Registered User
Joined: 1/28/2005
Posts: 6,049
|
In the most recent Worden report. Sir-Fan-Pattern
highlighted the value of time above a 50 day moving average.
----------------------------------------------------------------------
This can be checked in 3 basic ways.
A true false condition:
C>AVGC50ANDC1>AVGC50.1ANDC2>AVGC50.2
this will be true if the last 3 days are above.
You can carry it out as far as you like.
--------------------------------------------------------------------
We can also do a count:
ABS(C>AVGC50)+ABS(C1>AVGC50.1)+ABS(C2>AVGC50.2)
This will equal 3 if all three days are above.
2 if 2 are.
1 if 1 is.
0 if none are
You can carry it out as far as you like.
-------------------------------------------------------------------
I believe the most flexible way is with a sort.
Goto:
Chart Template
Add Indicator
Indicator
Middle (where you plot it is not important)
To test for the last 60 days
(mentioned by sir fan)
Create the indicator:
ABS(C>AVGC50)*60
Set the "smoothing average" to 60 simple
Click on the indicator and select sort by
actual value.
(make sure your watchlist has a "sort" column in it.)
The value you see is how many days the close
has been above the moving average.
In the Nasdaq 100 currently 4 stocks have been
above for 60 days:
ISRG
JNPR
BIIB
GRMN
----------------------------------------------------------
To change the length you want to look at.
Change to 60 in the equation and its
smoothing value.
(they should always match and smoothing
should always be simple)
So for 50 days you would use:
ABS(C>AVGC50)*50
(with a smoothing average of 50 simple)
--------------------------------------------------------------------
This equation uses a simple 50 day average.
If you want exponential you should use:
ABS(C>XAVGC50)*60
(the smoothing average in the indicator
will still be 60 simple)
-------------------------------------
Thanks
diceman
|
|
Registered User
Joined: 12/21/2004
Posts: 902
|
Along the same lines, I did some experimenting awhile back with the idea of the time MACD (not MACD-H) is above the zero line as a sign of strength and got some good results.
I made up a True/False PCF that tests for MACD over 0 for the last 45 days (in this market, that should be pretty good). Surprisingly there are 10 in the Nas100 - half of them are semiconducter or networking devices.
NVDA
BIIB
AMZN
JNPR
KLAC
SNDK
CSCO
ISRG
FAST
GRMN
Here's the formula for 12,26,9 MACD over 0 for 45 days - it can obviously be tweaked for longer or shorter periods - or maybe Diceman can tell us how to do an indicator sort.
((XAVGC12 - XAVGC26) > 0) AND ((XAVGC12.1 - XAVGC26.1) > 0) AND ((XAVGC12.2 - XAVGC26.2) > 0) AND ((XAVGC12.3 - XAVGC26.3) > 0) AND ((XAVGC12.4 - XAVGC26.4) > 0) AND ((XAVGC12.5 - XAVGC26.5) > 0) AND ((XAVGC12 - XAVGC26) > 0) AND ((XAVGC12.6 - XAVGC26.6) > 0) AND ((XAVGC12.7 - XAVGC26.7) > 0) AND ((XAVGC12.8 - XAVGC26.8) > 0) AND ((XAVGC12.9 - XAVGC26.9) > 0) AND ((XAVGC12.10 - XAVGC26.10) > 0) AND ((XAVGC12.11 - XAVGC26.11) > 0) AND ((XAVGC12.12 - XAVGC26.12) > 0) AND ((XAVGC12.13 - XAVGC26.13) > 0) AND ((XAVGC12.14 - XAVGC26.14) > 0) AND ((XAVGC12.15 - XAVGC26.15) > 0) AND ((XAVGC12.16 - XAVGC26.16) > 0) AND ((XAVGC12.17 - XAVGC26.17) > 0) AND ((XAVGC12.18 - XAVGC26.18) > 0) AND ((XAVGC12.19 - XAVGC26.19) > 0) AND ((XAVGC12.20 - XAVGC26.20) > 0) AND ((XAVGC12.21 - XAVGC26.21) > 0) AND ((XAVGC12.22 - XAVGC26.22) > 0) AND ((XAVGC12.23 - XAVGC26.23) > 0) AND ((XAVGC12.24 - XAVGC26.24) > 0) AND ((XAVGC12.25 - XAVGC26.25) > 0) AND ((XAVGC12.26 - XAVGC26.26) > 0) AND ((XAVGC12.27 - XAVGC26.27) > 0) AND ((XAVGC12.28 - XAVGC26.28) > 0) AND ((XAVGC12.29 - XAVGC26.29) > 0) AND ((XAVGC12.30 - XAVGC26.30) > 0) AND ((XAVGC12.31 - XAVGC26.31) > 0) AND ((XAVGC12.32 - XAVGC26.32) > 0) AND ((XAVGC12.33 - XAVGC26.33) > 0) AND ((XAVGC12.34 - XAVGC26.34) > 0) AND ((XAVGC12.35 - XAVGC26.35) > 0) AND ((XAVGC12.36 - XAVGC26.36) > 0) AND ((XAVGC12.37 - XAVGC26.37) > 0) AND ((XAVGC12.38 - XAVGC26.38) > 0) AND ((XAVGC12.39 - XAVGC26.39) > 0) AND ((XAVGC12.40 - XAVGC26.40) > 0) AND ((XAVGC12.41 - XAVGC26.41) > 0) AND ((XAVGC12.42 - XAVGC26.42) > 0) AND ((XAVGC12.43 - XAVGC26.43) > 0) AND ((XAVGC12.44 - XAVGC26.44) > 0) AND ((XAVGC12.45 - XAVGC26.45) > 0)
|
|
Registered User
Joined: 1/28/2005
Posts: 6,049
|
"Here's the formula for 12,26,9 MACD over 0 for 45 days - it can obviously be tweaked for longer or shorter periods - or maybe Diceman can tell us how to do an indicator sort."
--------------------------------------------------------
hohandy's scan actually tests for 46 days.
Using the methods I described above it can be turned into a sort
using:
ABS(XAVGC12>XAVGC26)*46
(set the smoothing average to 46 simple)
After the stocks listed above. It can be seen that next on the list
is RIMM at 45 days and PAYX and COST at 42 days.
--------------------------------------------------------------------------
If the MACD histogram is desired. This can be used.
I choose 42 days to represent 2 months of data:
ABS((XAVGC12-XAVGC26)-(XAVG(XAVGC9,12)-XAVG(XAVGC26,9))>0)*42
(set the smoothing average to 42 simple)
The best in the Nasdaq 100 is WFMI with 36 out of 42 above zero.
The worst is LAMR with only 2 days out of 42 above zero.
Thanks
diceman
|
|
Registered User
Joined: 12/21/2004
Posts: 902
|
Thanks, Dice!
|
|
Registered User
Joined: 9/21/2005
Posts: 566
|
Hohandy can you explain to me what Diceman added in that post directly above yours? Mayabe its late and my brain is tired but I am not clear on how to use his entries aas a supplement to your macd above 45 stuff?
Do you seek to scan for 2 months of data when using this macd > 0?
|
|
Registered User
Joined: 1/28/2005
Posts: 6,049
|
jcfla7
Did you read my first post?
If you did. It will make sense.
Thanks
diceman
|
|
Registered User
Joined: 9/22/2005
Posts: 849
|
Diceman,
It seems to me that this tenique can be applied to any indicator for which you can devise an equation.
Is there a way to apply it to TSV?
|
|
Registered User
Joined: 1/28/2005
Posts: 6,049
|
Yes bcraig.
It can be used for any condition that can be
defined by true/false.
For example TSV18>0 would be:
(over 42 days):
ABS(TSV18>0)*42
(set smoothing average to 42 simple)
GRMN is number 1 on the Nasdaq 100.
With 42 out of 42 days above zero.
-----------------------------------------------------
It can also be applied to price patterns:
ABS(C>C1ANDH>H1ANDL>L1)*42
(smoothing average set to 42 simple)
This counts over the last 42 days how
often the high, low and close were greater
than the last days values.
(all three at once)
GRMN was number 1 again with 20 days
out of 42.
-------------------------------------------------
Thanks
diceman
|
|
Registered User
Joined: 12/21/2004
Posts: 902
|
JC - in case you didn't get it - my indicator returns a simple "True/False" whether or not a stock's MACD has been over 0 for the entire time during a period. Diceman's formula is a chart indicator that you can sort a watchlist by to rank the watchlist by number of days during the time period that all the stocks MACDs were over 0, regardless of whether it is currently over 0, of course. When you do the sort, sort by "Actual value" rather than "Visual value" to see the actual number of days.
I originally learned to use 45 bars as my "medium term" time period length (I still use 45 in my mav calculations). I think Tobydad turned me onto the significance of the various zoom lengths and I've since settled on zoom=6 (3 months or 65 bars) as my "medium term" length. I think I was under the impression that zoom period length changed with the sizing of my chart window, so never had paid attention to it before.
So, as I adapt Diceman's formula for my own use, I will make the smoothing average be 65 rather than 45.
As the Bcraig/Diceman discussion confirmed, this formula method can be applied to most indicators that one can think of that can be used to identify strength and weakness. One that immediately comes to mind, which I think I will enjoy playing with, is number of days in a time period that stocks are up with higher than average volume as a means of identifying strength. And you can be sure that I will be torturing the MACD-H application to see what secrets I can devine there.
I just may have to name my firstborn "Dice".
|
|
Registered User
Joined: 9/22/2005
Posts: 849
|
Diceman
How can I turn the MACD histgram
(ABS((XAVGC12-XAVGC26)-(XAVG(XAVGC9,12)-XAVG(XAVGC26,9))>0)*42)
into a PCF for a crossover from negative to positive?
|
|
Registered User
Joined: 9/22/2005
Posts: 849
|
David
I, too, am a lurker who monitors many threads and seldom let mysef be known.
If I have a comment which I feel will add to the discussion or a question for my own edification, then I will enter into the dialog. Many times I just listen and learn.
|
|
Registered User
Joined: 9/22/2005
Posts: 849
|
Diceman
Where has Tobydad gone? Did he make enough to retire and move to Jamaica or did he just decide that he doesn't love us any more?
|
|
Registered User
Joined: 12/21/2004
Posts: 902
|
BCraig - can you explain what you're looking for in the PCF formula that you're asking for? The one tells how many days in a time period the MACD-H is over 0, regardless of where the MACD-H is currently - it's a counting measure. The other describes a current condition where the MACD-H is crossing over 0 - it's a timing measure. To me, the two just don't seem to mesh. I can conceive using them as two separate and complimentary PCFs, but I'm not sure what you're trying to do combining them into one PCF.
|
|
Registered User
Joined: 9/22/2005
Posts: 849
|
I'm looking for the crossover and just playing with this indcator. I have concluded that this one isn't appropriate for that purpose.
|
|
Registered User
Joined: 12/21/2004
Posts: 902
|
Bcraig - I have a basic MACD-H crossover PCF forumula on my computer at home - if Diceman or anyone else hasn't posted one before I get home tonight (about 7 edt), I'll be glad to put it up for you.
Btw - I am enjoying playing with this new indicator a lot - hope you are too
|
|
Registered User
Joined: 3/21/2006
Posts: 4,308
|
Diceman, I have read this thread a dozen times so as not to make a mistake or miss-interpretation and waste your time, but your indicator ABS(C>AVGC50)*60 Does not seem to account for when price falls bellow the 50 day pma during the sorting process. I revised the formula for 100 days - ABS(C>AVGC50)*100.
When I sort by actual value the ticker NVEC is reading a value of 71 even though its price has been below the 50 day pma for over a month now.
You could make the consession that only when the indicator is sporting a positive slope are the readings valid, but somthing tells me that is not what you had in mind.
|
|
Registered User
Joined: 12/21/2004
Posts: 902
|
Apsll - the indicator is only supposed to count the number of days (bars) in the time period that the condition is met (i.e, price is over the 50pma). If your time period is 100 bars, and a month is about 21-22 bars, then theoretically it could be below the 50pma for the entire last month yet still return a value of 71 for the entire period. Was it primarily over the 50pma during the rest of the 100 day period?
|
|
Registered User
Joined: 3/21/2006
Posts: 4,308
|
Thank you Hohandy, I was reading this indicator the wrong way and now you have opened my eyes.   
|
|
Registered User
Joined: 12/21/2004
Posts: 902
|
Apsll - as was discussed later in the thread, the formula methodolgy lends itself well to use with other indicators. I think for your pennance you should think of a cool application of the formula and share it with us 
|
|
Registered User
Joined: 3/21/2006
Posts: 4,308
|
Done:
I have added a 10 day linear regression line to the indicator and now you can sort by the LR line (visual slope) to find stocks that are still above the 50 day pma.
I tested it on the Electronics, Semiconductor, Specialized sector. Here are some stocks that are putting in some nice bottoms.-
BKHM
NMGC
SLAB
UCTT
|
|
Registered User
Joined: 3/21/2006
Posts: 4,308
|
I will think up a better one then above, but that will do for today I hope. Am I forgiven
|
|
Registered User
Joined: 3/21/2006
Posts: 4,308
|
Hohandy, have you ever heard of the three black crows candlestick formation? It kind of goes along with your MACD-H re-test thoery...
|
|
Registered User
Joined: 12/21/2004
Posts: 902
|
Sorry - I was at a drs appt. yes I've heard of and look for the 3 black crows. I do like the 3 white soldiers better, however...
How do they fit with the retest? I think of them more as a reversal than something at the end of a trend that's suitable for retest.
And yes, you are provisionally forgiven... depending on what cooler application you come up with tomorrow (With luck I'll drag this out all week and then we'll all profit )
|
|
Registered User
Joined: 3/21/2006
Posts: 4,308
|
I have been reading (my mag) and there is an article on pg 48 Counting Crows. They quote Oliver Velez From his book Strategies for profiting on every trade: Simple lessons for mastering the Market.
They say that a good entry is to find a strong trending stock that has recently made a new high, then to look for the three crows pattern (a pull-back) with three consecutive lower highs (bars). Then they say to buy above the high of the third bar. It could take a few more days or just lead right to a down trend but if the buy signal is trigered then it can be a very profitable trade. They give it a high probability rating.
I thought that maybee you could put a MACD-H spin on this..
|
|
Registered User
Joined: 12/21/2004
Posts: 902
|
is your mag TASC? Mine came this weekend but I haven't looked at it yet.
It sounds suspisciously like the HNC/Diceman/MACDH100 setup though - providing that the 3 crows are just a pull-back and not an out-and-out reversal. To tell the truth though, when I was doing a lot of investigating of the HNC/Diceman/MACDH100 setup earlier this spring and summer (I've put it on hiatus until we get a trending market again) I don't recall seeing a whole lot of 3 crow pull-backs though (although I wasn't specifically looking for them). I guess it's a situation that not every HNC/Diceman/MACDH100 will be a 3 crows, but many 3 crow pullbacks after a recent high may be HNC/Diceman/MACDH100s.
I wonder if we could corrolate the 3 crows with MACD-H level - a high MACD-H in the high before the crows indicates a pull-back opportunity, a low MACD-H in the high before the crows indicates a real reversal. Hmmmm....
|
|
Registered User
Joined: 3/21/2006
Posts: 4,308
|
I have constructed a percent true indicator to find the buy signal. The following chart might help you visually.
|
|
Registered User
Joined: 3/21/2006
Posts: 4,308
|
Yes TASC is the mag. After you read the article let me know what you think.
|
|
Registered User
Joined: 12/21/2004
Posts: 902
|
The chart makes me wonder if rather than corrolating the 3 crows with MACD-H, if we should corrolate with BOP or another accumulation indicator. Interesting.
|
|
Registered User
Joined: 12/21/2004
Posts: 902
|
Bcraig - here's a PCF formula for MACD-H crossing 0 (I'm sure Diceman has a much simpler one)
((.0001 * ( 638.2398 * C0 + 365.1241 * C1 + 172.5132 * C2 + 40.0649 * C3 - 47.8228 * C4 - 103.0645 * C5 - 134.7136 * C6 - 149.6092 * C7 - 152.8821 * C8 - 148.3502 * C9 - 138.8258 * C10 - 126.3551 * C11 - 112.4034 * C12 - 97.9987 * C13 - 83.8417 * C14 - 70.3912 * C15 - 57.928 * C16 - 46.605 * C17 - 36.484 * C18 - 27.5638 * C19 - 19.801 * C20 - 13.1255 * C21 - 7.4515 * C22 - 2.6856 * C23 + 1.2673 * C24 + 4.5006 * C25 + 7.103 * C26 + 9.1575 * C27 + 10.7399 * C28 + 11.919 * C29 + 12.7564 * C30 + 13.3066 * C31 + 13.6176 * C32 + 13.7314 * C33 + 13.6843 * C34 + 13.5076 * C35 + 13.2282 * C36 + 12.8689 * C37 + 12.449 * C38 + 11.9848 * C39 + 11.4898 * C40 + 10.9753 * C41 + 10.4506 * C42 + 9.923 * C43 + 9.3988 * C44 + 8.8827 * C45 + 8.3786 * C46 + 7.8892 * C47 + 7.4167 * C48 + 6.9628 * C49 + 6.5284 * C50 + 6.114 * C51 + 5.7201 * C52 + 5.3466 * C53 + 4.9933 * C54 + 4.6597 * C55 + 4.3454 * C56 + 4.0498 * C57 + 3.7721 * C58 + 3.5117 * C59 + 3.2676 * C60 + 3.0393 * C61 + 2.8257 * C62 + 2.6263 * C63 + 2.4401 * C64 + 2.2664 * C65 + 2.1046 * C66 + 1.9538 * C67 + 1.8134 * C68 + 1.6827 * C69 + 1.5612 * C70 + 1.4482 * C71 + 1.3432 * C72 + 1.2456 * C73 + 1.1549 * C74 + 1.0707 * C75 + .9926 * C76 + .92 * C77 + .8527 * C78 + .7902 * C79 + .7323 * C80 + .6786 * C81 + .6287 * C82 + .5825 * C83 + .5397 * C84 + .5 * C85 + .4631 * C86 + .429 * C87 + .3974 * C88 + .3681 * C89 + .3409 * C90 + .3158 * C91 + .2925 * C92 + .2709 * C93 + .2509 * C94 + .2323 * C95 + .2152 * C96 + .1993 * C97 + .1845 * C98 + .1709 * C99 + .1583 * C100 + .1465 * C101 + .1357 * C102 + .1257 * C103 + .1164 * C104 + .1078 * C105 + .0998 * C106 + .0924 * C107 + .0856 * C108 + .0792 * C109 + .0734 * C110 + .0679 * C111 + .0629 * C112 + .0582 * C113 + .0539 * C114 + .0499 * C115 + .0462 * C116 + .0428 * C117 + .0396 * C118 + .0367 * C119 + .034 * C120 + .0315 * C121 + .0291 * C122 + .027 * C123 + .025 * C124) ) > 0) AND ((.0001 * ( 638.2398 * C1 + 365.1241 * C2 + 172.5132 * C3 + 40.0649 * C4 - 47.8228 * C5 - 103.0645 * C6 - 134.7136 * C7 - 149.6092 * C8 - 152.8821 * C9 - 148.3502 * C10 - 138.8258 * C11 - 126.3551 * C12 - 112.4034 * C13 - 97.9987 * C14 - 83.8417 * C15 - 70.3912 * C16 - 57.928 * C17 - 46.605 * C18 - 36.484 * C19 - 27.5638 * C20 - 19.801 * C21 - 13.1255 * C22 - 7.4515 * C23 - 2.6856 * C24 + 1.2673 * C25 + 4.5006 * C26 + 7.103 * C27 + 9.1575 * C28 + 10.7399 * C29 + 11.919 * C30 + 12.7564 * C31 + 13.3066 * C32 + 13.6176 * C33 + 13.7314 * C34 + 13.6843 * C35 + 13.5076 * C36 + 13.2282 * C37 + 12.8689 * C38 + 12.449 * C39 + 11.9848 * C40 + 11.4898 * C41 + 10.9753 * C42 + 10.4506 * C43 + 9.923 * C44 + 9.3988 * C45 + 8.8827 * C46 + 8.3786 * C47 + 7.8892 * C48 + 7.4167 * C49 + 6.9628 * C50 + 6.5284 * C51 + 6.114 * C52 + 5.7201 * C53 + 5.3466 * C54 + 4.9933 * C55 + 4.6597 * C56 + 4.3454 * C57 + 4.0498 * C58 + 3.7721 * C59 + 3.5117 * C60 + 3.2676 * C61 + 3.0393 * C62 + 2.8257 * C63 + 2.6263 * C64 + 2.4401 * C65 + 2.2664 * C66 + 2.1046 * C67 + 1.9538 * C68 + 1.8134 * C69 + 1.6827 * C70 + 1.5612 * C71 + 1.4482 * C72 + 1.3432 * C73 + 1.2456 * C74 + 1.1549 * C75 + 1.0707 * C76 + .9926 * C77 + .92 * C78 + .8527 * C79 + .7902 * C80 + .7323 * C81 + .6786 * C82 + .6287 * C83 + .5825 * C84 + .5397 * C85 + .5 * C86 + .4631 * C87 + .429 * C88 + .3974 * C89 + .3681 * C90 + .3409 * C91 + .3158 * C92 + .2925 * C93 + .2709 * C94 + .2509 * C95 + .2323 * C96 + .2152 * C97 + .1993 * C98 + .1845 * C99 + .1709 * C100 + .1583 * C101 + .1465 * C102 + .1357 * C103 + .1257 * C104 + .1164 * C105 + .1078 * C106 + .0998 * C107 + .0924 * C108 + .0856 * C109 + .0792 * C110 + .0734 * C111 + .0679 * C112 + .0629 * C113 + .0582 * C114 + .0539 * C115 + .0499 * C116 + .0462 * C117 + .0428 * C118 + .0396 * C119 + .0367 * C120 + .034 * C121 + .0315 * C122 + .0291 * C123 + .027 * C124 + .025 * C125 )) < 0)
|
|
Registered User
Joined: 1/28/2005
Posts: 6,049
|
Try
(up through zero):
(XAVGC12-XAVGC26)>(XAVG(XAVGC9,12)-XAVG(XAVGC26,9)) AND (XAVGC12.1-XAVGC26.1)<(XAVG(XAVGC9.1,12)-XAVG(XAVGC26.1,9))
(down through zero):
(XAVGC12-XAVGC26)<(XAVG(XAVGC9,12)-XAVG(XAVGC26,9)) AND (XAVGC12.1-XAVGC26.1)>(XAVG(XAVGC9.1,12)-XAVG(XAVGC26.1,9))
Thanks
diceman
|
|
Registered User
Joined: 9/22/2005
Posts: 849
|
Thanks Diceman and hohandy.
I used to be fairly proficient writing in PL1 (IBM quit supporting it in the 60's), Fortran, and a couple of the basics but whatever you guys are using is beyond me and I don't have the time nor the inclination to learn it.
I’ll depend on you to set me straight.
|
|
Registered User
Joined: 9/22/2005
Posts: 849
|
Thanks Diceman and hohandy.
I used to be fairly proficient writing in PL1 (IBM quit supporting it in the 60's), Fortran, and a couple of the basics but whatever you guys are using is beyond me and I have neither the time nor the inclination to learn it.
I’ll depend on you to set me straight.
|
|
Registered User
Joined: 3/21/2006
Posts: 4,308
|
Hohandy, here is my next twist on the ABS(C>AVGC50)*50
Indicator. Keep in mind that this is basic stuff and will not cause a shift in the earths current rotation axis.
ABS(AVGC50>AVGC200)*15
I have read many times that the 50 day pma crossing the 200 day pma is important for chart pattern development. So I tweaked a few numbers from the original formula and came up with an indicator that will tell us (in the same way as dicemans indicator) how many days the 50 has beeen over the 200 (within a given time period, in this case 15 days). This indicator should go in the middle window and to the indicator add a 15 day Linear regression Line.
In the bottom window go with Volume bars and the original indicator - ABS(C>AVGC50)*50 (to the indicator add a 10 day linear regression line.
In the top window of corse we want a 50 & a 200 day pma on price (add BOP onto price hidden).
First thing is to sort a list (of your choosing) by the new indicator (actual value) discarding all tickers with a value of 0. Next sort by the 15 day Linear regression line (visual slope) and discard all values that are less than o (this time you want to keep all the zero's). Now you have a list of stocks where the 50 day pma is above the 200 day pma.
Now we also want price to be above the 50 day pma so sort the list by the 10 day linear regression line from the indicator in the bottom window (visual slope) and discard all stocks that their value is lower than 0 (keep all the zero's).
You should now be left with a managable list you can veiw one at a time and pick out the best. Use the volume indicator to find a break-out pattern. Below are a couple of charts without the volume bars, but you will get the idea.
Good luck.
|
|
Registered User
Joined: 3/21/2006
Posts: 4,308
|
|
|
Registered User
Joined: 3/21/2006
Posts: 4,308
|
The second chart is to show how the stock whent up in price from the initial find.
|
|
Registered User
Joined: 12/21/2004
Posts: 902
|
Hey - good stuff Apsll
|
|
Registered User
Joined: 12/21/2004
Posts: 902
|
Apsll - interesting 3 crows article - although the author seemed to go through a lot of testing not to say much. I'd say this would an idea to test against MACD-H.
Last night I looked at the DCEL chart that you posted yesterday, and sure enough, the high the day before the 3 crows was a multi-month MACD-H high (didn't reach 100, because of a lower MACD-H low earlier in the time frame - but still would fulfill the Elder Dictum). I would like to test some more against the MACD-H - and I'll look at the ones that he listed in the Excel spreadsheet portion of his article. How did you find DCEL?
Also - look at the mid-May drop in the OSUR chart you have posted above - is that a 3 crows there - or is the body on the first bar not long enough?
|
|
Registered User
Joined: 9/22/2005
Posts: 849
|
Apsll
I set up my template as you recommended above and used to evaluate my watchlist of one to ten dollar stocks. (851)
When I sorted by the ABS(AVGC50>AVGC200)*15 indicator in the middle window it gave 288 stocks with the indicator greater than zero. A sort on the LR15 gave 288 stocks with the indicator greater than or equal to zero.
Then a sort of the LR10 of the indicator in the bottom window ABS(C>AVGC50)*50 resulted in 240 with a slope of zero and 48 with a slope greater than zero.
It seems to me that we want to discard those with zero slope and keep those with slopes greater than zero.
Or did I miss something?
|
|
Registered User
Joined: 3/21/2006
Posts: 4,308
|
Hohandy, I found DCEL by using my indicator designed for the purpose of locating the condition three black crows followed by the buy signal the buy signal is that if price goes higher than the high of the third days high in the formation.
I am not sure if the is the three black crows formation for OSUR in May. The stock was not in an up-trend, but I do know that it was not followed up with a buy signal. (those are the key points).
Bcraig,I tried to duplicate your test above and came up with a similar read in your first step. I came up with 305 stocks after the initial sort, I then sorted by the 15 day LR line (visual slope) in the middle window and my numbers were 255 with a value of 0 or higher. I discarded the numbers lower than zero. In the final step I sorted the 10 day Lr line in the bottom window and was left with 63 charts that were higher than or equal to zero.
I am not sure were you whent wrong, but you want to keep the zero slopes for the Linear regression sorts because what they are saying is important. It signifies that the indicator is maxed out to the limits that you have set for the indicator and will remain so until the condition changes (the 50 day pma drops below the 200 day pma). In that case the visual slope would now change to a negative slope.
Let me know if you need some more help with this.
My design for this indicator only required a minor tweak to Dicemans design to give me a broader veiw of the 50 day pma's action.
Good luck.
Apsll.
|
|
Registered User
Joined: 3/21/2006
Posts: 4,308
|
Bcraig, I have to correct my methods. There is one final step I have over-looked in this process. In the middle window add the percent true indicator - C<avgc50
Sort the list of 63 charts by the percent true indicator (actual value). Hit the home key and discard all charts with the sort value of 100.
You were partially correct that in the last sort (The 10 day LR line) some of the visual slopes with the value of zero were found on the down side of the indicator in the bottom window and we do need to eliminate those particular charts. The percent true indicator that I have just added will identify them and now the revised list consists of 33 charts that we can now view one at a time.
Bcraig thank you for helping me to put the final piece of the puzzle in place..
Apsll..
|
|
|
Guest-1 |