| Welcome Guest, please sign in to participate in a discussion. | Search | Active Topics | |
Registered User
Joined: 10/19/2008
Posts: 33
|
The Kicking formula is given in many locations as: (O3-C3)/((.001+(H3-L3))>0.6)AND(O2-C2)/((.001+(H2-L2))>0.6)AND(O1-C1)/((.001+(H1-L1))>0.6)AND(C3<O3)AND(C2<O2)AND(C1<O1)AND(C>O)AND(O2<C3)AND(O1<C2)AND(O>O1)AND(C-O)/((.001+(H-L))>0.6)
I'm having trouble getting my head around the first term: (O3-C3)/((.001+(H3-L3))>0.6)
If it was written as: ((O3-C3)/(.001+(H3-L3))>0.6) I would understand it as a down candle who's body is greater than 60% of the total range of the candle and the first 3 terms of the whole formula would imply a low stochastic and approach candles with short wicks which are requirements for a kicker setup.
Maybe I just don't understand the order of operations in computerese. Could you help me understand the operation of the first term please.
Thankyou
|
|
Registered User
Joined: 1/28/2005
Posts: 6,049
|
(O3-C3)/((.001+(H3-L3))>0.6)
This asks if the (Open - Close) divided by the (High-Low) is greater than .6
This looks at the bars 3 bars back from the current bar.
The .001 is to prevent a divide by zero error, in case H-L=0.
The fact that .6 is positive means a higher close than open is required for a true condition.
(taking a negative value and dividing it by the H-L range will be negative)
Thanks
diceman
|
|
Registered User
Joined: 10/19/2008
Posts: 33
|
Thanks for the reply diceman but I'm afraid I'm still not getting it. Let's look at the 3 sets of parentheses in the denominator:
inner set (H3-L3)
second set (.001 + (H3-L3))
third set ((.001 + (H3-L3)) > 0.6)
The above third set makes up the whole denominator (including the >0.6) {thats the part I don't get}
which is divided into the numerator (O3-C3)
To make it do what you say it should do you have to put it into the form in my original post (don't you?).
ie. (O3-C3)/(.001 + (H3-L3)) >0.6 if you leave out the outer parentheses or ((O3-C3)/(.001+(H3-L3)) > 0.6) if you include them.
I'm further confused by your statement "The fact that .6 is positive means a higher close than open is required for a true condition." We are looking for a bullish reversal so the preceding 3 candles are all supposed to be down candles (close less than open).
I've now programmed in both ways in TC2000 and in a random list of 113 stocks, I found 12 instances of the original formula finding a "bullish kicker" and none for the revised version. According to Julia on the webinar on reversing candles a Bullish Kicker is supposed to be a very rare event. The hits are more than 10% so I wouldn't call it rare.
I can see that the instances where the formula is identifying a bullish kicker look reasonable but I suspect that TC2000 is ignoring the first 3 candle requirements. If you use the "test formula" feature in the condition or indicator panel with the original formula you get an error message about not enough data (can't be true because the whole calculation only involves 4 bars). I suspect the >0.6 being in the denominator is causing the error message.
I know I can be a bit thick in the head on some of these things but I'm sorry, I still can't understand the equation as it is written in the reference posts.
Thanks
|
|
Registered User
Joined: 1/28/2005
Posts: 6,049
|
I’m sorry pop33
Looking at this quickly I thought it was this:
((O3-C3) /(.001+(H3-L3))>0.6)
(because this actually “works” in that it gives a result)
Which would be reduced this to:
(O3-C3) /(H3-L3)>0.6
The piece you’re looking doesn’t work (syntax error), which means later
parts of the equation are important.
(that’s probably why its confusing)
What I meant to say was this value (O3-C3)
should be positive.
To be positive the open would need to be higher than the close.
(of course that was based on this:
((O3-C3) /(.001+(H3-L3))>0.6)
being the equation which its not)
Sorry for any confusion.
Thanks
diceman
|
|
Registered User
Joined: 10/19/2008
Posts: 33
|
Thanks Diceman
I created a stripped version of the equation leaving out the first 3 and the last items that contain the syntax error. The resulting equation caught all of the hits for the original equation plus some more so the error parts of the equation are having some effect.
Revising the equation to the form that I was suggesting is a non-starter as I scanned 400 charts from Feb 2010 to present and didn't get a single hit.
Assuming you are going to revise the formula, please note that the error also occurs in the Bullish Harami which I only noticed in passing. I have not examined the other candlestick formulas.
Thanks for your help.
|
|
Worden Trainer
Joined: 10/7/2004
Posts: 65,138
|
My best guess as to the what the formula should be is:
(O3 - C3) / (.001 + H3 - L3) > .6 AND (O2 - C2) / (.001 + H2 - L2) > .6 AND (O1 - C1) / (.001 + H1 - L1) > .6 AND O2 < C3 AND O1 < C2 AND O > O1 AND (C - O) / (.001 + H - L) > .6
-Bruce
Personal Criteria Formulas
TC2000 Support Articles
|
|
Registered User
Joined: 10/19/2008
Posts: 33
|
Thanks Bruce
Your equation is the same as my modified one which had no hits in 400 charts Feb 2010 to present. Mine has a few more parentheses but it is the same, ie.
((O3-C3)/(.001+(H3-L3))>0.6)AND((O2-C2)/(.001+(H2-L2))>0.6)AND((O1-C1)/(.001+(H1-L1))>0.6)AND(C3<O3)AND(C2<O2)AND(C1<O1)AND(C>O)AND(O2<C3)AND(O1<C2)AND(O>O1)AND(C-O)/((.001+(H-L))>0.6)
|
|
Registered User
Joined: 10/19/2008
Posts: 33
|
Bruce
After the last post I see that I have not fixed the last term. Formula should be:
((O3-C3)/(.001+(H3-L3))>0.6)AND((O2-C2)/(.001+(H2-L2))>0.6)AND((O1-C1)/(.001+(H1-L1))>0.6)AND(C3<O3)AND(C2<O2)AND(C1<O1)AND(C>O)AND(O2<C3)AND(O1<C2)AND(O>O1)AND((C-O)/(.001+(H-L))>0.6)
I'll make the change in TC2000 and see if that makes any difference. If it does I'll let you know.
|
|
Registered User
Joined: 10/19/2008
Posts: 33
|
Bruce
I checked about 100 stocks, still no hits.
|
|
Worden Trainer
Joined: 10/7/2004
Posts: 65,138
|
I'm not a candlestick guy (at all), but I can break down the formula. The requirements are relatively restrictive but do not appear to be impossible. I've re-arranged it slightly to show what it represents (it will match the order of the less restrictive version at the end of the post):
(O3 - C3) / (.001 + H3 - L3) > .6 AND C3 > O2 AND (O2 - C2) / (.001 + H2 - L2) > .6 AND C2 > O1 AND (O1 - C1) / (.001 + H1 - L1) > .6 AND (C - O) / (.001 + H - L) > .6 AND O > O1
You have three candles where the open is greater than the close with bodies which are at least 60% of the entire range of the candle.
(O3 - C3) / (.001 + H3 - L3) > .6
(O2 - C2) / (.001 + H2 - L2) > .6
(O1 - C1) / (.001 + H1 - L1) > .6
The body of the first candle is entirely above the body of the second candle and the body of the second candle is entirely above the body of the third candle.
C3 > O2
C2 > O1
The fourth and final candle in the pattern also has a body which is at least 60% of the entire candle, but the close is above the open.
(C - O) / (.001 + H - L) > .6
The entire body of the fourth candle is above the entire body of the third candle.
O > O1
If you eliminate the requirement that the bodies be at least 60% of the candle, the formula simplifes to:
O3 > C3 AND C3 > O2 AND O2 > C2 AND C2 > O1 AND O1 > C1 AND C > O AND O > O1
-Bruce
Personal Criteria Formulas
TC2000 Support Articles
|
|
Registered User
Joined: 1/28/2005
Posts: 6,049
|
pop33
Just my 2 cents.
I did a visual check with the first formula:
(O3-C3)/((.001+(H3-L3))>0.6)AND(O2-C2)/((.001+(H2-L2))>0.6)AND(O1-C1)/((.001+(H1-L1))>0.6)AND(C3<O3)AND(C2<O2)AND(C1<O1)AND(C>O)AND(O2<C3)AND(O1<C2)AND(O>O1)AND(C-O)/((.001+(H-L))>0.6)
Most stocks only get 2 hits over a period of 2 years.
Seems like a very restrictive scan.
Thanks
diceman
|
|
Registered User
Joined: 10/19/2008
Posts: 33
|
Thanks for all your help.
|
|
Worden Trainer
Joined: 10/7/2004
Posts: 65,138
|
You're welcome. Our pleasure.
-Bruce
Personal Criteria Formulas
TC2000 Support Articles
|
|
|
Guest-1 |