1st and 2nd derivatives of the MACD - page 48
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
And if not, is the bell not normal?)
Who could have known that this combination would work well in this area? In general, this is not the correct expression, they all work the same way, but this one is more understandable. And I didn't bother choosing, so I just did a more or less understandable picture from the ballpark in 20 seconds. You can use all 10 lines if you want. You have a loaf of bread, you can cut it lengthwise, or crosswise, into 2 parts, or into 20, whatever you prefer. The main thing is that there is nothing left over.
The third part is not disclosed at all. The first two are covered at the level of general ideas, let everyone figure out the specifics for themselves. Whoever thinks it up, won't post it anywhere.
I was also thinking that you don't know when to take a particular combination and how long it will "work".
So, it's also risky (though less risky) to trade on this decomposition, the data is taken from only one symbol.
I think the main advantage of this decomposition is that it is more optimal and convenient for further representing and selecting pairs of currency indexes where the accumulation of index pairs increases the number of frequencies with certain properties.
For me it is still a mystery (almost) how an index is selected from several pairs.
If you use only euro pairs you won't get the euro index, you need other pairs (of course you may get them out of pairs with euro, that's what people who ignore spread, small differences and other details do, like you do).
But I think the anticipation would be even stronger with respect to the index, if you would take not only your 17 pairs and get other pairs from them, but the originals of all used instruments.
I came from afar, first I tried to use multicurrency analysis to detect the anticipation of one pair.
Yeah...
Just for example.
"True" medium is useful for understanding ARXINUSOFFs D)
I shammed and shammed with your phrase, but I didn't understand what you were trying to say, what does the picture say?
if i understood you at all, you must be referring to the construction of a Brownian bridge. this is not exactly what you mean, but if you go to that notion, then i would say that it is the analysis of the dynamics of expansion and contraction of this bridge and identification of the densest places in the cross section of this bridge that is interesting, rather than the bridge itself and its direction.
it is possible that the coefficient of variation - the width of the average bridge for each pair - could be used in constructing the index itself or something like that
but it seems to me something else.
It is called GARCH
when making a forecast based on the GARCH model, it makes a forecast for the volatility not of the time series itself (e.g. stock prices), but of the forecast errors (epsilon).
if that is what you mean, it is probably useful sometimes, but that is not the point, I have something else, I am not sure what it is yet))
when making a forecast based on GARCH model, it makes a forecast for volatility not of the time series itself (e.g. stock prices), but of the forecast errors (epsilon).
If that's what you mean, it's probably useful sometimes, but that's not the point, I have something else, I don't know what it is yet))
I was also thinking that one does not know when to take this or that combination and how long it will "work".
So trading on this decomposition is also risky (although less so), the data is taken from only one instrument.
The filter doesn't care what is fed to its input. Instead of decomposing a pair, you can decompose both indexes. This is what I end up with.
I am interested in one reference, I'll throw it here for now, who is interested can discuss. I understand that the offtop, but not to get lost, let it hang here, and maybe in a separate thread turns into a topic for a branch.
This reminds me of something http://en.wikipedia.org/wiki/Black%E2%80%93Scholes
I am not convinced by these pictures. Counter example, I won't even draw it. We take a phase with two periods. Let us enter along the high phase along the low-period one. We draw vertical lines exactly at pivot points if both ZZs coincided or shifted if we had to wait for a small ZZ.
This is what is drawn.
The problem is that there will be a lot of false reversals and we will not be able to hold the trend. HP has the same problem. And any other filters. New bar gives new frequency decomposition and corresponding amount of false inputs.
And this is the problem of a method that does not take into account statistical characteristics of the residue that rules everything - the tail wags the dog. This problem is recognised by all who deny the efficiency of markets.
There is a big difference. The filter redraws smoothly and the further back in time, the less. PZ redraws all of a sudden and immediately in the opposite direction. It should also be accompanied by a beep with the words "What, SUKA, weren't you expecting?!!!".
The residuals problem doesn't seem like a big problem to me. You can think of price as C[i]=Cs[i]+Cs[i], where Cs[i]-the part due to the market's own fluctuations, Sv[i]-the part due to external influences. What you call the residuals will go completely into part 2. It will definitely have interesting and useful properties. For example, its expectation is 0. It will help to earn or hinder with 50/50 probability. It must have clear links to the news, trading session time or something else. It can be defined and drawn as a separate line and surely it will be of much use. But I have not got to that point yet.
There is a big difference. The filter redraws smoothly and the further back in time, the less. ZZ is redrawn suddenly and immediately in the opposite direction. It should also be accompanied by a sound signal with the words "What, SUKA, you weren't waiting for?!!!".
Right. There is no need to shift into the future. We should create a filter with an odd number of N summands and shift by (N-1)/2. You will get a filter with zero phase shift, i.e. a perfectly matched filter. Besides an ideal price match, there is another reason for doing it this way - I won't say about it yet. And with such filters divide the whole spectrum into parts.
However, I keep thinking that the last (N-1)/2 bars of your lag-free filters are drawn on the assumption that the price will not change in the future. This part of the filter is the most important because you should make some assumptions about the future and the corresponding trade decision. You can of course look at individual filters-MAKDs and find out if they move up or down in this last section, but we already know in advance that they predict C(0)=C(-1)=C(-2)=... Also the process of choosing a pair to trade is unclear - they all have the same prediction into the future, which means they are all equal candidates for trading, or more accurately "not trading".