Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Strategies & Market Trends : Technical Analysis- Indicators & Systems -- Ignore unavailable to you. Want to Upgrade?

To: Bob Anderton who wrote (3203)	12/6/1998 4:10:00 AM
From: Richard Estes	Read Replies (1) \| Respond to of 3325

I will try to drop in and talk system testing. But the first thing you need to know is: System Tests do not model reality, their prime use is to compare the relative differences between systems. No, we don't continue to invest the profits back into the same stock in real life or buy/sell at open. But since all systems are subject to the same measures, it is comparable. You are trying to find the best system that meets your mindset. Humans don't maximise profits, they optimise them in keeping with mindset. I responsed to Peter's post with a little tounge in cheek, thinking that he may have made an error in his computations. That type of return would beat all the supposed "established firms". The first item, I look at is avg length of trade.. we all have an idea of how long we could stay in a trade. Next you want to know about drawdowns, how low would my capital go during a trade can I live with this? MSWIN has MAE measure of the lowest point during a trade that the stock went down to during the trade. Its late. WOWPro allows testing on muti-stocks. Mswin requires you to test one at a time, but has better reporting. 5 years of data is minimum amount for a good test. Always use the same amount of money, and other settings for all tests for comparsion. I suggest 10,000 and buy/sell next day at open after signal. Write a book, if handled right you can sell 500,000 copies or more. Assuming 10% return on a $50 book, you get 2.5M for a few weeks work. That should help your trading capital. I would give the 90% to the people that gets the numbers of sales in that range gladly.

To: Bob Anderton who wrote (3203)	12/6/1998 1:07:00 PM
From: peter n matzke	Read Replies (1) \| Respond to of 3325

i tested all the of the systems on a base of 63 stocks, 127 stocks and 2500 stocks. The three systems previously mentioned outperformed in all bases. i used a $5000 base investment, $25 cost per trade. no stops, trade occurs on the day following the signal. original $200,000 base investment in 1989 in 40 members of the fortune 500 23 additional stocks were added in the following 9 years putting the maximum investment at a total of $315,000 the systems added: GTSG in 1998 CIEN,QWST,AMFM,AMZN,network associates in 1997 SEEK,XCIT,LU,YHOO,and lycos in 1996 NSCP and TBR in 1995 ASND in 1994 GTW, and morgan stanley dean witter in 1993 PSFT,AOL,SBUX,CPWR in 1992 Winstar in 1991 CSCO, and COST in 1990 members of the group that performed poorly were Drug, Oil, and Airline stocks such as LLY,PFE,BMY,AMR, Mobil etc. so there were some under performing stocks within the model. only one or two of the stocks were the top performer when compared individually against other system tests. But the models performed consistently better than any other models tested. regards peter