The Backtest Results Thread

kut2k2

kut2k2

It occurred to me that in the interest of not reinventing the wheel, we could use a thread collecting the results of backtests that have been performed on public-domain strategies. Whether they are winning strategies or not is immaterial. What matters is the knowledge base.

Hopefully many of you will contribute.

I'll start things off with this great post by Hook N. Sinker in response to the thread

" Just show me one mechanical technical analysis that has been profitably backtested"

http://www.elitetrader.com/et/index...fitably-backtested.147576/page-5#post-2218032
 
Attached are the results of an interesting Down trend reversal strategies tested on a basket of ETFs. The idea of the strategy is there has been significant selling on the Monthly and Weekly charts and then we have a big up day on the Daily chart. Stop and Limit price targets are included. Would be interested to see how this would perform with stocks. The strategy can be backtested and modified here.

http://www.fasterbull.com/members/strategy-builder.php?id=810

fan27
 

Attachments

  • DowntrendReversalStrategy.jpg
    DowntrendReversalStrategy.jpg
    253.1 KB · Views: 140
fan27 said:
Attached are the results of an interesting Down trend reversal strategies tested on a basket of ETFs. The idea of the strategy is there has been significant selling on the Monthly and Weekly charts and then we have a big up day on the Daily chart. Stop and Limit price targets are included. Would be interested to see how this would perform with stocks. The strategy can be backtested and modified here.

http://www.fasterbull.com/members/strategy-builder.php?id=810

fan27
More...

20 trades in 20 years ... does it sound like there is some possibility of over-fit?
 
dom993 said:
20 trades in 20 years ... does it sound like there is some possibility of over-fit?
More...

While that is a valid concern, I don't think that is the case for this strategy. Here is why.

The strategy analyzes three different time frames (daily, weekly, monthly). The monthly time frame is looking for price to close below the lower Bollinger band two months ago and then to have a 12 month low one month ago. In other words, there has been sustained selling on the long term chart. If I backtest just those two conditions on my basket of ETFs, I get only 80 signals. The original strategy also looks for 2 weeks ago and 1 week ago to be down. Lastly, the daily condition is that we have a big up day. The hypothesis of this strategy is that sustained selling will eventually lead to some sort of bounce. The big up day on the daily chart is our trigger. The fact that we only have 20 trades is due to this pattern not occurring that often. From what I have found, trades with very good odds do not occur often. If a broad range of stocks were included we would have more signals, but I suspect this strategy will perform better on Index based ETFs.

fan27
 
You must log in or register to reply here.
Back
Top