| Welcome Guest, please sign in to participate in a discussion. | Search | Active Topics | |
 Gold Customer
Joined: 10/19/2004
Posts: 56
|
Would you know of a PCF that would tell you a stock has been mostly sideways for let's says
the last 5 days? I'm having trouble writing it and was wondering if you guys might have one already.
|
|
Worden Trainer
Joined: 10/7/2004
Posts: 65,138
|
Not really. Any formula would depend on your specific definition of sideways.
-Bruce
Personal Criteria Formulas
TC2000 Support Articles
|
|
Registered User
Joined: 4/18/2005
Posts: 4,090
|
I've struggled with this allot. the best Ive been able to come up with is this, which was done for me by diceman.
((ABS(XAVGC3 - XAVGC8) + ABS(XAVGC5 - XAVGC8) + ABS(XAVGC10 - XAVGC8) + ABS(XAVGC12 - XAVGC8) + ABS(XAVGC15 - XAVGC8)) / XAVGC8) * 100
This finds stocks based on the spread between multiple moving averages. When all the fast moving averages are pressed tight togeatehr you probably have a consolidation.
ou can also plot the donchian chanels and sort by them or a formula similar to the above the measures the spread
((MAXH20 - MINL20) / ((H + L) * .5)) * 100
you could combine the two also.. Say run the Moving average spread scan and then with the chanels plotted on price sort by price compared to one of the chanels. This will likely bring up high bases or low bases.
Finaly you could plot bolinger bands and sort by those on price. Invert the list to see the narrowest spreads first and click through your list quickly
I personaly use the Donchian Channels and the movign average spread PCF. It can be put into scans and restricted or plotted as a custom indicator and sorted by. i actually put it in a scan with liquidity filters and TSV and MS criteria and then sort buy it and volume to quickly find the candidates i'm interested in.
|
|
Registered User
Joined: 1/28/2005
Posts: 6,049
|
You may be interested in this:
Thanks
diceman
|
|
Gold Customer
Joined: 10/19/2004
Posts: 56
|
Guys, thanks for your inputs. As you might have guessed, what I'm scanning for is range consolidation after a range expansion period, especially as price moves sideways into a moving average such as the 10 or 20SMA.
Such a condition, when it occurs as part of a larger pattern, such as the right side of a cup without handle pattern, often precedes another range expansion move. An example is APH (Amphenol) and BRP in the last few days.
Thanks again. I'll try your suggestions.
|
|
Registered User
Joined: 10/19/2007
Posts: 4
|
I entered the pfc's from above and am not getting sideways scans. I tried each separate and then an easyscan with both as conditions.
I'm new to all of this, so I'm not sure if it's something I'm doing wrong...
|
|
Registered User
Joined: 1/28/2005
Posts: 6,049
|
dclee123
The best way to do it is probably the video I linked
above. (unless you have a specific condition that an
exact scan would work better)
------------------------------------------------------------------------
Scott's scans above:
The first will be a low value when 3,5,8,10,12,15
exp. moving averages are close to each other.
(a period of low volatility)
It is more a test of low volatility than a "sideways scan".
The second is the max range over 20 days divided by
half its value. It will be lower when the stock has
not moved up or down as much over a 20 day period.
Hope that helps
Thanks
diceman
|
|
Registered User
Joined: 1/1/2005
Posts: 2,645
|
vasm,
Suppose we are interested in determining when Closing Prices move sideways for P days.
1) Please plot a line chart of Closing Prices on Logarithmic Scale. We wish to determine when the last P Closes resemble a horizontal straight line.
2) Please place a Linear Regression (LR) line of Period P on Closing Prices. We wish to determine when the LR line is near zero slope and last P Closes are near the LR line.
Calculate the LR Slope and bound it, i.e.:
ABS(LRSlope(LOG( C ),P)) <= S
Calculate the LR root-mean-square error and bound it, i.e.:
LRrms(LOG( C ),P) <= E
The constants S and E are chosen to satisfy what the individual believes is close enough to sideways movement of the Closing Prices for the last P days.
Note: The LOG( C ) was used so that the results would be insensitive to a scale factor change, i.e., for any a > 0:
LRSlope(LOG(a*C),P) = LRSlope(LOG( C ),P)
and:
LRrms(LOG(a*C),P) = LRrms(LOG( C ),P)
Thanks,
Jim Murphy
|
|
|
Guest-1 |