On the unequal probability of a price move up or down - page 128

 
Alexsandr San:

If I understand correctly, the author is waiting for a signal from several pairs?

something like that?

of course there is a difference from signals


there is an interesting difference -

- well anyway, it's a fortune teller

GBPUSDH4

 
khorosh:

It is difficult to make the filter lag little and at the same time effectively suppress short-term false price spikes. There are non-lagging MAs, but they overdraw, although they look very nice on history.

How do you like such a filter? The black colour is for the input signals, the red colour is for the filtering result.

 
Mikhael1983:
How do you like this filter? In black are the input signals and in red are the filtering results.

Where's Mihalych's moniker?
 

You'd say it's not particularly impressive, there's a lag, albeit a slight one? I would agree. However, I will note that there are no filter settings that somehow tell it what signal is expected to filter.

However, here is its operation on rectangular pulses or steps (hello short bursts):

The original signals (black curves) are not visible, because they are completely covered by the red ones (the result of filtering)

 

And finally, on the real quotes:

 
Mikhael1983:

And finally, on real quotes:

how shall I put this correctly...

er, if the phase shift is known (assumed) in advance, the filter at that frequency will not "lag".

It is not certain that it will show anything sensible, but in its madness it will not "lag".

 
Maxim Kuznetsov:

how shall I put this correctly...

er, if the phase shift is known (assumed) in advance, then the filter at that frequency will not "lag".

It is not certain that it will show anything sensible, but in its insanity it will not "lag".

The filter has no information about the input signal. No frequency information, no phase information (whatever that means in your mind), no volatility information, no sign or anything else. None at all.
 
Mikhael1983:
The filter has no information about the input signal. None at all. No frequency information or phase information, whatever that means in your mind.

I was just curious to see a filter like this...

And what does it filter then? It has some parameters - and in them you put your bad assumptions about the original signal and sweet dreams about the resulting signal.

 

On the Excel PRNG graph, draw a filter and press F9. Maybe you'll see something there too:)
But seriously, someone even tried to attach Kalman. As for macd variants I'm even not mentioning their innumerous number.

Of course, there are doubts about the market phase.

 
Oh, I feel like you're going to get swamped with digital filters. Michael... It was invented a long time ago.