Make sure you guys give it a shot with the random entry quiz at: http://www.elitetrader.com/vb/showthread.php?s=&threadid=10909 peace axeman
Here's results of a random entry with a trailing stop in the SP market using daily data from 1983 - present.
Nice work Alan. I think there are inherent limitations with running the logic on daily data, tic resolution being one of them but the stops appear to be large so it isn't a big deal. I think in order to really test random entries we will have to look at intraday data to bump up the sample size. I was going to run it quickly over a smaller compression but since I can't copy the gif I'll have to take the time to retype something similar shortly.
Do you have any rseults using trailing stop and target? Say 1 to 2, or 1 to 3? That's the first ever I've seen of any results with trailing stops. Very hard to find accurate data if like me you don't know how to program. Most data is all based on fixed stop. Would really like to compare just fixed stop and target vs trailing stop and target vs trailing stop no target Thanks for the chart
Here's a 11 year test on SP from 1/1/1990 - 12/31/2000 using 30 min. data and forcing exit on the close. This way new trades start every day. I also reduced the trailing stop from 5x the average true range to 3. I'm not sure any of this means anything. I think I'd need to do the test 1,000 times and build a distribution of profits/losses. In my past research I found that trailing stops are effective to the degree a market trends. Using a chi square test to test for trendiness, the SP is a poor candidate. I ran tests on the Yen and it worked much better. Some things are pretty clear. Using trailing stops gives less than 50% winners. The sizing of losing trades is less than winning trades due to letting winners run.