1st and 2nd derivatives of the MACD - page 5

 
gpwr:

Thephysical meaning in the 2nd derivative of the MACD is hard to find. MACD is calculated by subtracting the long waveform from the short waveform. Since the mask is a low-pass filter, with this extraction we get an intermediate-pass filter. So we are looking at 2nd derivative of bandpass filter. Or we are looking at the acceleration of a fast sweep relative to a slow sweep. But still I'm haunted by a question: why do we need all this?

1st derivative - speed, 2nd derivative - acceleration, do you agree (in the usual sense), by the acceleration we can analyze the force of inertia?
 
trol222:

The 1st derivative is speed, the 2nd one is acceleration, do you agree (in the usual sense), we can analyze the force of inertia by acceleration?

One friend of mine forced me to write an impulse indicator, or rather the second derivative of the price. Now he successfully uses it in other systems (not related to MT4)
 
gpwr:

The MACD is calculated by subtracting the long waveform from the short waveform. Since the bag is a low-pass filter, this results in an intermediate-pass filter.

Unfortunately this is not the case. The LPF shifts the phase. Because different wipers shift it differently, the IACD is not a bandpass filter. It passes lower frequencies that the PF is not supposed to pass.
 
AlexeyFX:


There are an infinite number of different filters, so you can find just as many pictures if you want. If the filter is applied to the price, there is little to judge from the picture.

This is roughly what 99% of traders in the world do. Let's take some indicator, run the price through it, observe it, come up with unsubstantiated entry and exit points. Let's adjust the parameters and use the tester. When we start losing profit, we again fine-tune it. For 5-10 times we will say that the market has changed and we will use another indicator. This can be done endlessly.

This all instead of setting something more simple (e.g., sine) and see what the indicator actually does with the input signal.

And it does something like this:

Here and further the green line is the input signal. These are filters, similar to MA. Absolutely all known LPFs behave similarly. You can tweak the parameters as much as you like, nothing will change in principle.

But these are MACD-like filters. It becomes clear why some people prefer to trade without indicators. It's better without indicators than with such indicators.

And this is what I think filters that are more useful than harmful should look like. They don't seem to be there, but they are there, just covered with a green line.

And your own words:

By the way, by anticipation I mean only being able to calculate some regular component of the price due to its past values. I have never tried to get some kind of signal based on data from other pairs or indices before a pair reaction occurs. I do not believe in it and do not see such a need.

 
AlexeyFX:

Unfortunately this is not the case. The LPF shifts the phase. Since different mashups shift it differently, the MACD is not a bandpass filter. It passes lower frequencies that the PF should not pass.

Well, the MACD is technically a bandpass filter. Albeit a lousy one.

trol222:
What is the 1st and 2nd derivatives of the MACD equal?

MACD derivatives are meaningless. Because MACD has no explicit (sinusoidal) oscillations. Or sine-like. Quality factor of such a filter is too low, or selectivity if you make a selective filter.

The first derivative of sine is cosine. Thus, you can look half a period into the future.

 

trol222:

If the filter is applied to the price, little can be judged from the picture.

This is roughly what 99% of traders in the world do.


And your own words:

By the way, by anticipation I mean only being able to calculate some regular component of the price due to its past values. I have never tried to get some kind of signal based on other pairs or indices before a pair reaction occurs. I do not believe in it and I do not need it.


And what is the problem? I meant that it is difficult to judge the quality of the filter by the picture where prices are fed to the filter input. And 99% of traders look at these pictures without knowing what to compare it with or what it should look like.
 
trol222:

The 1st derivative is speed, the 2nd one is acceleration. Do you agree (in the usual sense), using acceleration we can analyze the force of inertia, the initial problem - the price line, we replace it with the line of Maqdi, but not between the fast and slow MA, but between the MA and the price (lots).

Simply put :

speed V1=macd[0]- macd[1];

V2= macd[1]- macd[2];

A1=V1-V2;

And the meaning is very simple and straightforward. If V1+A1<0 on the current bar, then on the next one it will peak.

Good luck

 
vlad1949:

Simply put :

speed V1=macd[0]- macd[1];

V2= macd[1]- macd[2];

A1=V1-V2;

And the meaning is very simple and straightforward. If V1+A1<0 on the current bar, then on the next one it will peak.

Good luck


So it's speed + acceleration? Is that the right word?

I thought so cloze, ma1,ma2,ma3,ma4........

macd0=ma1/close-1, macd1=ma2/close-1, macd2=ma3/close-1, macd3=ma4/close-1 (

we get something like speed (for ma)(for macdi - path), then

acceleration0=macdi1-(maybe we should divide?)macdi0, acceleration1=macdi2-macdi1, acceleration2=macdi3-macdi2 (acceleration for ma, for macdi - speed)

overacceleration (oil-oil-oil)) acceleration for macdi)0=acceleration1-acceleration2, overacceleration1=acceleration2-acceleration1....

The highlighted line is interesting - at this point is it worth doing a separation into ma2, ma2-ma1, ma1-close

and in general, maybe it would be better to do the following

there is a quote, we derive ma from it, then

macdi0=ma1/cloze-1, macdi1=ma2/cloze-1, macdi2=ma3/cloze-1, macdi3=ma4/cloze-1 (

on obtained values we again build a trend (the sum of all positive values)

on this trend, we again build Ma, and then from these Ma values we again build

macdi0=ma1/cloze-1, macdi1=ma2/cloze-1,macdi2=ma3/cloze-1, macdi3=ma4/cloze-1 (

 
vlad1949:

Simply put :

speed V1=macd[0]- macd[1];

V2= macd[1]- macd[2];

A1=V1-V2;

And the meaning is very simple and straightforward. If V1+A1<0 on the current bar, then on the next one it will peak.

Good luck


By the way, what happens if C1=A2-A1,

E1=C2-C1

 
Guys, you won't get anything good from this, even if you take triple or decimal derivative. Yes, the first and second derivatives sometimes show the top very quickly (without delays), but it is not long-term. In general at one time long twisted this chip, came to a conclusion that it is useless.....