Distribution of price increments - page 11
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 have repeatedly written in this thread too - if you think that trade (play) with cheaters, then staying in the game, what should we call you?
It's not a fact that I think so. I offered only one version :)
P / S. And if you do not think that trade (play) with cheaters (but in reality so), then staying in the game as you can be called?
Don't argue over nothing, traders!
I hope that Forex is not a game, as long as the preconditions for such a statement are not in place.
What saves us is the non-marking nature of the pricing process - and we will definitely take advantage of that!
Don't argue over nothing, traders!
I hope that Forex is not a game, as long as the preconditions for such a statement are not in place.
What saves us is the immorality of the pricing process - and we will definitely take advantage of it!
In the complete set of the terminal all of the indicators imply unmarked pricing, i.e. the developers of the terminal (any one with technical indicators) are already aware of the presence/absence of the carrot.
There is a theoretical assumption that markets are fractal, on small TF one may observe the same processes as on large ones - it has not been argued yet, may be it is worth to argue? (humour) https://ru.wikipedia.org/wiki/Фрактальный_анализ_рынка . I.e. someone has already thought about and proved the unmorcidity of price formation, at least since the advent of candlestick analysis, when the "tick" waited for a week or a month - figuratively, no carrot.
Why would one collect an obviously individual tick sample from one dt/broker (another dt/broker's tick sample in the same interval may differ three or five times), if all the dt/brokers have samples from minutes to half an hour will have the same number of quanta? Minutes are enough to check the theoretical assumption, the history by minutes can be found for 3 or more years - more than enough for statistical elaboration. I assume that the aim is to find some smaller obviousness than the opening(closing) position before the double swap, i.e. discretization of 1/2 week, like new day 1/5(6) week, opening/working time of the exchange 1/3- 1/2-1 day.
For what I gave an example with the index (only 4 pairs) - dc/broker introduces in the conventionally original quotes some change, which has a certain pattern, from which dc/broker due to its business model can not (c)omit, which you can physmatically extract and use for infinite enrichment (yourself or dc/broker).
- Practically, a minute is enough, mt's terminal functionality (if only it didn't recalculate its indicators by ticks) is enough.
- If you become an IB, your personal scalping broker/dealer will close your eyes.
- Find a pattern/loophole between options premiums ws market - any bank will buy you.
- Not enough, do the filter calculation on xilinx - $200, couple of months and you're almost king of the stats.
- Stick Perelman in xilinx sell it on amazon it's like a bluetooth whistle to a smartphone/tablet mt4/5(niche free) - you'll be a dollar multi-millionaire.
In 2008 the mmwb/rts index started to collapse, never looked at charts of these indices before, caught an acquaintance watching this chart - my summary of the situation around 1200 was "buy at 900, at 600 and leave to 300 for everything and cover the whole ***", the summary of a physicist and qualified financier was "YOU WHAT, this is not possible!!!! It's economics" - maybe I'm a genius? no, just about any illancier would set up the (grid)order(s) like that. PE/finance majors didn't teach grid/ilan.
The pound isn't the same anymore (salt is sweeter, sugar is saltier, mql5 was invented etc), take dax look at it for three months - the idea of ticks will fall away by itself.)))
Forum on trading, automated trading systems and trading strategies testing
Price Increments Distribution
Maxim Dmitrievsky, 2017.11.06 17:10
Wouldn't it be like reinventing the wheel here, like RSI in another thread... by complicated scientific paths :) and trying to reinvent another market model
Or maybe just a new run as a proof of the Dow axiom?
Data collection for the two methods of calculating tick-tock timing is ongoing. I will only be able to do the analysis at the weekend.
Why collect data when you already have it?
About the question "Distribution found, but how to apply it?"
I am interested in machine learning to create a trading robot with neuronics. One of the problems is that neuron can be easily trained to predict some price increases, but there is no guarantee that it will work correctly on new data, this problem applies to any trading robot in general.
There is a theory that neuronka predictions should fall into the same distribution as the input data, i.e. for example we take thousands of ticks (or open prices), train the neuron to predict the price increase for each tick and then predict the training data to determine how much it is wrong. Usually r^2 metric is used to estimate neuronka accuracy, but besides that you can compare distributions of original price increments and predicted price increments; if distributions don't coincide then neuronka is most likely to fail on new data. I don't compare distributions myself yet, "similarity of distributions" is too cloudy metric for me. But it could be quite useful for those who understand it. And if I understand correctly, the previously mentioned garch model also does such comparison of distributions.
There is a theoretical assumption that markets are fractal and that the same processes can be observed on small TFs as on large ones - so far it has not been challenged, maybe it should be challenged? (humour) https://ru.wikipedia.org/wiki/Фрактальный_анализ_рынка . I.e. someone has already thought about it and proved the non-morkiness of price formation, at least since the advent of candlestick analysis, when the "tick" was waited for a week or a month - figuratively, no carrot.
Or maybe just a new run as to prove the Dow axiom?
Excellent and timely comment! Obviously, it is - how could I not have guessed it right away! (I say without irony - I can only be excused by the fact that I have relatively recently begun to seriously engage in forex processes)
And do you know what a fractal is in terms of probability theory? It is nothing else but a superposition ("mixture") of homogeneous distributions. Thus, we can conclude that the probability distribution of Ask or Bid price is a superposition of Student's distributions with 2 degrees of freedom.
So, I formulate basic hypotheses about the processes in the Forex market, which can be considered proven empirically and experimentally (in fact, the person who proves these hypotheses in analytical form, can safely go to the Nobel Committee for a prize :))))
1. The process of Ask and Bid price formation is non-Markovian.
In practice - all Expert Advisors, indicators and advisors that do not consider analysis of historical data (such as Bollinger Bands, Fast Fourier Transform, etc.) can not be considered from the word "at all".
2. The probability distribution of price increments (returns) in the flow of tick quotes is a discrete, asymptotically described by the Student's distribution with 2 degrees of freedom and the quantile functionQ(p) = 2*s*(p-1/2)*sqrt(2/a), where a=4*p*(1-p), s - nonparametric standard deviation.
In practice -all EAs, indicators and advisors that use Gaussian normal distribution in their calculations (as well as other classical distributions), "3 sigmas" rule etc. can also be ignored.
3. Theprobability distribution of the Ask or Bid price is a superposition of the Student's distribution with 2 degrees of freedom.
In practice, it is a cool task to extract specific distributions from the superposition.
Actually, based on analysis of historical tick data, or simply by averaging statistical parameters at certain sample sizes, a conclusion about the current price value exceeding certain historical boundary conditions is made. Only after that, the current distribution parameters - dispersion, kurtosis, skewness, etc., etc. are analyzed in order to find out whether a new Student's distribution has begun or has already finished. In the first case - the deal is made following the trend, in the second - against the trend.
Sincerely,
Alexander_K
Data collection for the two methods of calculating the timing of tick data reception is ongoing. I will only be able to do the analysis at the weekend.
For those who are interested in the topic, I can inform in advance about the criteria of the future analysis.
1. we will use different samples of a certain volume (in the Vissim simulation system - the maximum volume is unfortunately only 16384... Does anybody know what the maximum sample value is for calculating medians or weighted averages in MathLab for instance? ).
2. The current parameters would be moving median, moving arithmetic mean, moving variance, Pearson's moving asymmetry coefficient for a given sample size.
3. Averaged variance and averaged Pearson's skewness ratios on a statistically significant interval, a month, will be used as integral parameters.
And (VERY IMPORTANT on the subject) - just returns - price increment at a given step of the process - will be used as a differential parameter.
1) What kind of a thing is Vissim? Matlab takes arrays of huge size. I've loaded over 200k.
This is the cool stuff! It's a dynamic process modelling system. I use it in my work.
The only thing - it has a restriction on the window for calculating moving averages and medians (functions median(x) and mean(x)) - I don't know why - I'm afraid it's not enough for analyzing tick data in dynamics.
This is the cool stuff! It's a dynamic process modelling system. I use it in my work.
The only thing - it has a restriction on the window for calculating moving averages and medians (functions median(x) and mean(x)) - I don't know why - I'm afraid it won't be enough for analyzing tick data in dynamics.
Matlab has a similarSimulink package. The convenience lies in the linkage with Matlab itself.
Alexander, a methodological question: why take ticks? There is a lot of noise there. It is worth taking close prices, imho. Or rather the returns from close prices.