StockFetcher Forums · General Discussion · Genetic Programming to find out the best filters possible<< >>Post Follow-up
cocky_pusher
12 posts
msg #36746
Ignore cocky_pusher
7/6/2005 2:31:24 PM

Dear filter community at SF, it is so cool to have these forums to talk to people who are into financial modeling. I have been working with financial algorithms for some years already, most of them adaptative momentum algorithms and recently I finished my work on Genetic Programming using technical analysis indicators on the NASDAQ 100 shares. And since everyone shares filter and trading knowledge in here I want to invite you to check the research Iíve done. Donít worry this is not a black box, neither a gray one, pure financial modeling research. I hope you find it interesting and that you post your opinions.

www.tau-zweig.com



Koronbock
201 posts
msg #36749
Ignore Koronbock
7/6/2005 3:46:05 PM

Cocky,
I looked under the link you provided and have taken a quick look at the material. You have done some good work there. BUT it is my impression that this whole approach smells an awful lot like CURVE FITTING ( I may be wrong). After being in the business of buying and selling stocks for more than 20 years I have come across some mighty looking "grand schemes" which all blew up in real time. The market has a way of frustrating all attempts for "get rich quickly" approaches.




cocky_pusher
12 posts
msg #36750
Ignore cocky_pusher
7/6/2005 4:13:30 PM

Koronbock thnx, Genetic Programming is a way of finding solutions to big problems, thru past data processing. So curve fitting may somehow ring a bell, but the thing is that the process result; are the best solutions possible. In this case stock buying filters, this is no get rich quick scheme, only an approach to finding the best solutions available with technical analysis indicators. That same idea could be used with SF database and would surely find some incredible stuff. Anyway thnx for looking.


TheRumpledOne
6,358 posts
msg #36751
Ignore TheRumpledOne
7/6/2005 4:27:40 PM

Very interesting... Very cool.

Any chance you can include RSI(2)?




scvidar
51 posts
msg #36754
Ignore scvidar
7/6/2005 5:49:00 PM

Look very interesting. Sounded somewhat like a book I read called The New Technical Trader by Tushar S. Chande & Stanley Kroll. It was written 94, but alot of the new indicators at that time he came up with were similiar.

Thanks for the post


cocky_pusher
12 posts
msg #36778
Ignore cocky_pusher
7/7/2005 6:02:49 PM

Performed a simulation introducing rsi(2) into the variables, and another using rsi(2) only, the results are here:

http://www.tau-zweig.com/rsi2.html



TheRumpledOne
6,358 posts
msg #36780
Ignore TheRumpledOne
7/7/2005 9:30:18 PM

Am I reading this right?

GENERATION:2 *VALUE:0.920388 ACCURACY:0.523240 GAINS:1.706045 AVERAGE_GAINS:1.012911 WIN/LOSS_RATIO:86/100

GENERATION:99 *VALUE:1.275498 ACCURACY:0.583431 GAINS:2.107434 AVERAGE_GAINS:1.024192 WIN/LOSS_RATIO:97/100

Use RSI(2) and you win from 86 - 97 times out of 100!??!!

Wait till the StockHolyGrail sees this!





cocky_pusher
12 posts
msg #36794
Ignore cocky_pusher
7/8/2005 4:10:20 PM

The win/loss ratio is how many stocks in the NASDAQ 100 where the filter did perform did the filter make money instead of loosing.
86 of the 100 in 14 using that particular algorithm it did not make money. But the times it did can be read in the ACCURACY result.

I really think that Holygrail uses rsi2 in a different way, which my program could not find because it does not compare variables to constants so rsi(2) ( variable ) < 1 ( constant ) canít arise as a solution in the process of my program.



StockFetcher Forums · General Discussion · Genetic Programming to find out the best filters possible<< >>Post Follow-up

*** Disclaimer *** StockFetcher.com does not endorse or suggest any of the securities which are returned in any of the searches or filters. They are provided purely for informational and research purposes. StockFetcher.com does not recommend particular securities. StockFetcher.com, Vestyl Software, L.L.C. and involved content providers shall not be liable for any errors or delays in the content, or for any actions taken based on the content.


Copyright 2016 - Vestyl Software L.L.C.Terms of Service | License | Questions or comments? Contact Us
EOD Data sources: DDFPlus & CSI Data Quotes delayed during active market hours. Delay times are at least 15 mins for NASDAQ, 20 mins for NYSE and Amex. Delayed intraday data provided by DDFPlus