I have made some changes to the hedge fund portfolio, in terms of portfolio composition and parameters for trading strategy. I have increased the diversification of the portfolio by adding an international real estate (RWX) and bond ETF (BWX). I have added a US bond ETF that tracks longer term (20+ year) treasuries (TLT) to balance out my yield curve exposure with the existing short term ETF (SHY). I have also substituted a Russell 2000 US stock index ETF (IWM) for the S&P 500 index ETF (SPY). My reasoning is that the Russell 2000 index is more representative of the US economy than the S&P 500 index.
I have made a slight change in the parameters used to generate trading signals for this system. Thus I will update the system at the end of each week, instead of at the end of the month. The new parameters should make the system more responsive to changes in trend direction, but it still maintains the characteristics of a long term trend following system. Thus, I do not anticipate much of an increase in trade frequency.
I have also added disaster stops to guard against extreme moves. These stops are based on volatility and maximum allowable portfolio risk and will be recomputed each week. They are not true trail stops, as, for instance, if a long trend reverses, the stop will move down as the price declines, not remain fixed as with a true trail stop. Since I will be recomputing positions each week, the stops are designed to preserve capital during a huge intra-week move. If a stop is triggered, the position is eligible for reentry at the end of the week. However, if there was an extreme event that caused the stop to be triggered, I will wait for the markets to settle down before reentry. See my recent post “The Bane of Automated Trading Systems” for my concerns about extreme events. Note that collective2 has not properly updated my current positions. I will enter stops when they have done this.
Note that I did partially rebalance the portfolio due to the addition of the new ETFs.
Note that the system performed poorly in October with a 6.5% loss. The recent market swings between fear and relief as the European farce plays out have created an environment that is not conducive for trend following systems. However, as always, trend followers must be patient. Strong trends will emerge again.
|# Profitable||14 (38.9%)|
|# months tracked||15|
|Profitable months||8 (53.3%)|
|Avg trade duration||98.6 days|
|Annual return (compounded)||1.9%|
|Max peak-to-valley drawdown (historical)||21.52% (incorrect, see update for August 31st)|
|drawdown period||May 02, 2011 to Aug 19, 2011|
|Correlation w/ S&P||0.320|