a trading strategy based on Elliott Wave Theory - page 85

New comment
 
Well, the calculation of parabolas has been done, but here is a new surprise. Of course, I poked errors into the algorithm at first and fished them out at the third visit. But when I figured it out, I start getting a stable coefficient A for Y=A*X^2+B*X+C equal to zero. Checked the algorithm two more times and finally got sure it has no errors, but I still got A=0 (at least B=0) in all charts. Finally I have found the right moment on the EuroJPY (even PoundJPY was not available).



The only thing I should do now is to implement drawing of parabolas (for illustration purposes) although I still consider parabolas only for placing limit orders.

ZZZ The calculations are still forced
2006.07.13 22:44:33 VG_Ch9 EURJPY,H1: Channel 2 parabola A=0 B=0.0024 C=146.0852<br / translate="no"> 2006.07.13 22:44:33 VG_Ch9 EURJPY,H1: Channel 1 parabola A=0.0002 B=-0.0302 C=146.873
2006.07.13 22:44:33 VG_Ch9 EURJPY,H1: Channel 0 parabola A=-0.0013 B=0.0475 C=146.18
2006.07.13 22:44:33 VG_Ch9 EURJPY,H1: Time Optimized Algorithm 47 ms
2006.07.13 22:44:33 VG_Ch9 EURJPY,H1: initialized

Not only parabola calculation here, but a lot of other unnecessary mathematics :)
 
As for parabolas, yes a lot of interesting things I will try tomorrow to redo the coefficients finding in my own way, because now it is very cumbersome and a lot of errors.

And I think I've figured out the forced calculation algorithm too ;), well, it's all for tomorrow.
 
I tried to attach script ANG310 to draw parabolas, I struggled for about ten minutes, some kind of glitch or something in MT4 - it does not allow to drag the linear regression, and the whole chart keeps sliding.

 
Well, there's a parabola...



I have come to the conclusion that the parameter A, which is zero in the graph, is not actually zero, there are just significant digits after the 4 decimal point. Otherwise the parabola would not draw a parabola, it would draw a straight line.
It remains to check the derivative (for academic interest) and calculate the sum of gradients (deviations from the parabola) separately.
 
There you have it, the parabolas are there... <br/ translate="no">


I notice that you are drawing parabolas for those channels that you selected by criteria, and then the question arises what this parabola means. The question is not really clear-cut. It depends on what we will use it for, or rather I have not yet understood why we need a parabola built from data of a channel that was selected without this parabola or any of the field characteristics. That's why I decided to select parabolas using the same scheme as for regression channels, thus defining a pivot zone at the intersection of a parabola and limits of confidence intervals of the channels selected before, but here I have another problem, because there are a few channels and a few parabolas which to combine them with I do not know yet.

Well... I've almost finalized the forced algorithm for calculation of channels, I just needed to give it a couple of touches to make it work properly; I already see that it takes 78 ms to calculate 5000 bars and it used to take about 10000 ms with the same accuracy.
 
<br / translate="no"> I noticed that you are drawing parabolas for those channels that you selected by criteria, and then the question arises what this parabola means. It's not really a straightforward question.


I agree, and now I think "may I take an umbrella or may I not take it". We are used to it, of course, and it is not a problem to make another version of the indicator. Let's add the RMS from a parabola and go (moreover it is also considered analytically). Yesterday I understood Vladislav's phrase "iterative methods of channel creation". 99.9% sure.
 
Two more pictures with parabolas

EuroFunt


and EuroOzzie

 
Meantime I have 16 consecutive losing trades and 1/6 drawdown from 150k to 125k. No open positions. :(
Rosh
... I just understood Vladislav's expression yesterday 'iterative channel-building methods'. I'm 99.9% sure.

Tell me, do your ideas are very much different from the algorithm shown by solandr at the end of 16 pp?
 
We'll have to see how the situation turns out.

 
Meanwhile on the empire 16 trades in a row with -1/6 drawdown from 150 to 125k. no open positions:(<br / translate="no"> Rosh
... I understood Vladislav's expression yesterday 'iterative methods of channel building'. 99.9% sure.

Tell me your ideas are very different from the algorithm shown by solandr at the end of 16 pages?


I have an idea so far, but no implementation. Most of the points I agree with his plan on page 16. I would probably use (but I should investigate it first) the first derivative and position of price in channel. And maybe abandon Murray and replace it with another. For now I'm thinking of looking at the channels in their pure form, without additional filters.
1...787980818283848586878889909192...309
New comment