Most trading disasters are studied from one side only: the side that lost. What gets overlooked is that for every catastrophic loss in a futures market, someone on the opposite side of that trade was collecting exactly those losses as profits. The Metallgesellschaft case is one of the most instructive examples in trading history precisely because both sides are documentable. One side had a strategy. The other did not.
Do you think a systematic trading approach could have brought one of Germany’s largest companies to its knees? Unfortunately for Metallgesellschaft (MG), a German metals and oils conglomerate, it happened. MG was long crude oil futures on the New York Mercantile Exchange (NYMEX) most of 1993. They lost, depending on the estimate or source, $1.3 to $2.1 billion dollars. Those traders in short crude oil futures made the money MG lost. It’s a zero-sum game. If you trade like the mindless majority, you lose. MG traded like the majority.
During the course of 1993 crude oil futures had a steady slow decline from May through December (see chart A). A move such as the crude decline is not necessarily unusual, but this time was different. Why? There is always someone on either side of a futures trade. The difficulty lies in determining who is on the opposite side of a trade if one side is known. We know MG lost as crude prices dropped, but who won the other side of the trade and how? More important, can the individual investor or trader benefit from an understanding of who won and how?
In the aftermath of MG losses a variety of explanations developed. People were treated to academic mumbo jumbo from MBA programs as to why MG lost money and numerous articles condemning futures and their speculative nature. The actual explanation is simply that MG were not good traders. However, this case is a primer not because MG lost but because someone else won.
That distinction matters. The financial press covered MG’s collapse as a story about the dangers of derivatives. But derivatives did not cause the loss. A lack of discipline, no exit rules, and no systematic plan caused the loss. The instrument was neutral. The approach was not. Trend followers trading the same instrument, in the same market, during the same period, produced some of the best returns of their careers.
Trend Following Played the Hand in MG’s Defeat
Trend following played the hand in MG’s defeat. Trend followers left a performance trail of their activities over the course of 1993. The job of explaining this is made easy by the near 100% correlated performances (see chart B). Seven different trend following firms, trading independently, with no coordination, produced remarkably similar monthly results across the same period. The correlation is not coincidence. It is the signature of a common approach reacting to the same price signals.
Chart B: Monthly Performance of Trend Followers — 1993
| Trader | Jun 93 | Jul 93 | Aug 93 | Sep 93 | Oct 93 | Nov 93 | Dec 93 |
|---|---|---|---|---|---|---|---|
| Abraham | -1.2 | 6.6 | -5.3 | 1.2 | -6.6 | 3.5 | 12.5 |
| Chesapeake | 1.0 | 9.5 | 5.8 | -2.7 | -0.1 | 1.1 | 5.8 |
| EMC | -1.5 | 22.0 | 9.3 | -2.9 | -2.0 | -2.4 | 8.2 |
| JPD | -6.9 | 10.2 | -2.1 | -4.1 | -2.0 | 2.7 | 8.6 |
| Rabar Market | -1.3 | 14.8 | -3.9 | -4.1 | -6.0 | 5.6 | 10.1 |
| Saxon | -2.7 | 20.5 | -14.3 | -2.1 | -1.1 | 6.6 | 17.1 |
| Sjo | 8.0 | 4.3 | 12.6 | -11.4 | -1.3 | -0.9 | 6.0 |
There is slight variation in performance data due to differing levels of leverage employed by trend followers, but the key to the explanation lies in the months of July 93, December 93 and January 94. Those months don’t require much more than a glance at the correlation to confirm the similarity in the strategies employed. They all made money in July and December, and all lost in January.
Read that again. Seven independent firms. All profitable in July. All profitable in December. All losing in January 1994. No coordination, no shared positions, no communication. Just the same systematic rules responding to the same price reality. That is what a rules-based trend following system produces: consistent, correlated behavior across practitioners because the market is the common input and the rules are the common filter.
How did trend followers position themselves to benefit from a completely unforeseen event like MG’s bad trading plan? Here are some insights as to why:
- Trend followers don’t predict market movement, they react to it. May and June showed downward crude pressure and trend followers followed the trend. They have limits on market sectors and risk limits on the total portfolio as well. Position sizes are based on the amount of money under management.
- Every day trend followers know how many contracts must be on based on total capital. The key to their strategy is to place good profits at risk to participate fully in a trend. For example, after trend followers initiated positions and were rewarded with strong profits in July, they were willing to risk those profits again, which is what they did with their crude oil positions.
- Trend followers will risk 100% of the profit in a trade. For example, in August with nice profits in hand, they would have been willing to risk all of their profits and still lose a fixed % based on the original stop.
- Trend followers are willing to let profits on the table turn into losses. Trend followers let the market tell them when the trend is over (i.e. Jan-94).
- Trend followers don’t favor liquidation. They want to capture 60-70% of a trend, not just 15%. Big money is made in the big moves. For example, trend followers want to hold silver as it moves from $6 to $30, not just hold it from $6 to $9.
Each of those five points is a direct inversion of how most traders behave. Most traders predict. Most traders protect profits early. Most traders exit when pain builds. Most traders favor liquidation over holding through drawdown. Trend following does the opposite on every count, and the MG case shows in real dollar terms what that difference produces. For more on the risk management principles that make this possible, and the position sizing rules that kept trend followers correctly calibrated throughout 1993, see the full TurtleTrader rules.
The July Entry
Examine the crude oil chart and the performance chart. Trend followers all entered in the May and June 93 period and then, in July, the market nose dived. Trend followers were now in for the long haul with a fantastic and very profitable short position firmly established. What’s the lesson? If trend followers had not been profitable, but had lost their maximum percent allowable instead, they would have taken their losses and exited.
This is the other side of the discipline coin. The same rules that kept trend followers in a winning position through the summer and autumn of 1993 would have cut them out immediately if the trade had gone wrong. There is no asymmetry in how the rules apply. A losing position that hits the maximum allowable loss is exited without hesitation. That willingness to take a defined loss quickly is what allows the system to hold a winning position aggressively. The two behaviors are inseparable. MG had neither. They had no exit on a losing position and no plan for building on a winning one.
After July: The Waiting Game
In August trend followers are firmly established short in crude oil futures. Obviously, MG is long at this point. MG is desperately trying to hang in there (hoping for crude to stop falling) while trend followers wait patiently (and aggressively short). Trend followers are no doubt opportunistic predators. Their waiting game continues through the summer with no real movement up or down (see chart A). MG has met margin calls and stayed in the game in hopes of an upward crude push. They do not anticipate or understand the discipline of their opponent on the opposite side of their long trades. Trend followers are not exiting anytime soon since the trend is down. An exit would violate their most fundamental rule: Follow the trend.
Trend followers were not just short, they were aggressively short with profits reinvested back into additional short crude oil positions. On the other hand, MG has no apparent strategy. They refuse to take a loss early on. The whole MG debacle would have been a footnote in trading history if they had simply exited after the July losses.
That sentence is worth sitting with. A $2 billion disaster was avoidable. The decision to stay in a losing position rather than accept a manageable early loss is what turned a bad trade into a corporate catastrophe. This is the loss aversion trap described in behavioral finance, playing out in real time with billions on the line. Trend following has no loss aversion built into it. The rules decide when to exit, not the trader’s emotional state or their hope that the market will turn around.
The Late November and December Route
Moving through October and November, the situation culminates in MG’s near corporate destruction. Crude oil begins its final descent in late November and into December. At this time MG management closes out of all of their trades that fueled the November and December crash. Trend followers are still short from the May, June and July period and raking in profits at the expense of one ill-fated firm.
Ultimately all good trend following must come to an end. Trend followers would eventually need to begin their crude oil futures exit. The exit came sooner than later as January 1994 proved to be the month of trend followers’ liquidation. Look at the performance of January 94. All trend followers lost for the month as they extricated themselves from their history making winning crude oil trades of 1993.
That January loss is itself a lesson. Giving back gains at the end of a trend is normal and expected. It is the price of not exiting early. Trend followers did not know in November that December would be the final month of the crude decline. They stayed short because the trend was still intact. When it reversed in January, they exited. The total result across the full year speaks for itself.
Full Year Performance: 1993 Returns
The numbers below represent what systematic discipline produced across a full year while MG’s losses compounded toward catastrophe. These are not cherry-picked outliers. They are the documented annual returns of firms that applied consistent trend following rules across all markets, of which crude oil was just one.
| Trader | Returns % 1993 |
|---|---|
| Abraham Trading (Turtle 2nd gen) | 33.6 |
| Chesapeake Capital | 61.8 |
| EMC | 64.2 |
| JPD | 23.3 |
| Rabar Market Research | 49.7 |
| Saxon | 52.6 |
| Sjo | 8.9 |
Conclusions
Trend followers reacted to crude oil movements with a steady plan of systematic trading. They played short to MG’s long.
The MG example, with trend followers winning, is not 20/20 hindsight. Trend followers make money in the same months in the same stocks, commodities and currencies by employing consistently their trend following systems. Scandals and rogue traders in the trading game come and go, but trend following continues to produce great profits by taking advantage of other traders’ poor strategies.
The deeper lesson is structural. MG’s failure was not bad luck. It was the predictable outcome of a large directional bet held without rules, without exits, and without any mechanism for limiting loss. Trend following is the structural opposite: a system built entirely around rules, exits, and loss limits, applied the same way in every market regardless of size, opinion, or institutional pressure. Every time a large player fails in the way MG failed, the conditions that created their loss are the same conditions that create trend following profits. That is not coincidence. It is the architecture of the approach. For the full story of how systematic trend following was built and tested, see the TurtleTrader story.
Frequently Asked Questions
What happened to Metallgesellschaft in 1993?
Metallgesellschaft held a massive long position in crude oil futures on NYMEX through most of 1993. As crude declined steadily from May through December, MG accumulated losses of $1.3 to $2.1 billion. Rather than exiting the losing position, they met margin calls and held. The eventual forced liquidation nearly destroyed the company. Systematic trend followers on the short side collected those losses as profits.
How did trend followers end up on the right side of the MG trade?
By reacting to price rather than predicting it. Downward crude pressure appeared in May and June 1993. Trend following systems read that signal, went short, and sized positions according to capital under management. They had no knowledge of MG’s position or its size. The same rules that put them short would have cut them out immediately if the trade had moved against them.
Why is the correlation between seven independent trend followers so high?
Because they all use the same type of rules reacting to the same price signals. Seven firms, trading independently with no coordination, all made money in July and December 1993 and all lost in January 1994. The correlation is not coincidence. It is the predictable output of systematic trend following rules applied to the same market reality.
Why didn’t trend followers exit their short positions before January 1994?
Because the trend had not ended. Trend followers let the market tell them when the trend is over, not their own judgment or a profit target. An exit before the signal would violate the most fundamental rule: follow the trend. Giving back gains at the end of a trend is the price of not exiting too early and missing the full move.
Is the MG pattern unique or does it repeat?
It repeats. Trend followers make money in the same months in the same stocks, commodities and currencies by employing their systems consistently. Every time a large player accumulates a directional position without rules, without exits, and without loss limits, the conditions for trend following profits are created. Scandals and rogue traders come and go. Systematic trend following continues to find opportunity on the other side of those failures.
Trend Following Systems
Want to learn more and start trading trend following systems? Start here.
