Backtesting the ETF Screen

In the last post I identified 12 ETFs (6 pairs) fit fit my liquidity requirements and provided what I think is reasonable diversity for a first attempt at developing a system.   The next step is to do some rudimentary tests to see how the system performs.

It is important to see all of the variations in start date(s) for the backtest runs.  The system rebalances every four weeks.  So the first thing to try is changing the start date by moving it back 1 week, 2 weeks and 3 weeks as shown below.  Note that it is not necessary to go back 4 weeks as this would simply add one backtest period but other than that would give same results as not moving the start date.

As you can see below, the first two runs are similar but the third and fourth degrade with the loss around ~30% instead of ~45%.  Don't forget that the objective is to maximize loss as this is a short system.

It can also be seen from the above graphs that the last two runs also produce higher profits (not good for a short system).  Now lets have a look at gains (losses) for each of the last two years and compare them to the SP500.

 

I should point out that these results are with the first run and would not be as good with run 3 or 4.  As can be seen the system is profitable for both the last two years.  Unfortunately these are idealized results and would not be seen in practice.  For one thing there are fees/interest charged by the broker to maintain the short position that could easily eat up the profits for the most recent year.  Another problem is that the short positions need to be actively managed around splits/reverse splits and dividend payouts.

So although the results look very interesting I would not say that I have a tradeable system - at least not yet.

Steve

Comments

[Mike Millan]

i don't think it is the long term or short term strategy that is best. It all depends on your expertise and your investing needs to determine when where and how to invest. This in my point of view is too generic and a very narrowed analysis.

[StockMarketStudent]

Mike - think of me as a "system development junkie". I don't claim to be writing for everyone and in fact that would be impossible to do. I look for an advantage generally involving longer term trading (although not always) and try to develop systems around that advantage. I get a rush out of this and I hope that some others are interested in reading what I have to say. If not, you don't need to tune in. Thanks for the comment though.

Steve