mahkoh
1,065 posts
msg #104853
- Ignore mahkoh |
2/6/2012 6:03:38 PM
It would be interesting to find out how many times Monday's low was below the open, and if so how much lower. If a large enough number of the samples prints lower prices during the first trading day it would pay to monitor them and enter when price reverses (or seems to reverse).
Especially in the lower volume stocks with large opening spreads you may see quick gapfills which are accounted for in the statistics but one would never have been able to actually make that trade.
|
Kevin_in_GA
4,599 posts
msg #104854
- Ignore Kevin_in_GA
modified |
2/6/2012 6:07:38 PM
Agreed. I think the key is to use a limit entry. The trick is to have it set up to do so quickly at the market open (underscoring the need for full automation).
|
Kevin_in_GA
4,599 posts
msg #104856
- Ignore Kevin_in_GA |
2/6/2012 9:51:29 PM
OK, now I'm getting irritated ...
I noticed over this weekend that the screener used by Stratasearch was pulling up different numbers of stocks for the gap up and gap down lists.
The trades called by the SS list today were COMPLETELY different from what SF generated. Note that both are doing OK, but the discrepency is baffling.
This means that the backtest data generated by SS will not necessarily translate to the performance from the SF picks. Caveat on this.
Issue could be with SS - I'll keep digging. In the mean time the three trades called today are still in good standing. One (PPO) has already closed, and AREX actually HIT the limit order but the filter needs it to EXCEED the limit to call it closed.
|
evo34
82 posts
msg #104857
- Ignore evo34 |
2/7/2012 1:40:56 AM
One issue with any strategy based on open prices is that all basic data providers (Yahoo, etc.) record the consolidated open, which is the first ECN trade at >=9:30am. This is fine for Nasdaq stocks (whcih always open on primary at 9:30), but can cause quite significant discrepancies for NYSE stocks. No easy solution here, other than getting data from Bloomberg or NYSE itself.
|
Kevin_in_GA
4,599 posts
msg #104864
- Ignore Kevin_in_GA |
2/7/2012 2:31:59 PM
Today AREX closed its gap, but as of last night it disappeared from the filter output. Must be based on finalized prices.
I did see one reason for the difference between SS and SF output. I had set the close above 2 in my SF filter, and had kept the original "close above 1" in Stratasearch.
Still more digging on why the systems trade different stocks though. More on this as I learn.
|
Kevin_in_GA
4,599 posts
msg #104922
- Ignore Kevin_in_GA |
2/11/2012 1:36:54 PM
Updated lists for 10 Feb. These are from stratasearch, not from SF.
GAP DOWN CANDIDATES
GAP UP CANDIDATES
I think the differences are coming from the definition of what stocks are in the Russell 2000. Neither site gives a full listing, and I can't tell which is more accurate (not that it really matters). The shared lists will be the UNION of these two sets, rather than the intersection. More potential candidates that way.
|
mmaurice
51 posts
msg #104925
- Ignore mmaurice |
2/11/2012 4:11:13 PM
How does the filter perform if you include all equities (maybe except etf) above some average volume?
|
Kevin_in_GA
4,599 posts
msg #104926
- Ignore Kevin_in_GA |
2/11/2012 4:58:38 PM
Tried it, with the ETFs as part of the mix. Did not perform as well, although I'm not sure why.
BTW, it still delivered 25-30% per year.
|
duke56468
683 posts
msg #104931
- Ignore duke56468 |
2/12/2012 10:09:35 AM
Kevin do you recommend we run the GAP review filters monday AM and again several times during the day, or how are you finding the gapers on monday?
|
Kevin_in_GA
4,599 posts
msg #104932
- Ignore Kevin_in_GA |
2/12/2012 10:49:59 AM
Since I can't update the sharedlists here on SF (no idea why), I would not run the filter here as written.
Create two new personal watchlists as follows:
Gap_Down:
GLUU IO EMKR RBN ZIXI RTK JRN NOG QPSA CRIS FTK FVE DYAX MELA MBLX CVI FSII HEAT KOG GBX GPOR AIS AVNR SNMX MTZ NCT COMV BLDR GMXR FIRE AREX FXEN GEOI XTXI LPX FARO OPWV TWI SMP OCLR SPAR PANL ETM FNSR ATPG VOG TPLM NEWP BEE HLX ARII ICGE UPI HL ROC WWWW BONT RRR DSCO FLOW RPRX CPHD BSFT CNVO SLTM OCZ BRY LTM MHR NOR PKT MWA GLNG PIP MCHX HERO MOVE
and Gap_Up:
ZOLT ZIOP ZAGG YRCW WWWW WOLF WAVX VVTV VTG VOXX VOG VHC VDSI UTSI URG UNIS UEC TZOO TUES TSYS TRGL TPLM TPCG TLB SVNT SURW STEC STAA SPRT SPPI SPF SPAR SNTA SNMX SKY SCSS SBCF RTK RTEC ROCK RLOC RES RDN RBCN RAS QTM QPSA PQ PPO PNX PNSN PIP PANL OXGN OVTI OSG OREX OPXT OPK ONTY OMX OMEX OCLR NVAX NSPH NNBR MTG MOTR MNTA MNKD MNI MITI MHR MGAM MELA MED MEA MAKO LZB LXRX LPHI LNG LLNW LLEN LIOX KV.A KOG KEM JAG IRBT IPGP INSM HOV HOGS HNSN HNI HEAT HDY HAFC GTIV GSM GSI GPK GORO GNK GMXR GMO GKK GEOI GBX FXEN FOE FNSR FLOW FCEL FBN EXEL EVC EPL ENOC ENER END EGLE DYAX DSCO DHT DCTH CYTX CYTK CVO CTIC CRIS CONN COMV COCO CLNE CGA CERS CBLI CBK CAS BZH BTX BPZ BDSI BANR AVII ATEC ARUN ARII ANTH AMSC ALIM AHS ACLS ACHN AAWW
These are the candidates for Monday. Just change the syntax to "apply to watchlist(XXX)" instead of sharedlist.
SF seems to really be dropping the ball a lot lately on their "improvements". Really beginning to frustrate me that they focus on this stuff instead of improving the code. Despite hundreds of requests to do so, they instead try new stuff nobody had asked for and then screw it up.
K
|