Friday, March 28, 2008

My system switch

Last weekend I did some backtesting using my new system switch.
The results were less than promising.

The test period I chose initially was one where I feel is a particularly good period for running backtests. Its from 01-01-1998 until 01-01-2003. The market index (XAO) returned 6.7% pa annualised.

Monte carlo analysis was done over 5,000 portfolios. Starting with $50,000, the averages were:

Without switch:
Number of trades: 179
Profit: $294k
Maximum drawdown: 7.95% (worst 17.5%)

With switch:
Number of trades: 88
Profit: $70k
Maximum drawdown: 13% (worst 31.9%)

As you can see, the number of trades taken were halved compared to no system switch, but obviously the trade expectancy did not make up for this, and as such the CAR was reduced significantly.

The switch, though I haven't tested it sufficiently as yet, I suspect may help my system when in a prolonged bearmarket. A quick glance of the long term performance of the market will tell you that stocks do not spend most of their time in a prolonged bearmarket.

So my conclusion at this point is to just weather the drawdown and keep on trading the system. The only assumption I am taking here is that at some point, maybe after 1,2,3 years, is that the market will assume its long term bullish bias. I have no reason to believe otherwise.

In the meantime, I am developing other systems that will compliment my long term trend following system. I am looking at the forex markets due to its low correlation to equities.

I am still playing with the switch trying to make it work. Results will be posted here when the tests have been done.

No comments: