Now 47 years old, Briese was born in Minneapolis, where he grew up. After two years at St. John's College, in 1968, he left to fly helicopters in the Army. This was at the height of the Vietnam War. As he explained it, "My father had been a prisoner of war in World War II. My grandfather had served in World War I. I just thought I ought to go and find out what it was all about. There wasn't anything I found out there which made me feel we shouldn't have been there."
When he left Vietnam, Briese was the highest decorated officer in the history of his unit. It took him a long time to get his feet back on the ground. He said, "it took three years to return to some sort of normalcy. When I got back I couldn't stay put in one place. I was constantly traveling. Looking back at it a few years later, I could see that for three years I really didn't have my wits about me."
Steve believes his experience as a combat helicopter pilot had a very positive effect on the rest of his life. "I operated under extreme stress and under those conditions became the best that there ever was at something. So it did build self-confidence. I have lived the rest of my life feeling if I was going to get it, that was when I was going to get it. I don't have any fear of death. I don't have any fear of anything else. When I took around, I see that as a big advantage over a tremendous number of people who fear all kinds of things."
He re-enrolled in college, but again left, this time to help at his father's structural steel fabrication plant. Together they built the business to the point where it became a tourist attraction for Japanese who wanted to come and copy the most advanced steel fabricating plant in the world. As he described it, "A dozen people showed up with notebooks and tape measures. Before we knew what was going on, they photographed and measured our entire plant in detail, every piece of equipment. That's how the Japanese became such an industrial power."
In 1976, he left that business and for awhile flew helicopters in Alaska. Then he managed an employment agency for a year. Through that, he found a perfect job and placed himself as national sales manager for a specialty contracting company. He worked there until 1982, when he started his own commercial roofing business. Over the next seven years he built that operation to 60 employees.
On his 40th birthday, he decided to go on to something else. He found other jobs for his key employees and just closed the business down to concentrate on publishing the commodity newsletter he had purchased the previous year and trading for his own account.
At that time, Steve was not a newcomer to the commodity markets. He had started trading in 1973 after reading Anyone Can Make a Million Dollars by Morton Schulman. "The best and most explosive investment in the entire book appeared to be commodities," he explained. "I went to a local firm to see if any of the brokers was trading commodities. Sure enough there was a guy there who took Larry William's hotline and had commodities on his screen all day. I opened a $3,000 account and bought three silver contracts at $2.97. It went straight up.
"I rode it all the way to $3.60. I thought it had gone far enough for a while. I got out and then decided I might as well make some money on the correction. I went short and was ahead for about a day before the market locked limit up for three days in a row. By the time it was over, I had lost my $3,000, and I owed the broker another $3,000."
Steve didn't give up, however. "After I paid the broker back, I went about creating another account. At that time (1974), I discovered the Commitments of Traders Report. I subscribed to that and looked for opportunities. I was searching for extremes to go against the small traders.
"The nice thing about the Commitments report was it only came out once a month. It might be several months before you'd see a situation where there was a big spread between the trading positions. When I found something that looked cheap on the charts and small traders were in the opposite position or had no position, then I'd buy. I didn't have any real system, but I began keeping my own price charts with open interest, Williams %R and the Commitments data.
"Typically, I would make some money and then lose it back. I was making enough though to stay interested. I soon realized that even if I lost money, if I'd done just the opposite, I would have made a lot of money. So I kept working on developing some kind of methodology to trade by.
"By 1983 you could buy a personal computer. You couldn't do much with it yet, but it wasn't too long before there was software available to do some things. I started programing in 1983 with the first portable that came out. It was a CP/M machine.
"In 1987, after the stock market crash I decided that the commercial construction business was not going to be booming. I started looking around. I'd been interested in commodities all this time and decided maybe I could make a living at it and do what I really wanted to do. "I resolved to develop a methodology to trade. I did some research into who had organized a Commitments of Traders data base and where I could get it. I came across Curtis Arnold who had been publishing the Commodity Insider newsletter but had just quit. He couldn't make a go of it.
"I negotiated to buy his research which consisted of his data base and his formula for the COT index. His formula was a stochastic-type oscillator based on the commercial positions. It made a lot of sense. I started my own newsletter in March, 1988, publishing net position charts and COT indexes. In fact, I talked to you at that time and used your mailing list to get started.
"I decided the newsletter should pay all my living expenses and overhead. Then I wouldn't be under any particular pressure to live off my trading. That's the way I organized it. Even with five kids, within six months I was able to live off the newsletter and trade on the side."
Steve used his programming skills to create a number of commercially successful commodity programs. "I had people sending me systems all the time, wanting me as a partner to sell their system.
"One guy kept faxing me his signals every day for six months. He finally convinced me he had something. I made a deal with him, and developed a software product called System Tracker. The theory was to take many publicly available systems and methods, keep track of their trading, and then get into and out of the markets a little before they did. We rented the program for $300 a month to 50 people, with the customer being able to apply up to three month's rent to the $10,000 purchase price. 49 bought it. We still have one person who continues to lease the program for $300 a month.
"That same system developer later invented an algorithm for trading currency cross rate spreads. Steve turned it into a software product they sold called Cross-Current. It traded the Japanese Yen against the three European currencies. It came out in August, 1992, and purchasers made "tons of money" the first year. Although the currencies died down somewhat since, the system is still a highly-rated performer. It is currently available from, coincidentally, Curtis Arnold.
Steve also marketed a stock index system by Roger Altman called Private $tock. He hasn't run a track record on it recently but says, "We have people who continue to trade it." He also created a unique charting program primarily for his own use called, Auto Pilot. He has recently reprogrammed it for Windows. It contains some indicators available nowhere else such as lglehart Percent New and Hurst Envelopes.
My interview on April 25, 1996, continued:
Can you summarize your trading approach with respect to entries, exits and stops?
I think the bottom line to my approach is that if you are trading a methodology, a technical system, the technical system should dictate where stops and entries go. The typical trader, however, probably can't afford to trade with the risk that most systems create by having a large distance between the entry and the exit.
The route most system developers have taken around this is to adjust the stop loss. It becomes a money stop. "I can't afford $1,000 risk on this, so I'm just going to put a money stop in at $500." I think that's entirely wrong. If you're going to take a position, you should be able to stay with the position until the market tells you through your technical analysis that you are wrong. If you get out ahead of that, you're not really following any kind of technical system. I continue to believe that using standard chart patterns is a very lucrative way to trade.
You're talking about classical chart patterns. Triangles, head and shoulders, that kind of thing?
Yes. Edwards and Magee. I also have a manuscript of the charting course which preceded that book. It was never published in book form. It was a trading course published by an uncle of Magee. I've gone back to look at what has worked and tried to reduce the chart patterns down to a manageable few. I have two that I look for.
Which are?
The first is a failure swing top or bottom, which you can consider to be a double bottom or slightly-higher second bottom.
Bruce Gould's 1-2-3.
Or the last half of a head and shoulders top or bottom. Or the last two peaks of a triple top or triple bottom. These are reversal-type patterns.
The other one is a coil. That would include any type of triangle which is a consolidation. It is typically a continuation pattern, but it can also be a reversal pattern.
The problem with these patterns is that to trade a breakout and put your stop on the other side of the pattern usually creates more risk than I want to take and that most traders should take based on their account size. The typical approach to this would be to trade the breakout when it comes, but to put your stop at a money risk distance from the entry.
I take the opposite approach. First, I determine where my stop loss should be based on the pattern. I want to put my stop at the point where the pattern fails. Then I determine how much I'm willing to risk on the trade. In my case I never risk more then $500 per contract on any trade. I don't trade the S&P, where that would be difficult.
Is it a matter of personal comfort?
Yes.
On a 1-2-3 or failure swing, would your stop go below the first low?
Yes.
Go on.
I've established my stop, and I know what the exit point will be. Now I have to find the entry point that will keep my risk within the required amount. This will usually be at a point before a breakout verifies the pattern, still within the pattern, I use other factors to forecast which way a pattern is going to complete.
You don't wait for a breakout and then try to buy a reaction. You try to anticipate the completion of the pattern.
That's right. I use the Commitments of Traders data to predict the direction a pattern will complete. I'll give you an example in the Bond market. [See the chart.] In the Christmas, 1995 issue of my newsletter, I described a Special Situation on the short side of interest rate futures. Based on the weekly chart I was looking for a double top in T-Bonds. It was a huge pattern. I don't care how rich you are, you couldn't afford to wait for a breakdown below the 1994 low and put a stop above the 1993 high.
That's a worse-case situation. Most people aren't going to be trading patterns on a weekly chart. I was looking for a double top on the weekly chart because the commercials were selling very heavily into the rally. I elected to put on a position based on the weekly chart pattern with a stop above the previous high. Of course, that one worked to perfection.
You had to be pretty lucky to enter within $500 of the high.
When you know where your stop is and you know what your risk is, you are forced to wait for an opportune entry. The market must come to you. In this case it worked out because I traded options on this trade. It was possible to do even with futures if you were prepared ahead of time and had an order in to sell if the price got high enough.
The high on the weekly chart was 122-04. Do you remember where the futures were when you did your options?
Yes, they were above 121.
That type of situation doesn't come along every day.
I'm giving an extreme example. We're still not even half way down to the 1994 low, which is the point where this pattern will be confirmed on the weekly chart. People short from the top have made a tremendous amount of money already.
I find that the one thing even people new to the markets are able to do is determine where they want to be out--the spot where they know they're going to be wrong--their stop loss point. Typically, I ask them to find that point. I tell them that's where their entry should be. If you are getting in when everybody else is exiting, the majority of the time you will probably have an optional entry.
That's a very intriguing idea. Where would you put your stop than? That becomes a matter of risk.
I base that on a percentage of the account size, which I think should be one or two percent on any one trade. For smaller traders it's going to be a bigger number, but certainly no more than five percent.
This sounds contradictory. You started out by saying you don't want to use a non-chart-based stop. You use a pattern or some method to say where you should exit. That's where you put your stop. Then you enter at a point which will keep your risk to the amount you want. Now you're saying forget all that and enter at your exit point with a money management stop.
I'm using an outside indicator, in this case the Commitments of Traders Report, to predict the direction of the next move. Typically the entry point is much closer to the stop loss point than people realize. Quite often commercials are able to move a market to collect all these ready stops and have a maximum position before the market goes their way. So I think the stop is the most important thing, and the best entry is close to that.
Your second strategy requires patience because you're going to miss some trades when the market never makes it to your entry. It also requires courage because the market is probably looking pretty bad where your stop point is and you've got to be stepping up to the plate. The key to your methodology is that no matter how bad the market looks, it's deceptive.
The market always looks the worst before it turns near your stop. Of course, I use the Commitments data to predict that a major turn is happening. Under this procedure you're going to find the optimum entry.
So you wouldn't recommend that type of procedure unless you were using the Commitments of Traders Report to establish direction. Would that be a fair statement?
Yes.
Would trying to use it in other ways be dangerous?
I think so. I'm also saying this entry is superior to a breakout entry because the breakout entry is going to be too far away from where the prescribed stop loss should be.
I don't want people to be deceived. You're definitely saying to use this only in conjunction with Commitments of Traders directional suggestions.
That's right.
I agree with you in the sense that buying weakness and selling strength is ordinarily not a good way to trade. But you're going to make an exception here for your Commitments methodology.
Yes.
Before we get to your Commitments strategy, we still have to cover exiting. I think you've done a pretty good job of describing two different ways to enter. Now, once you are in a winning position, how do you exit?
The exit is the real art of trading. There are a lot of people who can enter right, but the exit is very, very difficult. I prefer to use a trailing stop loss. What I look for is the point when speculators become heavily involved in the market. You get what I refer to as a speculative bubble. You can see it in the Commitments Report. At that time, I like to move my stop up behind the latest correction low or high. I also have a system that looks at volatility. When the volatility doubles from entry, I will take off half the position.
Volatility defined as what?
As the average daily range. I use the same time frame as I use for getting into the trade.
What would that be?
I use various time frames depending on the market. For instance, let's say I'm trading a breakout system in both the Deutschemark and Swiss Franc. Since they are so correlated, I use two different time frames. I don't want to trade the same market cycle in both markets. For instance, if you're trading a breakout above a so-many days high, I might look for a breakout to a new 25-day high in one market and a breakout to a new 100-day high in the other. I would trade each market differently. The typical system developer looks to optimize a system to only one time frame. Doing that, there would only be a few complexes in which you're really diversified. You might have a choice of as little as six markets to trade before starting to double up on the same market forces.
What about the danger of curve fitting when you optimize each market differently?
The systems I use typically are so robust that in terms of a breakout they might work from 20 days to 120 days.
Why are they so robust? Because they have very few parameters?
Sure. Number one, they have very few parameters. And number two, the range of values for those parameters is so broad that it really doesn't matter which one you pick. You can pick with a dart what parameter value you want to trade. The key to increasing the smoothness of your equity curve is that within a particular market sector you should be trading more than one cycle. That means using grossly different time frames. As I previously described, you could trade a 25-day breakout in one market and in a related market that tends to move the same, trade a 100-day time frame. Of course, you must have a system that is robust enough to trade those non-correlated time frames profitably.
That sounds like trying to combine long-term and short-term systems in the same portfolio. There are people who will trade one short-term system and in the some account trade long-term as well. That would amount to the same thing. Is that the idea you're talking about?
Yes. To get back to my exit plan when volatility doubles, in order not to have an additional parameter, I use the same time frame to determine the average daily range as I used for entry. You lose degrees of freedom every time you add a new parameter to a system. (That's a statistical measure of reliability.)
The increase in volatility tells me this market has probably gone as far as it's going to go on this particular leg of the move. For instance, recently I had a system that went long Crude Oil on a breakout. Volatility doubled right at the top. I took off one contract at that point.
Then you try to get back in some time later?
Yes. I try to get back in when the market reacts very close to my current stop loss, so I have more or less the original risk back in the trade. On a long trade I look for a big down day followed by a small inside day near the bottom of the first day's range. If that brings the risk down, I put the trade back on.
For the rest of your position, you're trailing your stop below the latest reaction low?
That's right.
You have some important ideas about money management. Can you tell us about those?
I determined with Cross Current [his currency spread trading system] that by using a pyramiding scheme which adds to winning trades, you could drastically improve the reward/risk ratio. With that system you could about double it. Most system testing software packages don't allow you to pyramid. Or to get out of a trade in stages either. That was critical to turning a so-so system into a very successful system.
Another breakthrough came in measuring system performance. The available optimization routines use standard performance yardsticks such as return on maximum drawdown or total profits divided by total losses. When you optimize a system's parameters, you are optimizing based on those benchmarks. Those standard ones are about the only choices. I found they were not very good.
In system trading what you should be trying to do is to find a systematic method of building an equity curve that is as close to a straight line with as steep an angle as possible. The biggest drawback with most systems is the lack of smoothness of that line. You want to minimize the movement around a straight line equity curve. I developed a statistical measure which did that. It doesn't do any good, however, unless you can actually optimize for it, which is why I created software to do it.
Can you tell us more specifics?
No. It's proprietary.
How do you choose the best markets to trade?
I once analyzed a highly-rated system for its author. He found his system was profitable in three dozen markets and from those he had various portfolios for various size accounts. He based his portfolio decisions on the total profits generated over a 5 or 20-year period. Using the standard system performance measures, which are the ones he had available to him, such as return on total drawdown or total profits to total losses, it seemed/perfectly natural. But if you looked at my proprietary performance measure, you could see right away that many of these markets should have been poorly rated even though they were highly rated using the standard measures. I looked a little closer at the results. Sure enough there were some grain markets that were very profitable for one year every f'ive or six years. But they made money so infrequently and had such drawdowns in between, that they subtracted from the smoothness of the equity curve. Their equity curves were so far from a straight line that they were not good markets to trade with that system.
What about specializing in just a few markets or one market?
I once talked to a fellow who every spring invested $75,000 in Soybean options. I told him his strategy was going to hit it big eventually, but it would probably be the year he didn't have $75,000 left to invest. One of the best things you can do is include as many markets as possible in your trading that have shown consistent profit-making potential year after year.
What markets would you recommend? What are the best markets?
We're talking here in terms of a trend-following system. The currencies historically have been tremendous year after year. Some of the international foods are good. Cocoa is a big exception. Most of the grains are not good. Even though they look like it on a chart, when you try to trade them with a trend-following system, you find they just don't perform very well. Their big trends are too erratic and far between. They may have trended well over time, but they don't perform consistently well year after year. Corn and Soybean Oil are the two best markets within the grains.
Can you tell me how many markets is a good core number to trade without specifying what they are necessarily. Your system author had 36 markets, which was too many. When you weeded those down, how many did you end up with?
Eighteen.
Are you familiar with my book Trendiness In the Futures Markets?
Yes.
I first published it in 1988, and I update it every year. It is an objective examination of how historically trendy the various markets have been.
I would put the currencies first and interest rates second.
Exactly. Coffee's up there, and Orange Juice. I agree with you 100 percent that Cocoa is not a good trending market. I don't trade Cocoa with systems. I trade 21 markets with my various trend following systems.
I love the Cocoa market, but it has particular characteristics that don't allow you to trade it with a trend-following system.
I've suggested over the years that picking the proper portfolio of markets is equally important with picking a good system. It sounds like you agree with me on that.
Yes. It's critical. The problem is that the system performance measures available to the individual trader are not satisfactory.
Let me run another idea by you I feel strongly about. Part of the problem with almost all testing software Is that it only allows you to test one market at a time. You can't judge what happens when you're trading a portfolio of markets together. This applies to drawdown especially. My system software allows you to test a whole portfolio at once and see what the joint drawdown is. Do you agree that portfolio testing is crucial?
Yes, and I do that as well. Your software has always done that, and I agree that it is critical. But it still doesn't give you the right answer if you don't have the right performance measure.
That is certainly food for thought even if you won't tell us what it is. Let's turn to your Commitments of Traders methodology. Why don't you start with the theory of why It works?
In much the same way some people use the Elliott Wave Theory or various technical measures to determine how far the market will move and when it will reverse, I view the Commitments data as a measure of crowd psychology. In terms of crowds, I can break commodity traders into three fairly homogeneous groups. These three crowds happen to line up with the breakdown of the Commitments of Traders Report. By measuring their relative bullishness or bearishness, I can determine the likely direction of a market. What I'm looking for is extremes in sentiment. As soon as you take a position, you have joined a crowd. You start to lose personal initiative and adopt the crowd mentality. Let's say you're watching CNBC. If an analyst agrees with you, he's a very smart guy. If he doesn't agree with you, he doesn't know what he's talking about. He's all wet. You tend to follow your crowd's leaders.
For those not familiar with this, would you describe the three different crowds we're talking about here?
The first crowd is commercial hedgers. They are members of one of two sub-crowds in that they are either producers or consumers. Producers would include such entities as mining companies, mutual funds and grain elevators. These entities hold cash positions in a commodity and sell their commodity forward in the futures market at current prices if they're expecting a price decline.
On the other side are consumer commercials like manufacturers and food processors who have a continuing need for a particular commodity. If they believe the price of the commodity is going to rise, they buy forward in the futures market to cover their future needs.
What these sub-groups have in common is that they trade on inside information and they generally trade in only one direction. Mining companies are always long the physical metals, so they always short in the futures market. Dog food companies always have an inventory of meat byproducts or soybean meal, and therefore, they always buy in the futures market.
They're a homogeneous group. They are in the market to reduce their risk from future price movements. They have an offsetting cash position. If they are losing in futures, they are winning in cash. This gives them much more staying power than the other groups. They have deep pockets to begin with. In order to belong in the large commercial category, they have huge cash inventories and huge futures positions. That makes commercials a homogeneous crowd.
The second crowd is large speculators. Some books back in the 1970s, including Larry Williams' book, How I Made One Million Dollars Last Year Trading Commodities, recommended that you follow this group of large traders. At that time large speculators usually got that way through trading profits. Today the large traders are primarily the commodity funds. They're willing to assume the risk from commercials in exchange for an opportunity to profit. They are typically motivated to show a profit every month, and they're dominated by computer-driven, trend-following approaches. Thus, they also tend to be a very homogeneous group. They became large by having a great sales department, not by trading prowess. So I don't follow them, but they are a very close-knit group.
The remaining large group is small traders. That includes both speculators and small commercial hedgers such as farmers and others who don't have large positions. What they have in common is that they have smaller pocketbooks than either of the other two crowds. They lack inside information and, therefore, they're more susceptible to news. This crowd tends to be bullish if they are long rather than long if they are bullish.
These little guys are probably the most susceptible to crowd psychology. Once they're in a position, they make undisciplined decisions that after the fact they would find illogical. While they're under pressure from their crowd's mentality, they go along for the ride. They are the last ones in and the last ones out.
I look at the Commitments data as a very accurate way to measure crowd psychology. From the work I have done, I have determined that in most markets, of the three, commercials are by far the best group to use as an indicator for most market conditions. The reason is they take positions for the long haul and typically defend price levels for long periods of time. I did a little research study a couple of years ago that looked at extreme readings for each of the three groups to determine whether price had moved in agreement with their position three months later or whether it had moved opposite. There are many ways to determine whether a group is right or wrong. This is just one, but it confirmed my suspicions. Commercials were right about two-thirds of the time. Small traders, who many people think are a good contra-indicator, were right 45 percent of the time. That's way too close to 50 percent to be useful as a contra-indicator. Large speculators were very close to small traders. They were right 46 percent of the time. Based on this, in most situations I look at the Commercial position.
Watching the commercials does one more thing for us. Nobody can be an expert in fundamentals in more than a few markets. Even being an expert in fundamentals doesn't mean you're going to make money trading that market. People who trade with fundamentals are typically commercials who don't need to be as accurate as the rest of us. They have an offsetting cash position which gives them more staying power. I look at the commercial position as a measure of the unanimity of opinion among commercial hedgers. When commercial hedgers get very one-sided in a market based on historical patterns, you have a consensus formed that can determine major turning points in the market.
What about the argument that hedgers have to hedge, and they just go in and do it. They are not that concerned about predicting market direction. They're just trying to hedge their cash needs.
That makes sense until you talk to the people who are actually doing the trading for these large firms. It turns out the trading operation in many of these firms is now a profit center. Their livelihood depends upon making a profit on their hedges. While they can afford to be wrong, they don't want to be wrong. Their traders who are putting on these positions, like any other trader, want to be right.
They have several advantages. The biggest is an informational edge. If you look at the fundamental information available to traders that doesn't come from the government, most of it originates from the very commercial houses that we're trading against. In the best-case scenario, you assume the commercial houses have already acted before they release their reports and prognostications. In the worse case, you might imagine that what they actually release is not complete or accurate. They are also better financed. Commercials have a huge informational and financial advantage. All my research has shown they do make money on their hedging.
Most of the time they aren't one-sided enough to give a signal. What percentage of the time does their trading get one-sided enough to say something?
I haven't worked out the percentage. But the times they are one-sided tend to be key reversal points. The very best signals we get, the most reliable, are when a market consolidates and reacts from an ongoing trend. Typically when we see commercial buying come in on a correction to an extreme, it is a highly reliable signal that the major trend is going to reassert itself.
We don't want to make this sound like too much of a sure thing. It doesn't always work. Do you want to talk about that a little bit?
Sure. Commercials under the study I did were only right two-thirds of the time. What you find is that when they are wrong, they can be very, very wrong. In his new book on technical analysis, Jack Schwager talks about signal failures being the most reliable of all chart signals. When we have a signal failure on a commercial signal (such as where Commercials are as bullish as they have ever been in a market and it starts heading south), I view that as a signal failure. Those can be the most explosive moves.
You anticipated my next question. What about using a reversal stop when you're trading one of these commercial signals?
That's the last part of Schwager's discussion. He referred to novice traders who ride their positions into a big loss. A more experienced trader will have a prudent stop loss to exit his position. But the really skilled trader will do a 180 and reverse his position to go with the move.
What I try to do in my work is not only pinpoint where commercials are aggressively one-sided in the market, but also key points where I anticipate they are going to be wrong, I've learned to identify certain patterns in the Commitments data that occur at major market turns.
Take the T-Bond market as an example. [See the accompanying chart.] in 1993, I had spotted commercials as very bullish. They held an extremely bullish position going into the spring of '94. Their buying had been concentrated at about the 11 3 level, which I then highlighted. "If the market breaks through this extreme commercial support at 113," I said, "that will be our signal failure point-the point to go short." It did break through, and that was the worst year on record for bond prices.
Interestingly enough, commercials came in heavily as buyers in October, 1994, triggering a minor buy signal. Their activity called both ends of the move in a different way. They were wrong at the top but right at the bottom of that move.
I was looking for this rebound to become a secondary selling opportunity. As prices came closer and closer to the previous peak just over 122, 1 started to bet on a double top. Commercials were the most bearish at Christmas time, just as prices threatened the previous 1993 top. In this case the commercials were exactly right. This is where you get a very high-confidence signal: a correction to an ongoing major trend. I viewed the major trend as having turned down in 1994. Commercials came in as sellers on the test of the previous high and were exactly right. It is interesting to note what happened across the interest rate spectrum after the market broke from the second top but before that 3 -point down day caused by the unemployment report. Commercials were buying with a vengeance across the interest rate spectrum. They bought enormous numbers of contracts.
I viewed this as another indicator of a major top because what typically happens at these major highs is commercials put in the top through their selling but then try to hold the market within a narrow range by buying after a slight decline. When they buy in huge amounts, that's an indication of panic. What it means is they have such huge positions in the cash market that they try to use the futures market, where they can leverage, to hold up the cash market. What you find in these cases is firms putting on what are called Texas hedges. They're long cash and long futures, too. In this case I was able, as I have been time and again, to spot this panic buying as a red flag and a red herring. It was an indication that we were going to have a substantial move against the commercials.
It sounds somewhat contradictory. You say that the commercials are mostly right and so you do what they do. But then you also go opposite to what they do. You really have to know the specifics of when to follow them and when not to.
That's why my newsletter's renewal rate is so high. The Commitments data is readily available from a variety of sources. The CFTC has even made it available on the Internet. Anybody can collect it for free and took at it. All the chart services publish net position charts that show what the positions of these trader groups are. But analyzing these charts is not a straight forward, straightline procedure. It's no trick to determine when a particular trading group is bullish. The real trick is being able to determine when they're going to be wrong. I'm not suggesting you can bat 1,000, but I am suggesting that when commercials are wrong, they tend to be wrong in such a big way that those are the moves we don't want to miss. That's why you should try to spot the key commercial support and resistance zones. If they are penetrated, you want to go with that move.
Let's go on to the procedure. Tell us how you produce what you print in your newsletter.
I started with a data base I bought from Curtis Arnold. He published a newsletter before I did. He had many of the markets going back to 1983. It took me a year and a half to fill in this data base for all the markets that have enough open interest to follow. The data is monthly from 1983 to November, 1990. In 1990, I started a letter campaign to petition the CFTC to release the data more promptly and more frequently. That actually brought results. Starting in October, 1990, we got reports on the first and fifteenth of the month that were released within three days. The data became bimonthly from October, 1990 until October, 1992. Then they went to reporting weekly figures every other week. They calculate the net positions every Tuesday and release two-weeks-worth every other Friday.
Much of the historical data out there is defective. There's a data source called Pinnacle Data which claims to be the sole source of Steve Briese's COT Index, as well as providing the Commitments of Traders Reports. What they have are Curtis Arnold's COT indexes. He used a fixed time frame and did not adjust for changes in open interest, which have been extreme in some markets. So Pinnacle does not have Steve Briese's COT Indexes.
The government released the historical data Pinnacle uses in tape form in 1990. It goes back to 1986 and appears as twice-a-month data, which was not what was reported at the time. It was not audited, and it is full of errors. The government knew this when they released it. Most of the available historical data, including Pinnacle's, include these faulty numbers.
My data is all from the original reports. For many years the government couldn't always add correctly, so I provided error checking to them. To make sure they add up, I have error checking built into the program I use to read the numbers. I also have all the corrections the government issued subsequent to the reports.
Do you sell your data independently?
I do.
How much is it?
$95 a year.
So if you want to go back ten years, it's $950?
No. I give away the entire historical data base going back to 1983 with a one-year update order. You get a year's worth of future updates plus all the history for $95.
The updates are what you yourself download every two weeks using a modern? Couldn't a person just get the data themselves on the Internet?
Unfortunately, the Internet files are structured like a printed report. They're not useful unless you are able to read the numbers into a chartable file. What I do is take the electronically printed report, pull out the key numbers and put them into comma-delimited ASCII files in a standard price file format. You can then graph them in most commodity charting packages. They're in open, high, low, close, volume, open interest fields. You can import them and create charts on your own computer.
Where do you get the data?
From Martin-Marietta, which somebody else just bought out. I pay a hundred and some dollars a month to collect the data.
Why do you pay so much when it's free on the Internet or from the CFTC?
I am programmed to process the file format that comes from Martin-Marietta. I did revise my program to conform to the government's new format when they went to the Internet, but then they changed the format again. I wasn't about to keep reprogramming every time they change. So I stayed with Martin Marietta. Also, the Internet can be quite slow. Collecting that file has taken me up to 45 minutes for something that should take a minute. I need the data right away, so I pay for it.
If somebody subscribes to, say, Commodity Trend Service's chart service with the data in there, is there anything wrong with looking at their charts?
No, that's the easiest way to do it. Then it's a question of interpreting the charts. But you offer quite a bit more than just creating charts of the raw data. What I do in processing the data is create what I call the Briese Percent COT (Commitments of Traders) Index which takes the commercial position and compares it to an historical benchmark. The large commercial position is reported as so many longs and so many shorts. I subtract the shorts from the longs to arrive at a net position. I then compare that current net to the "maximum" and "minimum" net position historically. Those maximums and minimums are the most bullish and the least bullish positions over a specified look-back period. Commercials move from net long to net short in most markets. All I'm doing is creating an index which shows the current position's relationship to an historical period. How relatively bullish or bearish are the commercials currently?
The period you use is different for each market, and it's proprietary.
That's right.
The difference between what you do and what Curtis Arnold used to do is that he used the same period for all markets.
Correct. I have made one other modification which has become necessary over the years. There have been huge changes in the open interest in many of these markets. I have a proprietary open interest adjustment I use.
I don't understand.
Because we're measuring changes in the difference between longs and shorts, if the open interest increases substantially in a market because there is more active hedging from a particular group of suppliers or producers, then the net positions might not be from the same group you've been following historically. As the open interest rises, the net positions you see charted today may not be the same trader mix as comprised the previous data points on your chart. For instance, Futures Charts prints four years of data. As you see the open interest rise quite dramatically in some of the markets, the mix of traders changes. I found a way to adjust for that, so I get a more accurate ongoing reading of relative bullishness and bearishness compared to historical levels.
The percentage numbers you print in your newsletter are percentage bullish. The higher the percentage, the more bullish the commercials are in a particular market.
That's right. When I show a 100 percent COT Index reading, that means the commercial net position, the longs minus the shorts, is equal to or greater than the highest it has been over the look-back period. A zero percent reading means the opposite. That would be their most bearish position. The reason I use this index is that the mix of traders varies from market to market. You have to be able to read the net position pattern you see by plotting the net position of Commercials, small traders and large speculators for each particular market. Patterns in one market will not work in another market. For instance, in some markets Commercials are almost never net short. Yet in Silver, commercials have never been net long. The reason they have never been net long is there are more producers (mining companies) in the futures market than there are manufacturers who require silver. So the net hedge in Silver futures has always favored the short side. If you're looking at a Silver chart versus a Gold chart, in Gold you'll see that commercials readily move from net long to net short, whereas in Silver they never get net long. It's a completely different pattern. By using my oscillator, the COT Index, we can come up with a uniform scale for all markets.
So the table you have on the back of your newsletter is quite different from the numbers that you would see coming from the CFTC or charted in the Commodity Trend Service chart book?
Yes.
And in your opinion your numbers are a lot better to use.
If you're following two or three dozen markets and you have to assimilate the net position patterns in each one differently, that's a difficult chore. By simply changing it to a uniform oscillator so that zero and one hundred mean the same thing for every market, it's much simpler to employ.
When someone gets your data updates for $95 a year, does It include the COT Index numbers as well?
Yes, it does. It includes them for each of the three trader groups.
That seems like pretty valuable information for only $95 a year. A subscriber would receive on a biweekly basis the same numbers you publish on the last page in your letter, but not the three pages of analysis.
That's right. It creates a market interest in the Commitments data. What I have found through watching other analysts' work is that as this data has become more readily available and more widely reviewed, the quality of others' analysis is so poor that it increases our readership.
Other people are so bad at this that you're looking great by comparison. is that what you're saying?
Yes. And misinterpreting this data causes many Monday moves that go the wrong way. My subscribers are able to take advantage of this on Tuesday morning after they read my letter.
This happens because the data is released on Friday?
Yes, It's released on Friday after the close. You'll quite often see a reaction then. When you read on the news wire what happened, it was that the Commitments numbers came out.
If somebody has for instance, Omega Research's Supercharts program, they could download your data and then figure out how to display it under their commodity charts?
As a matter of fact, I provide specific directions for displaying net COT Index positions in SuperCharts.
There's another number in your newsletter we haven't talked about yet. It's called "move." What's that all about?
I look at two things in the commercial numbers. Number one, we want to look at the actual net position in relationship to the historical period to see how extreme the current position is. If it's extreme enough to move the index to about 90 percent, I call that, for simplicity, a major buy signal, although subscribers should not interpret it as advice to buy. It's simply an indication that commercials are buying in a major way. If it's extreme enough on the other end to get below 5 percent on our index, I call that a major sell signal, which again is not necessarily an indication for subscribers to sell. There is another way I like to look at the data. That is to find out how much the commercials' net position has changed over a relatively short period of time. If it's enough to move the index 40 points in either direction within a six-week period, that shows an unusual amount of commercial activity. I call these minor buy and sell signals to distinguish them from major signals which, as I previously indicated, mean the commercials have reached an historical extreme.
Typically, the minor signals provide key areas to buy bull markets after corrections where the correction never gets deep enough to reach 90 percent on the index. I look at how much the CoT index has changed in the last six weeks. If a minor buy or sell occurs because of a substantial change, I will identify that market with either a plus or minus sign.
Looking at your most recent letter, which is dated April 15th. This is just prior to some truly huge moves in the grains. Looking at your numbers for commercials, the commercials are long in Soybeans only 10 percent, in Wheat 33 percent, Corn 26 percent, Soybean Oil 17 percent, Soybean Meal 23 percent, Oats 32 percent. If one were looking at these numbers in your COT Index, it looks like the commercials are not very bullish. Yet the markets are going bananas. Why isn't this going to fake you out?
The commercials are value buyers. They were the most bullish when prices were closer to their value pricing levels. I had a buy signal in Wheat in March, 1995, at 345, a correction low. In Soybeans, I had a major buy signal in February, 1995, at 550. Beans are now above 800. Commercials are not buying aggressively at this price because the market has moved well above the value level where they previously bought.
But they're not selling aggressively either. Is that the point?
This has been an interesting market. We've had new all-time highs recently in Corn and Wheat along with historically high Soybean prices, yet commercials are not down to zero. That's why in that issue, even though there weren't any buying opportunitys, I told subscribers that this thing may have a long way to go from here because the commercials were not aggressively selling these rallies even at record prices. They can give important indications even when they are not at extreme levels of buying or selling.
A clear-cut example of this occurred in the Coffee market in 1993-94. That was a bull market many people thought came out of nowhere. Everybody assumed it was nothing but a speculative bubble. Back in December, 1993, we had put in a bottom after a long bear market. There was a small double bottom in September, 1992, and then a larger failure swing bottom in April, 1993. The commercials were aggressive buyers on that failure-swing bottom. In fact, it was just a couple of points short of a major buy signal on my index. Then there was a long triangular consolidation that lasted about six months. We broke down out of that consolidation in January, 1994.
What I saw was this. On that breakdown, commercials came into the market. Their net positions moved to the highest level in over two years. It was an indication to me that this was a bear trap. The international foods are quite famous for these because commercials dominate these markets. They typically hold two thirds to three-quarters of the open interest in the international foods. In Cocoa, Coffee and Sugar they are masters at getting a maximum position at the most favorable price. That's what they did there. They moved the price down through obvious chart support at the bottom of the long consolidation and then took it right back up. Everybody assumed the initial move was a speculative rally, that it was the commodity funds coming in and buying oodles of contracts. The major portion of the move was actually commercial buying. The large speculators were selling into the move before it ever got going. They were selling at well below $1. They started selling at 85 cents. As the market went up, volatility increased and the open interest dropped substantially. That was large speculators getting out. They couldn't stand the risk. The market continued to move on commercial short covering, and that's what propelled the enormous move in 1994--commercial short covering.
When you see that situation, it is commercial capitulation, which results in huge moves. For a similar current situation, look at a Cocoa chart. We recently saw a long consolidation and then an apparent downside breakout with a large contingent of commercial buyers ahead of that. You could have anticipated that if there were a downside breakout, it was going to be a bear trap. That's exactly what occurred. Then all of a sudden it was pushing 1400.
Would you summarize how you interpret the Commitments numbers?
Once I get the numbers and convert them to the COT Indexes, I look for two things. I look for the large movement index numbers to see where there has been a major change in the last few weeks. Secondarily, I look to see which ones are at extremes: above 90 percent and below 5 percent.
Would it make sense to analyze the Commitments data on a seasonal basis?
Commodity Research Bureau made that assumption in 1985, and it was later repeated in a couple of market books. They said the correct way to analyze the Commitments of Traders data was to compare current commercial positions with their seasonal average positions. It's a natural assumption to make. Farmers are planting grain in the spring and harvesting in the fall so there should be a seasonality to their hedging. I sent the data to a Ph.D. economist and asked him to check it for seasonality. He concluded there was none. That's why I decided not to print seasonal charts of the Commitments data.
I thought I should show a demonstration of this so I drew up the net positions of the commercials for Soybeans for ten years. People are amazed when they see the chart. It's all over the map. There's plainly no seasonal pattern to it. You can create seasonal charts by making averages, but if you check them statistically, you'll find there's no statistical reliability.
If somebody read your letter for awhile, how long do you think it would take before they could do their own analysis without you?
I've been analyzing this data for 22 years. I typically spend 12 hours on each report, which includes writing the analysis. Since I don't make specific trading recommendations, all I am doing is analyzing the report. I can't believe there are many traders who have that amount of time to spend, and I know there aren't any who have that kind of experience. I can't see where it's profitable for anybody to do their own analysis of the Commitments data when my research is so cheap and readily available.
Yes, I see your point. I don't want to sound impertinent, but you make this sound so easy and attractive. Naturally, our readers would want to know if this is so effective, how come you're not rich from trading it?
I think I am rich. I have five kids with just one left in high school. For the last dozen years I have coached basketball, softball and baseball. I have made enough money in the markets that my family has been comfortable and secure. I'm putting all my kids through college. I've taken out of the markets what I've needed to meet my primary goal which was getting these five kids off to a good start. I will have more time from now on to concentrate on the financial aspect of my work.
So within your own financial parameters you have been successful as a trader using this?
Sure.
Is there anything we've left out?
When you think about the Commitments data, people often think of it as a separate specialty area of analysis. This is basically open interest that we're looking at. There was a whole book written on volume and open interest. Here is a quick summary of the standard open interest and volume analysis rules. (1) When price, volume and open interest are all rising, that's a strong market. (2) If the price is rising but volume and open interest are declining, you'll have a weak rally. (3) If price is declining with volume and open interest rising, you have a strong downtrend. (4) Declining prices with declining volume and open interest means a weak downtrend. Those are the four rules of volume and open interest interpretation you can read in books.
Here is the reality. In the first place, volume is not a predictive indicator in commodity markets. I don't know whether it is in stocks or not, but I've proven it is no more than a coincident indicator in the commodity markets. It's coincident with volatility. When you chart the five-day average volume versus the five day average price range and invert one, you'll see a perfect Rorschach. [See chart.] You get a mirror image. As the price range increases, more traders become active, more stops are hit, and volume increases. Volume is directly proportional to volatility. Therefore, it is not a predictive indicator.
If you look at the four rules I spelled out, they work about half the time. They work often enough so somebody can show examples of their supposed effectiveness. I could show just as many examples where they don't work. For example, there was declining open interest over the entire 1994 rally in Coffee. I don't think anyone will deny that was a very strong rally, and yet it occurred with declining volume and declining open interest, which is exactly opposite to the classic rules.
Open interest is of paramount importance in market analysis, but you have to throw out all the old rules and look at the makeup of the open interest in ariving at your conclusions. By looking at the makeup of the various trading groups, whose breakdown the Commitments data provides, you can tell when those rules are going to work and when they're not.
What's the price of a yearly subscription to your letter and what does it include?
$295.00. That includes 26 issues, a 16-page subscriber's guide with long-term charts of the various markets, plus a video. I gave an all-day seminar in New York a couple of years ago and put the best of that into a 3-hour video, which goes to new subscribers.
Is there any shorter trial subscription?
Yes, a 3 month trial is $95.
That sounds like one of the best bargains around. My thanks to Steve for this enlightening interview. If I were limited to reading just one commodity newsletter, I think I'd have to make it his Bullish Review. You may subscribe to his newsletter or modem update service by writing to Bullish Review, 14600 Blaine Avenue East, Rosemount, MN 55068. Or call (612)-4234900. Fax (612)-423-4949.
Reality
Based Trading Company
1731 Howe Avenue, Suite 149
Sacramento CA 95825-2210
Copyright 1996 by Bruce Babcock, Jr. All rights reserved.