Out of the Starting Block Part 2

Here are the simulation stats I wasn't able to provide in the last post due to internet problems.  The results look fairly good but there are areas that need improvement.

First thing I would like to mention is that I forgot to point out the risk management rules I incorporated into the simulation.  The rule IndWeight < 10 is used to restrict the weight of any given industry to 10% at the time of purchase.  Since there are 30 stocks held this means no more than 3 stocks in any one industry.  The rule SecWeight < 25 which means that holdings are restricted to no more than 25% in any given stock sector.

The first thing I look for when running a simulation backtest is the percent profit per closed trade.   This simulation generated 17.6% profit per closed trade (see below).  Assuming future performance is less than backtest, 17% provides some room for poorer performance and still not get killed by commissions.  If the profit per trade were 5% for a backtest then I would have to think twice about deploying the system.  The backtest generated 57.7% realized winners with winners averaging three times the profit of losers.  I like what I see so far.

Now for the risk statistics.  The trailing 3 year Sortino Ratio is 1.04.  This is not bad considering the terrible time the markets have had the last year.  Alpha is 23.9%.  See figure below.

The next Item I like to view is the annualized performance by calendar year.  As you can see below the simulation generates some pretty consistent and healthy excess returns compared to the benchmark S&P500.   One of my goals is to be profitable every year.  2001, 2002 and 2008 were substandard but I have some ideas for how to improve these results.

From the screenshot below you can see the equity curve, % Max Drawdown and % Cash Invested.  I have marked the areas in the % invested where there is some spare cash when the stock market is declining  This provides an opportunity to buy up short ETFs as cash becomes available.  Doing this improves results considerably in 2008 and will likely provide better results in 2001 and 2002 although I have to analyze these years a bit closer.

The main thing that bothers me about this simulation is the drawdown.  Although capped at about 20% buying short ETFs using spare capital doesn't reduce the max drawdown in most cases.  The trading capital is fully utilized when the drawdown occurs.  So I need to study this problem a bit more.

The following are publicly accessible (to Portfolio123 members):

The simulation backtest is available here: http://www.portfolio123.com/port_summary.jsp?portid=401285.

The custom universe is available here: http://www.portfolio123.com/screen_summary.jsp?mt=7&screenid=20173

The ranking system is available here:  http://www.portfolio123.com/app/ranking-system/59261