product development and history

 Clarke Capital Management (CCM) has developed over 130 models in three broad time frames, intermediate, long and ultra long. In our intermediate models, winners are held and average of 15 days, losers an average of 5 days. The long time frame models have average days held for winners of about 45 days and losers about 15. Ultra long-term models have average holding periods for winners of approximately 140 days and losers about 45 days. Within each of these holding periods, there is considerable variation.

 During their development, the models have been tested against a database of around 100 different commodity futures products, including Financials, Currencies, Equity Indices, Metals, Softs, Energies, Meats and Grains with data that in some cases goes back to the 1940s.  Each model is tested against the entire database.  For a model to be considered a candidate for inclusion in a CCM program, it must be extremely successful in trading the entire group of markets with the same rules and parameters. We never tune a model to any individual commodity or group of commodities. With the exception that there is variation in the rules for going long or short, any particular model uses the same rules to trade Euro-Currencies as it does Wheat or Copper etc.

 We do vary the testing period for some models. For example, we may only test from 1970 on for some models, while others use everything since the 1940s if data is available. We also develop some models with an emphasis on performance during periods when trend-following CTAs have generally lost money. The reason for these variations in testing procedures is to create a more robust and diversified array of trading models in order to produce superior and consistent performance over a wide variety of market conditions.

 In our research, we try to develop models that not only work in the testing environment but will also work in the future, in real-time, where it matters. We endeavor to make sure a model develops sufficient numbers of trades relative to the number of rules in the model to insure its validity and that the rules themselves have an underlying logical sense to them rather that just being some little-understood idea that “just happens to work”. We perform extensive parameter shifting tests to underscore the reliability of any trading rule and determine and select the most robust settings, not simply the best performing settings.

 Once we have developed a model and feel it is a superior performer, the final step is to fit a “fuzzy logic” trend filter into the model. This proprietary technique allows us to determine certain patterns of trend that, if currently present during a newly received entry signal, will give us a higher rate of success and increased returns should we take this signal. We only take signals which pass this test.

 In developing robust models that work over long periods of time, it is our observation that there are periods where an individual model does not perform well. If a CTA was relying on only this particular model, the CTA’s performance would be poor and his clients may lose a great deal of money. From day one, it has been CCM’s strategy to blend models into programs in such a way that when conditions were unfavorable for Model “A”, then Model “B” would be doing well. Both models are in the long run superior performers but in any two to four year period a particular model may be in a drawdown.

 The more models that can be blended in this way, the better.

The Trading Programs

The first program that CCM developed was called the Domestic Diversified (DD) program. It used a blend of 7 intermediate and 7 long-term models to trade a group of 27 US futures products. In January, 1996, we added European and Asian futures to the mix of markets followed by the program and using the same models, took in our first clients in the Worldwide (WRW) program. We continued to trade the DD program simultaneously, not wishing to force any clients into WRW until the last client left DD in January of 1998.

 In early 1995, the MJC Aggressive Multi-Sector Fund LP (MJC AMSF) was launched by Commodity Pool Operator (CPO), RDB Holdings. We are now and have always been the sole designer and trader of this program. Solicitations and account handling has always been the sole responsibility of the CPO. The program was designed in request from a client of CCM who wanted broad diversity in models and commodities followed. The program currently follows 53 commodity products using a blend of 25 intermediate, long and ultra long-term models.

 In 1995, we had a discussion with one of the Introducing Brokers (IB) that we worked with. They proffered the idea that we could raise a significant amount of money if we designed a program for small retail accounts of $50,000. We considered that it would be a difficult task to create this type of program because the amount of risk to the nearest initial adverse exit point had to be very small so that a negative cycle of several losing trades in a row did not lose too large a portion of the clients fund to prevent the program from being a success. We were unwilling to increase the risk for these clients and still wanted the benefits of diversity. After looking into the situation we came up with 5 intermediate time frame models that, because of their very tight “stops”, would look for only ideal situations in which to enter a position. These situations are low volatility trending environments. Since the amount of initial risk on any one trade has to be kept very low in relation to the amount of capital funding this account so that drawdowns can be contained at a low level when conditions prove unfavorable We were able to come up with a program called Global Basic (GB) which accomplished the task. It started in February, 1996.

 We had a great deal of interest in GB and many clients were opening accounts with multiple units of GB. So in August, 1997 we introduced Global Magnum (GM) in order to get similar returns to GB but with a smoother equity curve due to the increased number of models. It originally contained 10 models similar to those in GB. It currently contains 9 models and because of its unit price of $100,000 is slightly less aggressive than GB.

 Due to a request by a client of the MJC AMSF, we launched a Managed Account version of the MJC AMSF in January, 1998. This program is called Millennium (MIL). Although the two programs use the same models, MJC AMSF trades a few additional products that, because of lack of liquidity, are not suitable for Managed Accounts.

 The Orion (ORI) program was started in July, 1999. It was originally designed to be similar to WRW when we were concerned about WRW reaching capacity and we wanted to have a multi-year track record in place in a similar product should this happen. As it turned out, WRW’s assets did not continue to grow after its losing year in 2001 and is currently well belowcapacity. Orion has since evolved to be the most conservative product in the CCM family of products with very moderate drawdowns. Like WRW, it is a blend of intermediate and long term trading models, but the models are a bit more selective than those of WRW.

 In May, 2000, Futures Investment Company, a CPO, launched a fund call the Alpha Fund LP. Except for a brief period, we have been the sole CTA trading this fund. This program uses a blend of intermediate and long-term models. Several times a year, the CPO will liquidate or partially liquidate the fund after a equity run up and restart trading with new signals, generally at the same time. It should be noted that we also offer this “liquidate and restart” capability to all clients of CCM managed accounts, but reserve the right to restrict this should a client liquidation/restarts be excessive.

 In January, 2007, the Omega Program (OMG) was launched. It is a managed account version of the Alpha Program. While CCM retains the sole right to make “liquidate and restart” decisions for Omega, such decisions often coincide with the “liquidate and restart” decisions made for the Alpha Fund. As such, new Alpha clients may start only on the first day of the month, redemptions are allowed only at month-end and Omega positions are aligned to current levels at that time.

 The FXF program was launched in May, 2004. It trades only Financials, Currencies and Equity Indices with a blend of long and ultra long-term models. The commodities followed are the most liquid commodities and the program was designed to have a very large capacity.

 In May, 2005 we launched our most diversified product, Jupiter (“JUP”). It trades all the products that our previous most diversified program, Millennium, trades and adds Equity Indexes, which Millennium does not trade. It uses 66 models as compared to Millennium’s 27 models. The mix of the models, although a blend of long and ultra long-term models like Millennium, is more weighted to the ultra long as opposed to Millennium 50% long-term / 50% ultra-long-term mix.

Past results are not necessarily indicative of future results.