Kevin_in_GA
4,599 posts
msg #114909
- Ignore Kevin_in_GA |
8/16/2013 9:00:07 AM
Just put the following text (exactly as written here) into the selection method box:
/*FIRST DETERMINE HISTORICAL RATIO OF S&P STOCK TO THE SPY OVER THE LAST 16 DAYS*/
SET{PRICERATIO, CLOSE / IND(^SPX,CLOSE)}
SET{RATIOMA16, CMA(PRICERATIO,16)}
SET{RATIOSTD16, CSTDDEV(PRICERATIO,16)}
SET{DIFF16, PRICERATIO - RATIOMA16}
SET{ZSCORE16, DIFF16 / RATIOSTD16}
SET{THRESHOLD16, RATIOSTD16 * 2}
zscore16 above -1
select by zscore16 ascending
If you want to use this intraday, you'll need to replace ^SPX with SPY.
|
tonyctl
7 posts
msg #114922
- Ignore tonyctl |
8/16/2013 1:00:38 PM
This is excellent info guys, thanks! Pardon my stupidity, but where is the selection method box you are referring to, Kevin? I can't seem to find it in the Backtesting settings. Also, I have tried replacing SPX with IXIC and generated rather interesting results :)
|
tonyctl
7 posts
msg #114925
- Ignore tonyctl |
8/16/2013 3:31:14 PM
NVM, just found it, thanks!
|
tonyctl
7 posts
msg #114926
- Ignore tonyctl |
8/16/2013 3:38:12 PM
"/*FIRST DETERMINE HISTORICAL RATIO OF S&P STOCK TO THE SPY OVER THE LAST 16 DAYS*/
SET{PRICERATIO, CLOSE / IND(^SPX,CLOSE)}
SET{RATIOMA16, CMA(PRICERATIO,16)}
SET{RATIOSTD16, CSTDDEV(PRICERATIO,16)}
SET{DIFF16, PRICERATIO - RATIOMA16}
SET{ZSCORE16, DIFF16 / RATIOSTD16}
SET{THRESHOLD16, RATIOSTD16 * 2}
zscore16 above -1
select by zscore16 ascending "
But shouldn't it say:
zscore16 below -2
select by zscore ascending
for selection criteria when entering a stock? I thought zscore above -1 is an exit criteria?
|
Kevin_in_GA
4,599 posts
msg #114927
- Ignore Kevin_in_GA |
8/16/2013 3:43:29 PM
You're partly right, and I'm partly wrong. That criterion is for the exit, but is ignored in the selection method box. That line can be deleted since it is not used.
|
tonyctl
7 posts
msg #114929
- Ignore tonyctl |
8/16/2013 3:55:03 PM
Thanks Kevin for clarifying. Would you recommend applying this filter as is to the Nasdaq index? I tried backtesting as an experiment, and added a Global filter of volume > 100000 to exclude stocks that are not very liquid and the results came out even better than S&P. Would optimization by stratasys be required once again since it is a different index we're applying it to? Thanks.
|
mdl060374
94 posts
msg #115559
- Ignore mdl060374 |
9/26/2013 7:24:33 PM
Kevin,
Do you have any sort of advice with a similar search for daytrading candidates?
It seems that a lot of these studies are more for swing/longer term trading.
Any advice you have for intraday would be great.
|
Kevin_in_GA
4,599 posts
msg #115630
- Ignore Kevin_in_GA |
9/30/2013 2:45:36 PM
There's no reason this basic approach wouldn't also work with intraday data. I have not tested it that way, and would probably look at a fair amount more optimization before trading anything. The challenge will be that the intraday moves will be of smaller magnitude while the commissions will stay fixed.
|
four
5,087 posts
msg #115832
- Ignore four
modified |
10/11/2013 10:10:34 AM
http://www.amazon.com/Design-Testing-Optimization-Trading-Systems/dp/0471554464/ref=cm_rdp_product_img#reader_0471554464
Design, Testing, and Optimization of Trading Systems [Hardcover]
Robert Pardo (Author)
4.3 out of 5 stars See all reviews (12 customer reviews)
PS Before you ask... just found this... haven't read the book
|
Kevin_in_GA
4,599 posts
msg #115837
- Ignore Kevin_in_GA |
10/11/2013 1:11:11 PM
I downloaded this one a while back - not bad. Mostly focuses on using Monte Carlo analyses, and walk-forward analysis to validate trading systems against OOS data.
If that doesn't make sense to folks reading this then they probably should get the book.
|