In the summer of '04, I developed a second program trading model for Profund funds. Like Aggressive Hedge I, it trades the Profunds long and short funds based on the price changes of the underlying index and purely technical indicators. It is simpler than Aggressive Hedge I in terms of free parameters, is less suseptible to whipsawing and yet has been (or was) nimble and efficient in identifying market turns. It spends more time in cash than Aggressive Hedge I. However, it is much less volatile and hence easier for me to live with as a possible guide for trading my own account. It has done well both during the bear market of '00-'02, and during the bull market and in the flat market of '04 since then. Its date of inception is Aug 13, 2004. The basic idea came to me while doing a swim and did NOT involve fine tuning a significant set of free parameters to maximize returns sensitively dependent on these parameters. Instead, the model is robust; the tunable parameters are few and you can see the result of varying them on the test page. The original formulation is the one I settled on. In the charts, prior to this date the model results are backtest. This model has shown less volatility than Aggressive Hedge I.
Applying this strategy to single long/short Profunds gives strong results through most of 2005. However, I felt it likely that it would perform even better and with less volatility if instead of putting the entire allocation each day into one fund, I split it between several Profund vehicles and making allocations based on the underlying indexes independently. The chart at right shows the 3 performance plots for Aggressive Hedge II applied to the NDX, to SPX and diversified between several indexes, since Feb 3, '04. The diversified strategy has as little volatility as the SPX version, and yet returns just as well as the NDX version.
As of the end of 2005, since the inception date of Aug 13, 2004, the Aggressive Hedge II model has only done about as well as the NASDQ 100 itself, and heading into 2006 the performance has deteriorated badly. As with Aggressive Hedge I, this purely mechanical model is no longer capturing the trading behavior of market participants. With the rise of hedge fund dominance in the total trading volume and the effort put into mechanical trading systems, the pace of evolution in this financial ecology is speeding up rapidly. A gut-check and it's clear I see no prospect of trading this with real money. So, I've ended this experiment and focusing on the discretionary models.
The performance of Aggressive Hedge II is shown on the Mechanical Models page chart.
Annual Performance:
2000: +171% (vs. NASDQ 100: -37%)
2001: +212% (vs. NASDQ 100: -33%)
2002: +208% (vs. NASDQ 100: -38%)
2003: +164% (vs. NASDQ 100: +49%)
2004: +64% (vs. NASDQ 100: +10%)
2005: -21% (vs. NASDQ 100: +3%)
2006: -52% (vs. NASDQ 100: +7%)