StockFetcher Forums · General Discussion · back testing<< >>Post Follow-up
shelupinin
120 posts
msg #34755
Ignore shelupinin
1/5/2005 2:59:43 AM

Hi, guys!
I'd like to ask about "back testing".. I'm not satisfied with "back testing" technique wich used on SF, on SF I can only "back test" filter for one given period of time. For example: I need to test filter
Fetcher[RSI(15) is below 30 and RSI(10) is below 30 and RSI(5) is below 30 and price dropped more than 15% for 1 day and price above 1 and volume above 20000 and data offset is 5]

, at SF I can only test it for last 5 days, but if I want to check filter performance
one month ago I can do it only manually. I'd like to find possibility to "scan" whole year with this filter, calculate performance, then divide value for number of days and get average performance for given filter for given data offset for last year. Is there any possibility to do it with StockFetcher?
If no, may be someone knows some web service with stock database which allows to write scripts with querys like "scan last 365 days for filter "bla-bla-bla" with data offset 5 days then divide value for 365" ? If it is possible some way then it is a great tool to check real average filter performace, to compare filters and find best one not only for last 5 days, but to get average value with several hundreds or even thousands querys !!!
any suggestions welcome
thanx in advance
Alex


cegis
235 posts
msg #34759
Ignore cegis
1/5/2005 11:59:17 AM

Perhaps this thread will help: Technique: Calc your own returns - for any period.

I just added to it, so it's at (near) the top of the list of threads in this forum.

HTH,

C


yepher
353 posts
msg #34765
Ignore yepher
1/5/2005 10:06:45 PM

search the forum for "G5 geodepe" I believe there is a tool there that will use your filter and compute what a typical gain is.

Beyond that site you could have a look at MetaStock if you want to do more relistic back-testing for a given trading strategy.

If you happen to be technically inclined you should have a look at TA-LIB.

If these solutions don't work for you then send me an email at sf_backtest@yepher.com and maybe I can help you out.


shelupinin
120 posts
msg #34775
Ignore shelupinin
1/6/2005 6:10:47 AM

cegis, yepher, thank you for answer but I still did'nt get results :(
the ipropel.com link at "G5 geodepe" post not working, and the "Technique: Calc your own returns - for any period" not just what I wanted to ask... sorry, correct me if I wrong, but "Technique: Calc your own returns - for any period" doesn't allow you to calculate average filter results for, say, last 1 or 2 years, I mean not a data offset for 1 year, data offset should be 5 days, I mean to get average value of performance for some filted for 5 days period during last 1-2 years, so I need to calculate filter performance for every day of last couple of years with data offset 5 days... Is it possible to do automatically? I guess there should be some "web based" service with stock price database for last several years, with possibility to write some querys.
I know that software like MetaStock allow you to do it, but I don't have MetaStock, I'm just beginner...
if you have any ideas how to do it, please, post..
thanx in advance
Alex


cegis
235 posts
msg #34783
Ignore cegis
1/6/2005 4:32:44 PM

Alex,

My technique allows you to get the filter's results as of a specific date, then calculate returns for the next 5 days (or whatever) within the filter results. If you want the sum of the results for the past year or two, you'd have to run the resulting filter for the time period you want, and download each run into a spreadsheet. I know it's a pain, but that's the best you can do (that I'm aware of) using SF...

HTH,

C


knowsenough
54 posts
msg #34784
Ignore knowsenough
1/6/2005 5:20:44 PM

Have a question on backtest time frame limit...
I keep reading about all this belly-aching over SF's backtest feature being limited to only 5 days offset? I get correct performance results on 30/60 so I don't understand the problem. Take TAGS for instance. Comes up on my screens (60) day backtest @ .79, and if I count it back on the chart that's exactly right, and so is the 200% plus gain. Am I reading this wrong?


shelupinin
120 posts
msg #34833
Ignore shelupinin
1/10/2005 11:52:09 AM

Hi, again!
As I understood from posts it is impossible to do automatic back testing for some filter with a lot of querys for some period of time with SF... So may be some one know what is a simples solution(I mean cheapest one :) to back test some filter automatically? I mean I need some stock price database with history for 1-2 years with ability to write some querys like "scan last 365 days for filter "bla-bla-bla" with data offset 5 days then divide value for 365"...
It's good that SF allows to "back test" filters, but it's much better if we can
test filter performance not with one-two querys but with hundreds querys, to get real avarage volume for filter performance...

any suggestions welcome
thanx in advance
Alex


StockFetcher Forums · General Discussion · back testing<< >>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