Inspired by others' contributions to the site, I'd like to open a new topical thread.
A strategy that has been profitable for me recently relies on finding stocks that have gapped down a significant amount.. some of these tend to bounce back a good few points, and some of them continue their downward voyage.
While there exist online tools and scans that will find gaps like these, they only work on today's data. I've found that the gap-down stocks that bounce back usually take a few days to recover from the drop before the price begins to climb. By then, of course, such a stock won't be caught on a gap scanner (like on stockcharts.com)
So I've clobbered together a solution that finds past gaps in historical data.. I've limited it to only hard gap-downs of 10% or more (ie, where the low of previous day is at least 10% higher than the high of the next day), and filtered out all stocks below a dollar (most of these are all gaps anyway..)
Mind you, this is not a load-and-fire tool.. only a few of these stocks bounce back; further analysis should help pick these from the list. Sometimes a significant gap becomes a resistance level -- it usually takes longer than a bounce-back, but upon breakout there's more confidence in a continued bullish movement.
Anyway, to illustrate the concept, here's a short list: stocks that have gapped down more than 10% in October, and closed between $1 and $30 yesterday.
NOVN NPSP CWTR NKTR JADE
If you find any stocks that have taken a big drop and look ready to pounce back up, post them here. I will update this list as I refine the tool, and as new gaps occur.
-------------------- the market is not your mother Posts: 1282 | From: Montreal | Registered: Mar 2006
| IP: Logged |
CRGN...Huge takedown after company discontinued the cancer drug Velafermin...Seems to have found bottom here...Been accumulating under $1.00 awaiting news about their other cancer drug Belinostat...
posted
Glad to see some others trying to find some good picking methods besides "Hey this one is being pumped on site XYZ!"
What do you use to pull your data? How far back do you pull/store data? How often will you run it?
Posts: 5508 | From: Southeastern PA | Registered: Jan 2006
| IP: Logged |
posted
Oh and as far as stocks that fit the bill. LVLT. I know nothing about it other than a friend at work saying this morning "F***ing LVLT. B*stards missed their targets and it gapped down like 20%. D*mn it all!" Posts: 5508 | From: Southeastern PA | Registered: Jan 2006
| IP: Logged |
posted
I've got an end-of-day feed for all American markets, about 5 years of reliable historical data. I've written a server-system to manage and analyse the data. The gap finder is another addition to the rickety collection of code..
I intend to run the tool every night and post the results for scrutiny. Hopefully we can make some mulah off a few stocks it finds..
-------------------- the market is not your mother Posts: 1282 | From: Montreal | Registered: Mar 2006
| IP: Logged |
posted
Cool. I do everything in Excel, so it's pretty inefficient, i think. I just kicked off a run to see if I agree with your list, or if I have more to add based on your criteria. Looks like ti will take abotu 1/2 hour to pull all October data, and then should take somethign around that to anlyze it.
how long does your stuff take?
Posts: 5508 | From: Southeastern PA | Registered: Jan 2006
| IP: Logged |
posted
very nice...this is the kinda plays i like...i will also try to contribute to the gappers
-------------------- If your not on the edge of ur seat, sell it! Everything posted is my opinion! Posts: 992 | From: H-town | Registered: Nov 2006
| IP: Logged |
posted
HSWI rebounding off gapper, but dont no targets or much about it at all..up 11% today
-------------------- If your not on the edge of ur seat, sell it! Everything posted is my opinion! Posts: 992 | From: H-town | Registered: Nov 2006
| IP: Logged |
quote:Originally posted by PCola77: Cool. I do everything in Excel, so it's pretty inefficient, i think. I just kicked off a run to see if I agree with your list, or if I have more to add based on your criteria. Looks like ti will take abotu 1/2 hour to pull all October data, and then should take somethign around that to anlyze it.
how long does your stuff take?
Interesting.. cross checking the gappers with form-4 filings? This would be to catch when insiders load up on suddenly cheap stocks? I like...
The scan itself takes about five minutes, but then there's the manual labor part, walking through the results and throwing out mismatched ones.
One thing I didn't think of was volume -- gaps are more significant if they happen on strong volume. I'll have to code that in..
-------------------- the market is not your mother Posts: 1282 | From: Montreal | Registered: Mar 2006
| IP: Logged |
posted
I wasn't refering to the form 4 program, but a TA program that I wrote and posted about last night in another thread.
I just kicked off the analysis on my 2nd computer for the last 15 years just to get a feel for how they perform without looking at anything other than price and gap down. I'll post the results when they finish.
quote:Originally posted by fourseven:
quote:Originally posted by PCola77: Cool. I do everything in Excel, so it's pretty inefficient, i think. I just kicked off a run to see if I agree with your list, or if I have more to add based on your criteria. Looks like ti will take abotu 1/2 hour to pull all October data, and then should take somethign around that to anlyze it.
how long does your stuff take?
Interesting.. cross checking the gappers with form-4 filings? This would be to catch when insiders load up on suddenly cheap stocks? I like...
The scan itself takes about five minutes, but then there's the manual labor part, walking through the results and throwing out mismatched ones.
One thing I didn't think of was volume -- gaps are more significant if they happen on strong volume. I'll have to code that in..
Posts: 5508 | From: Southeastern PA | Registered: Jan 2006
| IP: Logged |
posted
Ah, I see. If you put this into your backtester, it will generate mostly horrible losses -- most gap downs are for a reason, and the stock continues falling..
-------------------- the market is not your mother Posts: 1282 | From: Montreal | Registered: Mar 2006
| IP: Logged |
posted
Actually, it's doing shockingly well. After running through about 700 stocks so far it's returned an average of about 5% per month for the last 15 years.
Downside is that you would be making an average of like 16 trades per day so you' have to have a sh*tload of available capital.
Maybe I'll play around tonight and try to pare it down some without losing any of the gains.
quote:Originally posted by fourseven: Ah, I see. If you put this into your backtester, it will generate mostly horrible losses -- most gap downs are for a reason, and the stock continues falling..
Posts: 5508 | From: Southeastern PA | Registered: Jan 2006
| IP: Logged |
posted
5% per month? That's suspiciously high.. what are you using as exit signals?
I don't think the gap strategy as presented here is a winning one all by itself. It's just a filter to catch stocks where bears have expended considerable resources. It would make sense that the bulls would take over, if just for a short while. So it pushes the odds in our favor a bit.
-------------------- the market is not your mother Posts: 1282 | From: Montreal | Registered: Mar 2006
| IP: Logged |
posted
A simple refinement to the gap-down filter: in addition to previous day's low being at least (p%) higher than the following day's high, the previous day's volume must be (q%) less than the next day's volume.
The rationale here is that the gap-down is more significant if it is made up of a larger-than-average amount of trades.
-------------------- the market is not your mother Posts: 1282 | From: Montreal | Registered: Mar 2006
| IP: Logged |
posted
BHIP...Have not taken a position but have this one on my radar...Big gap down on Monday due to private placement financing PR...Small float, should bounce well once it finds bottom...
posted
Final tally for stocks with symbols starting with A-L (long story why I had to do it that way, but it shouldn't make much difference):
Overall over the 15 years, 22,705 buys. If you sold after 1 month you'd end up with 4.1% per trade (so 4% per month since you'd hold each for 1 month).
Looked extraordinarily promising, but I then looked at it by year.
So what this means is basically if you had done this in 1992 you would have gained almost 325% for the year, but the last few years it has become a much less exciting predictor.
Like I said, I will play around tonight with some other cuts of the data (for example your volume test) and see if we can't get this to work better over the last few years.
quote:Originally posted by fourseven: 5% per month? That's suspiciously high.. what are you using as exit signals?
I don't think the gap strategy as presented here is a winning one all by itself. It's just a filter to catch stocks where bears have expended considerable resources. It would make sense that the bulls would take over, if just for a short while. So it pushes the odds in our favor a bit.
Posts: 5508 | From: Southeastern PA | Registered: Jan 2006
| IP: Logged |
posted
I just re-read your post. Seems you and I were talking about slightly different gaps. I was testing only the open relative to previous day's close.
Posts: 5508 | From: Southeastern PA | Registered: Jan 2006
| IP: Logged |
posted
ok, i've modified the scanner so that it picks up gaps for which this is true:
-> more than a 10% difference between previous low and current high -> volume on gap day at least 20% higher than previous day -> stocks that have traded 97.5% of trading days in the last year (another words, no sparsely-traded stocks)
with that, i searched for stocks that closed between 0.50 and 6.00 today
i had to throw out a couple noisy ones, but the rest made it into these three lists:
1. gaps that could've worked if played BRLC, EFJI (twice), FMT, CMGI, THLD, COR, ANX
2. gaps that would've failed if played EFJI (once), HEPI, COOL, FNSR, WTSLA, AGIX, SOIGF, NIS, BJCT, COR, TES, HORT
that a gap "worked" means here loosely that in retrospect, the chart looks like there was an opportunity for a short-term bullish trade during the upswing that followed.
some stocks made it to more than one list, because they've had more than one gap over the last twelve months
out of the potential ones, some are kinda sucky.. my own personal likes, narrows it down to NPSP, MEMY, ANX, TTDS, LVLT and DCTH (the last two are from today)
note that there were almost twice as many gaps that haven't worked than ones that did, in the stocks that i found over the last year.
-------------------- the market is not your mother Posts: 1282 | From: Montreal | Registered: Mar 2006
| IP: Logged |
posted
Nice job. This is the kind of thing I've been doing in my "free time" for a while now. Certainly can't hurt to have another person looking at it from a whole different perspective.
A few comments though. Your 20% higher than previous day volume doesn't seem like it makes much sense to me on it's face, given how much volume moves from one day to the next. On the hard gaps that I've seen, volume is more like 10X the previous 10day moving average.
While it's not a bad idea to look at these individually, the firms that I've spoken with stress a method that you could tell someone else and they could easily reproduce your results. If your goal is personal trading, then no big deal, but if you're interested in being some kind of money manager with a fund or anythign in the long run, keep that in mind.
Try to set a hard and fast sell rule. I can't tell you how frustrated I've become in doing this when I see something that looks like it works, but only when I use hindsight to set my sells. It's fairly easy to find a good buy point, and infinitely harder to consistently sell at the right time, or often even a profitable time, despite how a chart looks like you could easily have cashed out a good gain.
Posts: 5508 | From: Southeastern PA | Registered: Jan 2006
| IP: Logged |
posted
I'm a bit confused by the 20% parameter, also...
seems to me, (previous) market cap should be factored in... in other words, I'd be interested in one that peeps can clearly see that the market-panic has simply gone too far.
That would rule out many bio/pharmas, I would think, because...who the heck knows? Know what I mean? "FDA approval not granted" etc... OK...what's the real value, then?
Anyway, I remember a good example from a few months ago: Severe panic following failed-merger PR (as I recall), but anybody who bothered could see the main company was worth way more than the "panic price."
yeah, hindsight would be useful if we had time travel.. um.. yeah. a strategy only gives you an edge, or saves effort.. doesn't guarantee anything.
this is totally for personal trading.. i've been building this system to help out with my trades, and finally something that looks useful is coming out as a byproduct.
good call with the backtesting.. the way my system is set up, it shouldn't be too hard to put in some tests..
-------------------- the market is not your mother Posts: 1282 | From: Montreal | Registered: Mar 2006
| IP: Logged |
posted
Hey fourseven, when did JADE gap down 10%? I ran my program a few minutes ago to try to reproduce your initial list, and came up with a very different list.
Here is mine with the conditions of $1-$30
AMCE CWTR EFJI EXAS ISLN JZC
Did you have some more conditions that you didn't mention?
Edit: just realized I had only run the data for the A-L. that's why I missed your "N" symbols. I'm running the second half now
Posts: 5508 | From: Southeastern PA | Registered: Jan 2006
| IP: Logged |
the date listed, is that the date of the gap? What is the number listed under drop? they all seem to be less than 1%.
Like NANX is listed on 10/19, but i don't see a gap in NANX at all in October.Also, can you opint me to the gap for JADE? I don't see one in October.
Posts: 5508 | From: Southeastern PA | Registered: Jan 2006
| IP: Logged |
The date is gap date, close is the closing price on gap day, drop is how many % it dropped (close-to-open, just to be confusing), vol is volume ratio (gap to pre-gap)
NANX gapped down on earnings on the 19th, JADE on the 18th
-------------------- the market is not your mother Posts: 1282 | From: Montreal | Registered: Mar 2006
| IP: Logged |
UBBFriend: Email this page to someone!
Printer-friendly view of this topic
![[Smile]](smile.gif)
![[Wink]](wink.gif)
![[Big Grin]](biggrin.gif)
![[Cool]](cool.gif)
