• Performance
• Services
• Company
• Learn About
• My Attain

The Crude Oil Rally: Comparing Trend Following Styles

August 23, 2004

You can't turn on the news these days without hearing about record high Crude Oil prices. It's a hot topic to be sure, as the commodity has seemingly made new all time highs week after week as it marches towards $50 a barrel. Volume on the New York Mercantile Exchange (NYMEX) where Crude Oil futures are traded has surged to all time highs, with 4.6 MM contracts traded in June, and a seat on the exchange is now priced at $1.65 MM, which is about $250,000 more expensive than a NYSE seat.

All of this attention is not unwarranted. The rally has been impressive to say the least. A chart of the increasing price of Crude Oil could be the posterboy for an upward price trend. The Crude Oil rise began in earnest in November of 2001, as the US prepared to attack Afghanistan and the specter of a war with Iraq loomed ever larger. Using CSI back-adjusted data, Crude Oil was trading at an adjusted price of 13.57 on November 1st of 2001 and 46.72 as of the market close last Friday. That's a gain of $33.15 per barrel, or a 244% rally. Wow!

Given the prominence of Crude Oil in the news and minds of investors, many investors pondering a trend following system want to know how each of the different trend following systems have handled this rally. While looking at the profits and losses of a single market when comparing trend following systems provides skewed and unreliable information on the overall performance of each system - looking at the Crude rally does provide a good opportunity to compare the different styles each of these trend followers employ.

We compared the activity of six different trend following systems over the time period November 1st, 2001 through August 20th, 2004 on the rallying Crude Oil market to pinpoint some of the differences between these systems. The systems included in our example are Aberration Plus by Trade System, Inc. , Andromeda II by Andromeda Futures, Axiom (Long Term) by Trading Visions Systems, Inc., Brixx by Alfaranda CTA, LaJolla by On the Close, and Trend Channel by Scientific Trading Solutions.

Each system was run with the developer recommended risk filters, if any. A risk filter is a type of risk management built into many long term systems which does not allow a system to take a trade if the risk per contract of that trade is greater than the filter level. For example, Aberration Plus uses a $3,000 recommended filter. The system signaled a long trade on August 4th, 2004 with a risk per contract of $4,039. Because this risk was greater than $3,000 — all investors using the recommended risk filter did not take Aberration Plus' long Crude Trade on August 4th. As can be seen in the table below, every system except TrendChannel utilizes a risk filter.

Using the table below, many differences between the six trend following systems become apparent. The first telling statistic is the number of trades for each system over the past two and a half years. While it may seem that such a strong rally would imply a relatively small number of trades, the average number of trades across the six systems was just over 11 trades. This tells us that not every system is looking to simply ride the trend higher.As an example, compare just three trades for Aberration Plus with the twenty-one trades for LaJolla. Aberration filtered out some trades due to its risk filter, as mentioned above, and won't get out of a trade until prices return to the 80 day moving average(which can take a while) In contrast, La Jolla is a reversal system which will attempt to lock in profits and pick the end of the trend by going the other way. Notice in the charts below how LaJolla reversed short towards the end of May this year — the only system to dare go short in 2004.

The next important statistic in the table below is the percentage time in market column, which shows the percentage of trading days each system was in a position. Any time you are in the market, your money is at risk, so with all else being equal it is better to achieve profits with as little time in the market as possible. You will quickly see that while Brixx led the way in terms of profits, it was also the riskiest system in terms of time in the market, being in a position 88% of the time. To take this risk into account, we can divide the net profit by number of days in the market to get a ratio of dollar returns per day in the market. Using this statistic, we can see that Axiom was slightly ahead of Brixx on a risk adjusted basis, even though that system lagged in total profit.

Sticking with risk measures, we also measured the drawdown of each system in dollar terms, as well as the daily standard deviation of returns for each system. As would be expected, the system with the largest risk filter, Andromeda($4,000), posted the largest drawdown, while TrendChannel, which has no risk filter at all, posted the second largest drawdown. As for the standard deviation of daily returns, the statistics mirrored the time in market findings, with those systems in the market the least seeing the smallest deviation in daily prices (Aberration - just $213 per day)

The last important number is the days per trade statistic. You will see that Brixx achieved its profits by holding onto trades nearly twice as long as any other system, while LaJolla saw the shortest trade duration because of its reversal strategy. It is interesting to note both Andromeda and Axiom's shorter trade durations, 36 and 34 days per trade. These numbers would be much larger if not for profit taking logic within both systems. The Andromeda chart below shows how Andromeda exited its long position in early May once it achieved $10,000 in profit per target, while the Axiom chart shows that system exiting its long position at approximately the same time.

The conclusion to be taken from this exercise is that higher returns come at the expense of higher risk. The three systems with the highest gains also ranked among the riskiest systems in one statistic or another. Brixx was in the market the most, while Axiom and TrendChannel had higher drawdowns. This also shows that some rallies, while extreme, are difficult for trend following systems to capture. It seems counterintuitive that a trend following system could have trouble catching a trend, but that can happen in a volatile market such as Crude Oil. That volatility means higher risk trades which many of these systems filter out, thereby keeping investors out of long Crude trades. Crude Oil's two and a half year rally also saw several sharp sell offs down to - or beyond - the 100 day moving average, which can stop out trend followers with small profits before the market returns to new highs. The ideal scenario for trend followers is a slow, steady crawl higher (or lower), where prices stay above the 50 day moving average and volatility is low enough to warrant a low risk entry.