Registered User
Joined: 10/7/2004
Posts: 2,181
|
I have found this to be handy in determining the likelihood of an anticipated successful reversal.
Set stochastics at 2,2,3; leave the "Visible" box UNchecked; then add a 20-day linear regression. It's a simple little confirmation signal that doesn't clutter up one of the windows (too much "noise" --for you electronics nuts out there-- with stochs set at 2,2,3).
When the bar is pointing upward and above "50", the odds of very short term (ie one day) upward movement is in our favor.
A price that has hit or violated the lower 20-day bollinger band and resting on a trend line, accompanied by the signal above is frequently a good one day move. If the markets are favorable I will buy at the open and set a very tight stop or a trailing stop.
CAUTION: Don't try this at home! No, seriously, as Diceman says, "...don't fish in these waters..." if you're not sure what you're doing. There are additional considerations. Paper trade this for a while first.
You may want to combine this with some of the info on the recent HNC threads.
|
|
Registered User
Joined: 1/28/2005
Posts: 6,049
|
Another interesting idea. Instead of a straight linear
regression line. You can try adding a 20 period
moving linear regression line. (at each point in time
it equals the last point of a 20 period linear regression line)
The advantage is you can see the "tip" move thru time.
If you look at a stock symbol like : CLI
(I just happen to have the ishares real-estate index open)
On 11/28/06 the indicator turns down before the top and
on 12/20/06 the indicator turns up before the price run up.
Indicating it may have some "leading" qualities.
How often this happens and how reliable it is I don't know
but it seems worth looking into. Also you don't "lose"
information because this will cross 50 the same as the
straight line.
Thanks
diceman
|
|
Registered User
Joined: 10/7/2004
Posts: 2,181
|
Diceman;
Thanks for the recommendation, it adds valuable insights.
In my mind, it's a purely subjective issue so I'm not arguing the point; but for those that like the std linear regression better (I do), you can use the [ and ] keys to move the chart back and forward a few days to see the progression and direction of the LR.
I probably do it that way cause I'm used to it.
|
|
Registered User
Joined: 10/7/2004
Posts: 2,181
|
Diceman's post brings up a good point:
I have been saying "20-day this" or "30-day that"; I shouldn't.
Diceman is correct in saying "20-period". Most of us know this; but for the newer members a 20-period setting on anything will be 20 days or 20 weeks or 20 months depending upon your time setting.
File it in the FYI.
|
|
Registered User
Joined: 4/17/2006
Posts: 271
|
Interesting indicators...both the straight LR and moving LR combined. Will give it a try. Seems to hold promise. Thanks tobydad and diceman.
|
|
Registered User
Joined: 3/21/2006
Posts: 4,308
|
Good morning to all.. I have been testing both versions of this leading indicator and I just want to make sure that I am doing this correct.. I will just use NTBK as an example, because it is the first ticker that I attempted to backtest this indicator, and it failed the test. (Tobydad please let me know if I am doing this wrong)..
Middle window:
Stochastics 2,2,3 (not visable) Linear Regresion (period 20)
When the line is pointing up & above 50
Bottom window:
Stochastics 2,2,3 (not visable) Moving linear regression (period 20)
When the tip is pointing up
NTBK: on 1/09/07 this condition was true for both indicators, and the stock droped in price.
And now today just the moving linear regression line is pointing up. I am not trying to start controversey, but of course I would be interested in any indicator that will lead price...
If I am doing this wrong then please let me know..
Apsll...
|
|
Registered User
Joined: 1/28/2005
Posts: 6,049
|
Check out CLI and see if it
confirms my statements.
Remember there is no
"certain" leading indicator.
It is just an idea.
Thanks
diceman
|
|
Registered User
Joined: 9/22/2005
Posts: 849
|
It doesn't work every time but it seems to be fairly
reliable when it is bouncing off the lower Bollinger
Band (20,20).
Try it on
CTYR
DRRX
USEY
|
|
Registered User
Joined: 10/7/2004
Posts: 2,181
|
CTIC today is a perfect example of the setup I look for when using this indicator.
And, everyone please note Diceman's comment, "...there is no 'certain' leading indicator...".
In fact, please recall that I referred to this as a confirmation. I never have tried to use it on its own.
Also, bcraig had it right when he said "...when it is bouncing off the lower bollinger band (20,20)." Exactly right!
A down day on a range bound stock touching down on the lower bollinger band (hopefully the first half of a bounce) and the LR pointing up to or even above 50 is the setup. When I see that, I like to check to see if there's a supporting trendline as well.
Then, (and this is about the only time that I don't use conditional buy stops above the high of the previous day) I buy at the open and I only look for 1 - 1.5% and then get out.
This is not nearly as reliable as other systems I use but it can be a good one with some practice.
|
|
Registered User
Joined: 1/28/2005
Posts: 6,049
|
tobydad
Notice on AVNR. the "moving" version topped out on
approx. 10/12/06. (well before price)
Bottomed on approx 11/03/06 and has shown a positive
divergence to price ever since. (could bode well for your
position)
Thanks
diceman
|
|
Registered User
Joined: 1/28/2005
Posts: 6,049
|
apsll
"NTBK: on 1/09/07 this condition was true for both indicators, and the stock dropped in price."
I would consider the December and early January tops of the
"moving" indicator to be in sync with the stock.
(although the Dec 22 06 top could be call leading. It turns
down before price and the Jan top is lower)
It is interesting to note however that there is a
divergence during the AUG 06 run-up.The indicator tops
out in mid-Aug and heads downward into Dec.
If we "trust" the indicator. It would have suggested a good short
in approx OCT 26 with a break thru the 5.63 support/resistance
level.
(notice also the indicator in positive mode long before the
Aug 06 run-up)
Thanks
diceman
|
|
Registered User
Joined: 3/21/2006
Posts: 4,308
|
I agree with Tobydad and yourself that this is a nifty little indicator, and can be used as an early warning method to let one know that a given stock might need to be looked at in more detail and given serious consideration. But (and I am pretty sure that you and TB agree)In no way would one use this as a buy signal..
I have a template set up with these two indicators along with Bolinger Bands 20,20 and will be conducting extensive back testing.
Keep these good ideas coming, we need all the weapons we can get in our arsenals...
|
|
Registered User
Joined: 1/28/2005
Posts: 6,049
|
I would probably consider this indicator an alert.
Once a buy or sell is given. look from conformation
from traditional indicators and price action.
Thanks
diceman
|
|
Registered User
Joined: 3/21/2006
Posts: 4,308
|
Agreed... 
|
|
Registered User
Joined: 9/22/2005
Posts: 849
|
You should note that CTIK broke the boyyom line of a decending right triangle yrsterday, Bearish
|
|
Registered User
Joined: 3/21/2006
Posts: 4,308
|
I am not trying to beat a dead horse here Diceman but bear with me here... (I followed your exact steps from above and you were correct in the actions of the indicators during those time frames, But when ever I try it on a particular stock, I seem to get negitive results..)
I bought CTCH yesterday, close to the same price that it is now..
With these two indicators (Tobydads version, and yours) lets start on 1/04/07 and scroll foward in time, one day at a time.. Now on the 4rth we can agree that both indicators are very positive... On the 5th & 8th however your indicator starts to turn negative while TBs indicator stays possitive.. By the 18th now they are both negative
My question I guess is why so negative durring this 3 week consolodation when at the end of the consolodation ther is a 10% gain in price pending? Granted on the 24th (the first day of the price increase) your indicator started to turn up, but TBS was still negative. and now your indicator shows negative action again (should I sell?)
I see the current action in price as the begining of another consolidation..
I want this indicator to be reliable, so I am not trying to be dificult. It just seems that when I use these indicators they do not appear to work for me..(maybee they do not like me)
Just my observations.. Apsll...
|
|
Registered User
Joined: 1/28/2005
Posts: 6,049
|
apsll
Realize that you cant compare the moving LR to the straight
LR directly. Basically when the moving LR is > 50 than
the straight is up. When the moving LR is < 50 then
the straight is down.
(the moving LR is the "tip" of the straight each bar)
So you will never see a bottom up-turn or top down-turn
confirmed by the other indicator. This was the point
of making it a moving version. You have all the info of the
straight and as an added bonus you get to see where
turning points occur.
By plotting just the moving you are not losing any information.
As far as CTCH goes. It appears the top in NOV leads the
early Dec selling by a slight margin. The bottom on 12/13
leads the buying by a slight margin.
The "bottom" on 1/22/07 was 58.08 (still positive on the
straight line)
If we forget about the "leading" qualities the straight line
buy (above 50 was 12/21/06) and has remained in the buy mode.
Certainly acceptable.
Realize that I only view this as the moving version.
Since where it is is the tip of the straight line. There
is no reason to view both. (at least from my point of
view)
thanks
diceman
|
|
Registered User
Joined: 3/21/2006
Posts: 4,308
|
Now I get it.. I was reading it wrong the whole time. I was not considering that the moving LR needed to be >50..
I was just mesuring its action by possitive or negative direction..
I will leave the moving LR in the bottom window and in the middle window I have replaced the straight LR with a different take on the situation wich appears at first glance to be responsive..
I still have stochastics 2,2,3 not visable, but with a 10 day exponential moving average.. And on the moving average I have a 16,16 Bollinger Band.. I will back test this whole situation and see what I come up with..
Thank you for clearing this all up for me Diceman..
|
|
Registered User
Joined: 9/22/2005
Posts: 849
|
Apsll
You might want to change your Bollinger Band to 16,20. The second number, 20, is the standrd deviation. 20 is two standard deviations. If you keep it at 16 then about 38 percent of the prices will lie outside of the bands(19 percent below and 19 percent below). If you change it to 20 then only 5 percent of the prices will lie outside the bands.
|
|
Registered User
Joined: 9/22/2005
Posts: 849
|
Apsll
You might want to change your Bollinger Band to 16,20. The second number, 20, is the standrd deviation. 20 is two standard deviations. If you keep it at 16 then about 38 percent of the prices will lie outside of the bands(19 percent below and 19 percent below). If you change it to 20 then only 5 percent of the prices will lie outside the bands.
|
|
Registered User
Joined: 10/7/2004
Posts: 2,181
|
A little added thought (and, admittedly, I have not had time to read the rest of the posts so forgive me if this repeats or no longer relates)
I forgot to mention that one of the ways I have used this has been to buy at the close of the day. CTIC, again, would have been a perfect example. Monday at the close, it hit the lower 20,20 bollinger band landing on a supporting trendline with the LR20 pointing nicely above 50. Next morning opened up a couple of percent.
It's a little risky but, in many instances, it has given me a quick 1-2% gain.
My strategy is this, if I keep 3-5 of my longer term plays going (ie 1 week - 2 months is the average); coupled with a few of these "quickies"; then it gives me a nice balance.
|
|
Registered User
Joined: 3/21/2006
Posts: 4,308
|
I am sorry to be hogging the forum, but there are just so many good opportunities right now.
If you have the time Diceman, let me know what you think about GTCB (Technicals only). It is a trianle pattern that I have been watching...
Top window Bollinger Band 20,20 with triangle trend lines in place
Middle window Stchastics 2,2,3 hidden / 10 day xma with Bollinger Band 10,20 on the 10 day xma..
Bottom window Moving LR 20
Now consider buying into weakness in an ascending triangle. Notice that price is near bottom with the BB 20,20 and the triangle trendline, also notice that in the middle window the 10 day xma is also at the bottom of the BB 10,20
In the bottom window look what happens to the moving LR 20 every time that price touches the bottom of the triangle trend line (It turns up)
I think that a good buy will be when it touches that bottom trend line again around $1.09 And the good thing about this approach is that triangles are very low risk entries. You put your stop right below the triangle.. In this case there is some minor support right at $1.07 so if it breaks the triangle to the downside, or that support, then you are out of the stock...
That is how I am going to play this one, I would just like your opinion...
Thanks Apsll...
|
|
Registered User
Joined: 1/28/2005
Posts: 6,049
|
Sounds like a plan to me. It hasn't show any sustained strength
for a while so you may have to be quick.
(but part of that will be taken care of by buying a
cycle low)
Thanks
diceman
|
|
|
Guest-1 |