day moving average forecast. The fourth network predicts a five-day moving average for two days in the future, while the fifth network predicts a ten-day moving average for four days in the future.

VantagePoint’s predictions are not based on a single-market approach that optimizes moving averages. Nor are its predictions linearly related to past single-market price activity. Instead, VantagePoint takes into consideration the past ten years data from nine related markets that nonlinearly affect the market being traded. The result is a powerful “leading” indicator, comprising two predictive moving average crossover oscillators which, when viewed along with the neural index, gives the trader a clear indication of what the market is expected to do over the next one to four trading days.

VantagePoint’s predicted intermarket information is detailed in an easy-to-read, one-page daily report, which is updated each day after the markets close. All you need to do is collect the open, high, low, close, volume and open interest data by modem for the ten markets (target market plus nine related inter-market) that comprise each VantagePoint program.

When the oscillators turn positive, VantagePoint expects the market to go up. Similarly, when the oscillators turn negative, VantagePoint expects the market to go down. Changes in the magnitude of each oscillator from day to day afford an early warning of an overbought or oversold condition and impending change in the strength or direction of the trend.

Depending on a trader’s account capitalization, risk propensity, and style, she can act on changes in one or both of the oscillators. She can either close out an open position if there is any indication of weakness, or only close out the position if the weakness exceeds a certain “threshold” amount, i.e., if one or both of the oscillators narrow by a minimum number of ticks.

The neural index is used to confirm these oscillators. It is based on a predicted three-day moving average of today’s, tomorrow’s and the following day’s closes. When the neural index is 1.00, VantagePoint expects the market to go up over the next two days. When the neural index is 0.00, VantagePoint expects the market to go down over the next two days. It presently makes these predictions with up to 78 percent accuracy.

A trader can also look at daily reports from related markets for additional confirmation. For instance, the Eurodollar, 5-year T-note and 10-year T-note daily reports offer additional insight for bond traders into what VantagePoint expects to happen in the interest rate complex. Similar relationships exist between various energy markets, stock indices, and currency markets that VantagePoint covers. The bottom line is that predictive intermarket information, based upon the pattern recognition capabilities of neural networks, offers traders a broader insight or “vantagepoint” (hence the name of the software) on the markets than can be realized by focusing solely upon the internal dynamics of each market alone.

Neal:
So how can we translate this concept into profits?

Louis:
Through financial forecasting incorporating intermarket analysis, traders can gain an anticipatory, not just a retrospective, vantagepoint on the markets. With single-market system testing, it’s easy to discern where the markets have been and to discover simulated trading strategies which may have worked on past data. But the real payoff is in being able to anticipate future market direction consistently so that you can act decisively and confidently when real money is on the line.

With randomness and unpredictable events inherent in the financial markets, no one, regardless of financial, intellectual and computational resources, will ever be able to make 100% accurate predictions. A maximum achievable level, in my estimation, is at best 80 to 85%. From a decision-making standpoint under conditions of uncertainty, even considerably less accuracy would offer an enormous competitive advantage over more traditional methods of analysis.

Neal:
In some ways intermarket analysis is much like spreading. Would you say you’re taking it to the next level?

Louis:
Based on my experience of more than 25 years’ involvement in the financial markets, and nearly as long as a commercial trading software developer, it is my opinion that intermarket analysis is on the brink of broadening the definition of technical analysis, just as system testing did in the early 1980s. The markets have changed drastically since then, and single-market trading approaches, although still popular, particularly among novice traders, now leave much to be desired. By incorporating predictive information based upon intermarket analysis into your trading plan, and by recognizing that success requires more than simply reading a trading signal off a computer printout or “eyeballing” price charts from related markets, you will become a more confident and effective trader in today’s complex world of futures trading.

Neal:
What advice would you give traders today?

Louis:
Avoid the herd instinct. If the oft-quoted industry loss statistics (that 90 to 95% of individual traders end up as losers) are even remotely accurate, then to succeed, you can’t just do what everyone else does. Otherwise, the most likely scenario is that you’ll end up just like them. You’ve got to think, analyze and act differently. After being involved with intermarket analysis for more than a decade, and with the world financial markets now more interconnected than ever, I believe that every trader needs to have at a minimum an awareness of what’s going on in related markets. This is a critical piece of the analytic puzzle that can no longer be ignored if you want to avoid becoming a trading casualty. But overly simplistic intermarket analysis can be fatal, such as assuming that bonds and stocks always trade inversely with one another on any given day or hour. What we’re talking about here is part science, part art. To be successful takes a lot of hard work. There is no easy fix. Traders who think so, who echo the grandiose claims too often made in this industry, are just kidding themselves – and will probably end up contributing to the industry loss statistics.

Neal:
Can you show me some practical examples of how traders would use the output from your VantagePoint software program?

Louis:
Sure. VantagePoint provides our clients with its predicted information in two ways. One is a graphical representation of the output in the form of a daily chart, and the other is a numerical representation of the output in a daily report. VantagePoint clients have the flexibility to look at this information either way, since some traders are more visually oriented, while others are more comfortable just looking at the numbers.

On the graph (see Figure 13-1), you can see the PTM (projected 10-day moving average) touching the TRNDM (actual 10-day moving average) at the bottom left. Each time those two lines converge and begin to cross, it is an indication that the market is about to change trend direction. When the two lines are moving together in an upward direction, VantagePoint is projecting to the market to continue an up-trend. On the flipside, as the two lines begin to converge and move to the downside, VantagePoint is expecting the market to trend down. This occurred on this graph in mid-October. The PTM and TRNDM crossed to the downside, indicating to VantagePoint clients a change in market direction.

The numerical daily report (see Figure 13-2) has three sections. The first section is the index, which will be a number between 0.00 and 1.00. This indicates when the market is expected to make either a top or a bottom.

The main point traders look for in the index is an increase or a decrease in value. For example, when the index switched over from a 0.00 to a 1.00, this was an indication that VantagePoint was forecasting a change in market direction to the upside. Since the index is never more than 1.00, once it reaches 1.00 and remains there, VantagePoint expects the market to continue in an upward trend. You’ll notice from 10/05/98 to 10/06/98, there was a decrease from 1.00 to 0.82. Again, that indicates an expected turn in market direction, this time to the downside. It dropped further to a 0.00 and remained at 0.00 for several days, indicating a downward trend. The other two sections of the report go hand-in-hand. The PTS Diff is the difference between the actual 5-day moving average and a projected 5-day moving average that VantagePoint has predicted through its intermarket analysis capabilities. The PTM Diff is the difference between the actual 10-day moving average and a projected 10-day moving average. If the numbers in the PTS Diff and PTM Diff are positive and increasing, this indicates that VantagePoint expects the market to trend up. If the numbers are decreasing (they can even become negative numbers), this indicates an expected change in trend direction to the downside.

In these examples, you can see that when the Index was showing positive upward strength, the PTS Diff and PTM Diff were also showing positive upward strength. When the Index was showing weakness, the PTS Diff and the PTM Diff were also showing weakness. When these three indicators are used in tandem with each other and are confirming each other, it increases the probability of a successful trade and gives the trader more confidence, since these indicators are independently derived through the use of several neural networks within VantagePoint.

In addition to the information above, VantagePoint predicts the next day’s high and low to assist in identifying entry and exit points and help with setting stops. All of the information generated by VantagePoint is generated each day by taking into consideration intermarket relationships that exist between various selected markets and the forecasted market. This broadened context has allowed our clients to expand their technical analysis perspective to incorporate these intermarket dynamics.

Neal:
Is there a Web site where readers can reach you?

Louis:
Yes, it’s http://www.profittaker.com.