Recognising images ( rhetorical theme ) - page 2
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Yura... I'm new to programming, MKL is the only language I know a little bit at beginner level, and believe me, I've tried all the graphical shapes, triangles, vertices, head-shoulders and many other patterns
ZS even managed to describe a diver, though I thought it would fail with my knowledge of the language
probably ... why not?
I had an experience of searching in a repeating pattern with a fan of averages
i got a fan of averages and as soon as the line tensions on a certain range (a certain number of bars ) were the same, i got a signal
--
In general, almost half of indicators, if not more, have a standard average in their database...
You need to collect statistics on the figures for the last 5-10 years. You can take a batch of 5-15 bars and play with their 3 (or 4) parameters in the tester. If it turns out that there are combinations that pour in high quality trades, then a set of these combinations can be tried to be applied in an EA. I did something similar but was disappointed with the results of statistics.
You need to collect statistics on the figures for the last 5-10 years. You can take a batch of 5-15 bars and play with their 3 (or 4) parameters in the tester. If it turns out that there are such combinations that fill with high quality of trades, a set of these combinations can be tried to be applied to an EA. I was doing something similar but was disappointed with the results of statistics.
Gathering statistics must be based not only on the repeatability of shapes but also on their scale
The market changes ... let's take 2007 - the Euro was moving from the high to the low at the average rate of 60 pips.
2009-2010, the Euro was averaging 120p.
it is the same with the figures - they should be scaled up at least by the average move in the last sector
Gathering statistics should not only focus on the repeatability of figures but also on their scale
The market is changing ... let's take 2007 - the Euro moved from the high to the low by about 60 pips on average.
2009-2010, the Euro was averaging 120p.
the figures are the same - they should be scaled up at least by the move average in the last sector
Yuriy, answer me the last question too :) Really interesting.
It is the same with figures - they should be scaled at least by the average stroke parameter on the last section
Yuri, I don't get it backwards... what do you mean "customers want"? They want to recognise shapes on the price line, but necessarily with indicators? (then the question is - do they care as long as it is recognised correctly). Or do they want to recognize figures on the indicator lines? Or search for divergences?
I think they don't care how to recognize it ...
because judging by the problem statement: SELL AT A DOUBLE TOP...
so in the order there is no DESCRIPTION of how to recognize it, what is a double top ...
the customer is relying on me to figure out how to recognize the double top...
and if I do not recognize it correctly :-) then of course I am wrong...
--
I usually don't take an order if the customer writes:
"Here it curved like this and here the average entered from the bottom you... look at the stochastic, it bent like this, in this case sell!!!
and before that, on H4, there were two fractals from above yesterday and the day before that (or maybe not) and even the fractals may not have been there
--
Divergences, have less clear criteria...
Once again, different divergence indicators have different CRITERIA, (because different indicators may be based on it)
the first indicator will show a divergence in one place and the other one will be silent in the same place
Right, it's one thing to do with shape and another to do with size. But, it is not a fact that if you change the size with the shape unchanged, the "value" of the bar will not change.
K= (HL, LO/HL, LC/HL);
here the hylum parameter can be made relative rather than absolute. I.e. consider it as a fraction of the maximum highlaw value in the history of trades. However, the maximum highlove will be "non-market", i.e. some kind of "hairpin". In addition, the analysis accuracy will be significantly reduced.
I describe candlesticks and bars with 3 parameters.
The parameter HL is absolute and expressed in pips. The other parameters - LC and LO - are relative and expressed as fractions of HL, HL is taken as a unit. A complex number is obtained, e.g:
K= (HL, LO/HL, LC/HL);
or
K= (360, 0.128, 0.746);
Thus any bar is described by a complex number and is easier to analyse.
Pattern recognition programs are complex and cannot be written "in an hour". This work is expensive. But, if the complex task is broken down into several smaller ones, it is doable. The only question is what the "price" will be for this recognition. Most likely it will turn out to be somewhere between 55/45 to 45/55.
By the way, Yuri, have you ever tried to build bars on the zig-zag?
K= (360, 0.128, 0.746); - here Open < Close because 0.128<0.746
K= (360, 0.746, 0.128); - here Open > Close because 0.746>0.128
This is quite enough to describe the size and shape of a bar (or candlestick).