ERUUSD spectra - is this proof of non-stationarity? - page 7
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
The market will be different on each of the next 240 bars.
If we follow Peters, we should take that part of BP which has memory, build a TS on this part and the TS will be invariant with respect to the following displacements along the BP. IMHO the size of the window is such that it represents all the frequencies that characterise BP, and it is neither shorter (not all frequencies) nor longer (some frequencies are repeated several times).
Mark Twain, How I Edited an Agricultural Newspaper
"(real editor).
- You don't distinguish between harrows and furrows; your cows lose their plumage; you recommend the domestication of ferrets, as these animals have a cheerful disposition and excel at catching rats! You write that oysters behave calmly while music is playing. But that remark is unnecessary, completely unnecessary. Oysters are always calm. Nothing can throw them off balance. Oysters know nothing about music. Oh, thunder and lightning! If you made it your life's goal to improve in ignorance, you couldn't be more distinguished than you are today. I've never seen anything like it. Your report that the horse chestnut is rapidly gaining ground as a marketable commodity could ruin the paper forever. I demand that you leave the paper immediately. I don't need any more leave - I wouldn't be able to use it under any pretence anyway as long as you sit in my place. I would be trembling with fear at the thought of what exactly you would advise the reader in the next issue of the paper. My eyes would go dark as soon as I remembered what you wrote about oyster cages under the heading "Ornamental Horticulture". I demand that you leave at once! My holiday is over. Why didn't you tell me you knew nothing about farming?
- Why didn't you, pea pod, cabbage pod, son of a pumpkin? This is the first time I've heard such nonsense. Let me tell you something: I've been an editor for fourteen years and this is the first time I've heard that a man has to know something in order to edit a newspaper."
A stable spectrum is when the market has cycles with a constant period with a constant phase change. Cyclicality is a very strict condition, like how day and night on earth are stable in time :) Just periodicity is a less strict condition. It is e.g. when some process has a fixed time period, but it does not have to repeat cyclically - it can start at almost any time. There is cyclicality in volatility and it is related to the different presence of players in the market at different times of the day. But cyclicality in periods of growth/decline, where does it come from? It does not exist even in a short sample - it is a coincidence.
>> tell your grandmother about the results of your research
What's the instrument(s)? Let me check for a signal.
A stable spectrum is when the market has cycles with a constant period with a constant phase change. Cyclicality is a very strict condition, like the Earth's day is stable :) For example, just periodicity is a less strict condition. It is e.g. when some process has a fixed time period, but it does not have to be cyclic - it may start at almost any time. There is cyclicality in volatility and it is related to the different presence of players in the market at different times of the day. But cyclicality in periods of growth/decline, where does it come from? It does not exist even in a short sample - it is coincidences.
I'm not talking about cyclicality, it's not electricity. Rather I am talking about the representativeness of the BP sample. The resulting spectrum characteristics will be roughly equal at shifts not necessarily per period.
I'm not talking about cyclicality, it's not electricity. Rather, I am talking about the representativeness of the BP sample. The resulting spectrum characteristics will be roughly equal at shifts not necessarily per period.
They will not stand out against other oscillations close in the spectrum. The signal will be inseparable from the noise. Only if there is a set of frequencies (or periods) that will be outnumbered by others.
They will not stand out from other oscillations close in the spectrum. The signal will be inseparable from the noise. Only if there is a set of frequencies (or periods) which will be outnumbered by others.
On this forum a curious state was once shown. On a length of 1 day they measured the lengths of candlesticks and in the same axes they plotted these dimensions over several days. The absolute magnitudes of the candle lengths differed from day to day, but their shapes were remarkably similar! I think the spectrum of each day differed little as well. If one could find such a stretch of BP, the SPM would have similar characteristics - that's what it's all about.
On this forum, a curious state was once shown. On a length of 1 day, they measured the lengths of the candles and in the same axes they plotted these dimensions over several days. The absolute magnitudes of the candle lengths differed from day to day, but their shapes were remarkably similar! I think the spectrum of each day differed little as well. If it were possible to find such a stretch of BP, the SPM would have similar characteristics - that's what we're talking about.
Yes maybe the spectrum and periods do stand out for certain stretches of history and even that would not be a coincidence. For example, it's been flat lately and most likely you can indeed find "important" periods in the increments. But it does not mean that you will find such moments in time and you will be able to profit from them. The same may be for some trend areas. But you will always find this spectrum with a delay and apply it when it has already changed (you will learn about it later). You can find out the moments when the market has changed only after the fact.
Yes, the spectrum and periods may indeed stand out for certain parts of the history and even this will not be a coincidence. For example, it has been flat recently and it is likely that "important" periods can indeed be found in the increments. But it doesn't mean that you will find such moments in time and earn on them. The same may be for some trend areas. But you will always find this spectrum with a delay and apply it when it has already changed (you will learn about it later). You can find out when the market has changed only after the fact.
So, using sliding window price filtering is not reasonable, because the spectrum is constantly drifting? Or is it possible to use adaptive filters? What is the criterion for adapting them?