Well, it turns out that I left about $15,000 on the table last month.  There was a revision to my system that I had been meaning to implement for the sake of clarity and intuitive sense, but I failed to do it during January 2007 because this revision didn’t seem that it could have a material effect on the trading results.  Well now that January has come and gone, and the biggest event was the volatility surrounding the Bank of England rate hike on January 11th, I revisted how the system filters signals around economic events.  I had the system coded as “do not trade on Wednesdays”, but what I really meant was “do not trade on the day that the EIA releases crude oil inventories.”  It turns out that the EIA loves the 3 day weekend, so that anytime there is a Federal holiday, the inventory number doesn’t come out until Thursday.  This is relevant because there were two holidays in January that threw the schedule off (New Year and MLK day).  The “do not trade on Wednesday” caused me to miss two hugely profitable trades…if those trades had been included in my system results, I could have made in excess of a 70% return for the month.  Ouch!

So I’ve closed the lazy coding gap, and I re-did the backtest for 2006, and no fooling, the results went up by 40%.  Live and learn.  I’ll have a new version of the system active on Sunday night then, and a new backtest result posted in the sidebar.

 On another note, the Employment Report on Friday caused a short lived flash of volatility in both directions, but neither move resulted in an entry signal, so for February, I am still 0 for 0 with a 0% return. :-)   So much for predictions.

tf_clock_clean_20070202.png

The spike up followed by the spike down on the chart above may not look like much, but I’m sure it was enough to clean a lot of short term traders’ clocks.