Jeff, You say you backed into the correct settings. If you don't mind sharing, what are your revised settings? BTW, which software package are you using to do the backtest? Thanks
Hi Jan. I use eSignal. I made several subtle changes to the article; - Entry at the close - Changed to 10 day hold, didn't like the 20 day drawdown - BB now set at (11, "Low", 1.6) Attached is a pic of the timeframe discussed to tweak the BB. --Jeff
Did not realize the author is a member here. Cool Vishnu, I am still trying to get my mind around that huge 20 day stop. What was the basis for that number? Even though the avg hold time was only 1.25 days isnt it just a matter of luck that the 20 day stop was never hit? Thanks, bhds
Yes. As an anecdote, one of my favorite hedge funds is Rennaissance Medallion, started by Jim Simons. He was a world famous mathematician who decided he would basically study every relationship in the markets he could find and then trade on the best of them. His fund is amazing: 35%+ / year returns. He returns money to investors every year. One of his investors tells me he even has systems for trading when the snowfall in Central Park hits more than 3 inches. I like his quote, "when I walk into the office in the morning I don't wish to be smart that day but to be lucky."
i am not subscribed to active trader. can anyone be a little specific about the strategy - if that is not asking too much.
The answer is that you can't avoid a catastrophic loss with that type of exit. Examples: 16th October 1987, 20th July 1990, 26th March 1992, to choose just three I checked up right now (I am sure there are others). Drawdowns were approx 26%, 14%, and 7% respectively. The system may well still turn a profit, but no system with almost open-ended risk, compared to much smaller upside, is very appealing. Also, the fact that there have been no losing trades during the test period ought to tell anyone that the test period is too short. This one has "black swan" written all over it. An simple modification would be to use a stop or a put option in order to cap the downside to a certain maximum. But this would obviously incur costs which would lower the return. I'd be interested in seeing the numbers once some form of genuine risk control has been implemented, but as it is, the system strikes me as the sort of thing Nassim Taleb delights in poking fun at.
Exactly. The worst case nightmare in this system is an entry made, then Close and BB's go vertical. You can decide to hold the 20 days or, as you said, put in a second backup Stop to the program which is your Max pain level. As you cited, the situation is a rare occurence but it will happen. A quick glance at the backtest probably is a setting of around 6% fixed Stop. The other side is the Bull market. If you downloaded the chart posted on Page 2, you will see awesome entry points on the pullbacks. After entry, I see a tremendous potential putting in trailing Stops to maximize profits. That's what those horizontal red and blue bars are. Not trailing Stops but another method I use. The 200 ema can be used to decide the market condition and whether to activate the Trailing Stops in a Bull market or run on a single Bar in a Bear as the original program intended. I got kinda burnt out on doing TA on some other stuff. I quickly coded this several days ago and had discarded. At least this thread got my creative juices flowing again. bhda, thanks for starting. Back to lurkdom. --Jeff
Gentlemen, Would it be possible if someone familiar with this system posted rules in a plain english so even beginner like me can try to follow it ? Thanks.
10 period bollinger band based on daily lows. 1.5 standard deviation. Buy on a close below lower band Exit on any close greater than entry price or after 20 days