Thoughts on some of the absurdity of multi-currency analysis. - page 25
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
I made an M1 ticks unloader with smoothing according to my own formulas - clearly there is a correlation between the pairs, but at what time and for how long
X - the pair was moving up, O - the pair was moving down - there was no change since the previous bar closing (it may be considered as the previous reading), time GMT+2
EURUSD and EURJPY move together very often
EURUSD and EURJPY see a joint move very often
I wrote yesterday an Expert Advisor, if a EURUSD and USDJPY candle closed on the downside (joint movement), then on the opening of the next candle we sell EURJPY.
I lost a lot of money :(((
I wrote an expert yesterday, if a EURUSD and USDJPY candle closed lower (joint move), then sell EURJPY at the opening of the next candle.
I am losing money :((.
I wrote an expert yesterday, if a EURUSD and USDJPY candle closed lower (joint move), then sell EURJPY at the opening of the next candle.
I am losing money :(
No, the principle is different here - while my Expert Advisor sees the joint move online on the zero bar, the tick collector just writes it into the file at the close of M1
I think I should make an opening on an open bar and not on EUROBAX because it has too much noise, it works much better on Aussie, I think I will extend my research to the USD a bit later
And what did you want to gain by opening a position based on a single candle that closed lower...
corrected by
And what do you want to get when you write an EA? Profit? No, I just wanted to check how inertial the market is. And I analyzed not one candle, but two, or rather one candle at a time, but on two pairs. I.e. the condition of selling the yen against the dollar and the fall of the euro against the dollar on the previous candlestick. Of course, the Expert Advisor was not written for 15 charts.
Sorry for the offtop, can anyone suggest a currency index indicator that has gold in it? And is there one, because currencies + gold buffers don't seem to be enough...
I've seen a modified one by Semsemic somewhere here, where nzd has been replaced with gold. But in principle it's not a problem to add more than 8 instruments. Of course, there are 8 mapped buffers, but no one forbids counting them along with 2 or 3 dozens of other instruments. I have rewritten CCFp in general for me - any number of arbitrary instruments (not necessarily Forex).
Thanks, that's the direction I'm going))
What is the correct formula for calculating the currency index? How to correctly calculate USD and EUR indices to see the movement in relation to them?
I tried Indexes_v8L, but it shows the same graphs, I think it's wrong.
So what's the problem - just download ticks using a formula and see if they correspond to terminal quotes or not
Thanks, that's where I'm going ))
What is the correct formula for calculating the currency index? How to correctly calculate USD and EUR indices to see the movement in relation to them?
Tried redoing Indexes_v8L, but it shows the same graphs, I don't think it's right.
http://www.forexltd.ru/ru/study/dollarindex/
http://www.spekulant.ru/archive/Indeks_dollara_rasschitat_ili_ugadat.html
I will try to outline my vision of index construction. However I doubt whether it is an index, but something very close to this concept.
This question periodically arises on a forum, I will try to set forth an approach to this question from the point of view of statistical test of hypotheses.
So let's take the data of 28.06.2010
EURUSD(Open=1.23781, Close=1.22803, Delta=0.00978)
In other words, EURUSD has grown by 978 points in one day.
(symbol (1) - growth, (-1) - fall, (0) - unchanged)
Form a complete group of hypotheses (H)
H0: EUR=0 USD=0 (unchanged)
H1: EUR=0 USD=1 (EUR has not changed USD has risen, etc.)
N2: EUR=0 USD=-1
N3: EUR=1 USD=0
N4: EUR=1 USD=1
N5: EUR=1 USD=-1
N6: EUR=-1 USD=0
N7: EUR=1 USD=1
N8: EUR=-1 USD=-1
As the exchange rate has risen, we can reject some of the hypotheses, and the hypotheses remain
N3: EUR=1 USD=0
H4: EUR=1 USD=1 (both rose but USD rose less)
N5: EUR=1 USD=-1
H8: EUR=1 USD=-1 (both are falling, but the dollar is falling faster)
Staying in the EURUSD two-dimensional space we cannot give any preference to any of these 4 hypotheses, we need to analyze other exchange rates. But once we move to the N-dimensional space the hypothesis becomes more complex, we should not forget about it.
If we take N currencies then H will look like this
N0: EUR=0 USD=0 JPY=0 GBP=0 GHF=0 AUD=0 .....
N1: EUR=0 USD=0 JPY=0 GBP=0 GHF=0 AUD=1 .....
etc.
Solving a problem head-on - testing hypothesis Н0 against hypothesis Н15 encounters mathematical difficulties (I failed), in statistics it is called a complex versus a complex (multiple choice) hypothesis test. In one of the books I came across the following recommendations. Try to test the simple hypothesis against the complex hypothesis, it is not always possible to do this, but in our case it works.
I.e. we have to check the simple hypothesis
H0: EUR=1
against complex hypothesis
N1: USD=0 JPY=0 GBP=0 GHF=0 AUD=0.....
In this formulation it is possible to solve, it is necessary to set first and second type error probabilities and to choose a criterion (maximum likelihood, ideal observer...). And with a certain probability the hypothesis H0 is accepted or rejected, and so for all currencies.
What do we get out of it?
Basically, it does not give much, we may actually determine that the movement is due, say, on Monday AUD and with a certain probability, not very high. Other hypotheses have an even lower probability. We can only assume that the AUD fall will continue - sometimes this works, sometimes it doesn't, it worked on Tuesday (it might not have).
My opinion is that this will give a slightly higher probability of a trade outcome, not 0.5, it is slightly better than nothing (0.5 is the spread).
Responding to the author of the thread, about the absurdity of multi-currency analysis, and the construction of indices. I would like to say that I don't see anything wrong or rediculous here. It has its place, and if this approach gives even a slightest advantage, one should use it. The above described approach to multicurrency analysis is not the only one, there are many of them ... and to reject them all is a bit presumptuous ...