StockFetcher Forums · Filter Exchange · HOW TO DESIGN A SYSTEM (NOT JUST A FILTER)<< 1 ... 14 15 16 17 18 ... 43 >>Post Follow-up
duke56468
683 posts
msg #101923
Ignore duke56468
8/1/2011 6:50:52 PM

Marco...The problem occurs in trying to predict market direction. The backtest just buys at the open. Trying to buy at open today based on the futures was a bad choice for me so far.

fortyfour
189 posts
msg #101924
Ignore fortyfour
8/1/2011 7:35:57 PM

I intend to test this "divergence" in up/ flat and down cycles...

Intuition and beliefs are more often wrong but it
may be that " after VXX / volatility has stopped rising..... the picks, (Zscore16 < -2) from days ago
that have continued to make new lows may perform better than todays Zscore16 < -2"

Some may even double down on thsoe that contionued to make new lows.

I mispoke to call this a "system"...I was referring to Kevins system.

And of course all may make new lows tommorrow and the next day etc....

mahkoh
1,065 posts
msg #101926
Ignore mahkoh
8/2/2011 4:47:22 AM

The z-score measures the dislocation of the stock versus SPX with the idea being that the stock will reverse to the mean (zscore=0). Like a rubber band: the lower the zscore, the more stretched the rubber band is. A zscore rising does indicate that the reversion has begun, but remember that you are dealing with two variables, the stock and SPX.

Duke, as an example CVH was selected on Friday. If you bought at the open at 33.83 you are down 6% already. After 10:00 it started moving up and you could enter around Friday's close near 32. Even better would be to wait until it starts trading above the 20 ma on the 5 min chart which would have you entered in the 31.50 area. If it runs from the open don't chase, try the second on the list or wait until the next signal.

duke56468
683 posts
msg #101927
Ignore duke56468
8/2/2011 10:07:48 AM

I agree mahkoh... and I do use the 5min chart and the ma(20) in trading just like you describe. Problem is there is no backtest to tell us that that will work better than buy at the open for this filter,

mahkoh
1,065 posts
msg #101936
Ignore mahkoh
modified
8/2/2011 5:40:21 PM

True, but you'd get each stock the system generated and that went lower after the open at a better price. I believe the odds are in your favor to outperform the system.

duke56468
683 posts
msg #101938
Ignore duke56468
8/2/2011 6:53:19 PM

You can tell the back test to only enter if below close with

set{ep, close yesterday - 0.01}
ep

but there is no way to trail it down like you might try in real life.

novacane32000
331 posts
msg #101942
Ignore novacane32000
8/2/2011 9:26:41 PM

Mahkoh-Your best bet to get a better price is simply to put in a .5% or 1% limit order at the open . Problem is you will miss out on stocks that take off from the open. My backtest show about 1 in 7 picks will not hit the limit order and will instead take off from the open for a big gain .

Average gain on those stocks is about 5%

Eman93
4,750 posts
msg #101959
Ignore Eman93
8/3/2011 2:45:34 AM

Duke never buy on a big gap up... most of the time it a fade.

if it is a gap and go .. like RDN today .. you wait for the first 5 min bar then trade long above that with a stop right below the low of that first bar... it has helped me. you have a plan for the open.

Get charts that show premarket data. you want to make sure you break above any overnight of pre market high.

Gap downs are the best to buy the open..

GLT.

duke56468
683 posts
msg #101962
Ignore duke56468
8/3/2011 9:35:55 AM

I agree.. thanks Eman

Kevin_in_GA
4,599 posts
msg #102015
Ignore Kevin_in_GA
8/5/2011 9:31:56 AM

Been on vacation (still am). After yesterday's 513 point drop I doubt that any long only system looks good right now. Makes me think that this system might benefit from incorporating a market-based metric as well. I'll look into this when i am back. My initial thought will be around adding a rate of change element ( roc(X,Y) above 0) so that you stay out of down periods entirely. might reduce overall return but should add in some protection aginst prolonged drops.

Also, it occurred to me that in using ^SPX in the Z-score calculation instead of SPY, you will always be using YESTERDAY'S ^SPX value until after the market closes (since it is not updated until 5:30 PM or so). Just a caveat for intraday traders ...

StockFetcher Forums · Filter Exchange · HOW TO DESIGN A SYSTEM (NOT JUST A FILTER)<< 1 ... 14 15 16 17 18 ... 43 >>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 2022 - 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


This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.