Trying to distinguish between real currency mark-up and illusory currency mark-up and recognising the simultaneity of the pairs' moves. - page 6
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
Do you mind if I join the discussion? Something is getting interesting.
By the way, yes, the reduced session time in the forum adds to the overall picture.
The currency state vector proposed by Alexey has information entropy - assuming equal probability outcomes of its states - 14.26466 bits. But apparently, the probability density function of its states is not uniform, so the entropy will be less.
How to formalise the trend in this formulation? Sum of elements greater than zero for upward trend and less than zero for downward trend?
Further on, I see the following generalized scheme of actions. We calculate the probability of the trend events considering the sign, of course. And then we move on to predicting the trend of the currency.
saying after each action - I'm not crazy ..... I'm not crazy....)))
I didn't want to get anyone off the ground with my posts. The point I wrote earlier: we introduce a working hypothesis: if there is a trend for a currency (for a set of pairs), it will continue. We check it on the data, formalizing the notion of a trend in one number, and this number characterizes the movement vector of currency pairs. Then we check if our hypothesis is viable by opening deals one step ahead (for example, all calculations on one hour bars and a deal opens on one hour) on all pairs.
I will write below in small print, because the topicstarter did not understand this nonsense. If he doesn't mind, let's keep raving.
Here again we have TH, which few people want to hear about :( But now everything is even more twisted: there are already two different entropies - thermodynamic and informational.
On the subject of formalisation: I don't know yet myself. If we define a macrosignature as a dumb sum of codes (for example, for <+1,0,-1,0,+1,-1,0,0,+1> it is 1), then it is logical to consider as a trend a microstate that is representative of a macrosignature macrostate with a macrosignature that is significantly different from zero.
Again on the terms:
A microstate is a vector in which all currency codes in all pairs are ordered. It is for example <+1,0,-1,0,+1,-1,0,0,+1>. The signature of this microstate is {-2,4,+3}, i.e. two -1, four 0 and three +1.
A microstate always corresponds to the real situation on the market and vice versa. The probability of absolutely any microstate is a constant, due to the very way the codes are calculated (by quantiles), and it equals 3^(-9) for nine pairs. But we all know very well that a strong trend and an ordinary flat are not at all equally probable states. So we have to move to macro states to count the probabilities.
A macro state is the totality of all micro states with identical (=equivalent) signatures. There are a lot of them here, but I can cite a few of them to make it clear:
<+1,0,+1,0,+1,-1,-1,0,0,-1>
<0,0,0,0,-1,-1,+1,+1,+1,+1,+1,+1>
<+1,+1,+1,-1,-1,-1,0,0,0,0,0>, etc.
The macro-state is entirely determined by the signature, unlike the micro-state. The macrostate is needed to estimate probabilities (that's how it seems to me).
Transitions between equivalent microstates (i.e. within the same macrostate, with the signature intact) are free and unpredictable.
Transitions between macrostates are also unpredictable, though not so free.
There is a hypothesis that the thermodynamic entropy of the macrostate is a characteristic tending to inertia, i.e. it may make sense to talk about it in the sense of prediction.
Here is one of the strangest examples of entropy (thermodynamic) inertia: sometimes before very strong news like trailers there is an extremely hard flat (dead calm), when almost all pairs fluctuate within some pips. Suppose at some moment this is a microstate <0,0,0,0,0,+1,0,0> to which corresponds a macrostate with macrosignature {-0,8,+1}. This is a very unlikely macrostate with a probability 130 times smaller than the probability of a typical flat {-3,3,+3}. Its t/d entropy (logarithm of the number of equivalent microstates) is ln( 9!/(0!*8!*1!) ) = ln(9) ~ 2.20.
The market explodes on the quid (suppose the quid rises) after the news is announced. This is, let's say, the microstate <+1,+1,+1,+1,+1,+1,0,+1,+1>. It corresponds to the macro state {-0,1,+8}, whose t/d entropy is ln( 9!/(0!*1!*8!) ) = ln(9) ~ 2.20, i.e. the same!
T/d entropy has not changed at all, although the nature of the market has changed dramatically. But something has changed: it is the macrosignature. It was close to zero (dead calm), and now it's 8.
I think it would be very interesting to see how the micro-state of the system changes over time. Maybe there are some patterns.
Is it already after the login?
Below I will write in small print, for in this nonsense topikstarter did not understand. If he does not mind - we will continue to rave.
Here we have again a TI, which few people want to hear about :( But now it is even more twisted: there are already two different entropies - thermodynamic and informational.
On the subject of formalisation: I don't know yet myself. If we define a macrosignature as a dumb sum of codes (for example, for <+1,0,-1,0,+1,-1,0,0,+1> it is 1), then it is logical to consider as a trend a microstate that is representative of a macrosignature with macrosignature significantly different from zero.
Again about terms:
Microstate is a vector in which all currency codes in all pairs are written in an orderly fashion. It is for example <+1,0,-1,0,+1,-1,0,0,+1>. The signature of this microstate is {-2,4,+3}, i.e. two -1, four 0 and three +1.
Microstate always corresponds to the real market situation and vice versa. The probability of absolutely any microstate is a constant, due to the very way the codes are calculated (by quantiles), and it is 3^(-9) for nine pairs. But we all know very well that a strong trend and an ordinary flat are not at all equally probable states. So we have to go to macro-states to calculate probabilities.
Macro-states are the totality of all micro-states with identical signatures (=equivalent). There are a lot of them here, but I can cite a few of them to make it clear:
<+1,0,+1,0,+1,-1,-1,0,0,-1>
<0,0,0,0,-1,-1,+1,+1,+1,+1,+1,+1>
<+1,+1,+1,-1,-1,0,0,0>, etc.
The macrostate is entirely signature driven, unlike the microstate. The macrostate is needed for estimating probabilities (it seems to me so).
Transitions between equivalent microstates (i.e. within one macrostate, with signature preserved) are free and unpredictable.
Transitions between macrostates are also unpredictable, though not so free.
There is a hypothesis that the thermodynamic entropy of the macrostate is an inertia-prone characteristic, i.e. it may make sense to talk about it.i.e. it can be useful to speak about it in the sense of forecasting.
Here is one of the strangest examples of entropy (thermodynamic) inertia: sometimes before very strong news like trailers there is an extremely hard flat (dead calm), when almost all pairs fluctuate within some pips. Suppose at some moment this is a microstate <0,0,0,0,0,+1,0,0> to which corresponds a macrostate with macrosignature {-0,8,+1}. This is a very unlikely macrostate with a probability 130 times smaller than the probability of a typical flat {-3,3,+3}. Its t/d entropy (logarithm of the number of equivalent microstates) is ln( 9!/(0!*8!*1!) ) = ln(9) ~ 2.20.
The market explodes on the quid (suppose the quid rises) after the news is announced. This is, let's say, the microstate <+1,+1,+1,+1,+1,+1,0,+1,+1>. It corresponds to the macro-state {-0,1,+8} whose t/d entropy is equal to ln( 9!/(0!*1!*8!) ) = ln(9) ~ 2.20, i.e. the same!
T/d entropy has not changed completely, though the market character has changed dramatically. But something has changed: the macrosignature. It was close to zero (dead calm), and now it's 8.
I think it would be very interesting to see how the micro-state of the system changes over time. Maybe there are some patterns.
Alexei what are you like little.... What's with all the ponzi scheme with the shrinking of the font... I very much appreciate your rare statements about my thoughts and I do not feel any aggression or disrespect to you..... It's just if one person knows that 2+2=4 and the other does not - you should not judge him for that I am in search of....
I'm sorry I am not gifted with mathematical wit.....
For God's sake stop raving (in a good way).
By the way, more pictures would be great.
Corrected the font in the original post.
Problem with pictures: entropy (any) is not easy to visualise:)
But from you, Valera, pictures would be extremely useful: the topic is yours.
Corrected the font in the original post.
Problem with pictures: entropy (any) is not easy to visualise:)
But from you, Valera, pictures would be extremely useful: the topic is yours.