What I found out from this experiment is that I could execute the system with consistency. That the system as implemented resulted in neither gains nor losses. Making a tweak to the parameters of the system turned it from break even to profitable and that to exploit my edge this is where I would address the shortcomings over the next sample size. In addition I noted that there were a few data points that would help future iterations so I adjusted my spreadsheet to document them.
Unlike the first series I failed slightly in the consistency department. Immediately I botched the first few trades as I was still adjusting to the timing and the mechanics. I ended up choking off trades too quickly resulting in missing a significant portion of the move on two of the most profitable positions within my sample size. Additionally, I was slow to react on a number of signals and was forced to set time aside to investigate my trading software to improve efficiency of execution. With the initial hiccups out of the way the results are in.
 |
| First Trail Run |
 |
| Second Trail Run |
With a simple adjustment of adding a trailing stop instead of closing out the position in the last half hour of the trading day a break even system became a profitable system. Knowing that a simple adjustment to the execution of the first sample size would have made it profitable and then executing that over another sample with consistent results, win/loss, and stops gives me greater confidence in continuing with this.
Over the next trial run the refinement to the system will be focused on a proper position size algorithm for each trade. As I'm only willing to tweak one parameter per run I'll look to first increase size while maintaining the current stop. Currently I am using a fixed position size and stop loss regardless of price and to improve profitability trading a $4 stock the same as a $40 stock is inefficient and wasted opportunity cost.
No comments:
Post a Comment