From theory to practice - page 150
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
Detrending and working from the borders of the channel to its centre - you call that a bicycle invented in the time of king peas the most ingenious solution?))) "How many wondrous discoveries the spirit of enlightenment prepares for us").
The success of this algorithm is relative. It will be successful at flat conditions. If you manage to identify the trend-flat and timely move from the counter-trend strategy to the trend, then it may be successful.
I would like to see in practice this no-return trend and how this algorithm deals with it.
I suspect that you simply calculate the observation time window (sample size) incorrectly.
For example, the 99.5% confidence level for the EURUSD pair is about 15.000, and for EURAUD - as much as 40.000!!! I.e., if I, foolishly, start to observe the EURAUD pair at volume = 15.000 ticks, which corresponds to approximately 95% level of confidence probability, then sooner or later the remaining 5% will finish my deposit - I agree.
Above I have posted charts for your data, which show that there is a memory of almost 40,000 ticks!
Yes, I have seen it. Thank you SanSanych - it matches my calculations.
A new tick comes in. Do we recalculate? On every tick, do we recalculate? The tick that was number 1 has become tick number 2 and is not in the sample. It is quite obvious, that ABSOLUTELY new ticks sampling will be made, as numbers of ticks in two exponents, which differ by one tick shift, will be different, i.e. all ticks are new! What does the presented figure refer to then? It turns out that the figures presented to us exist exactly for one tick!
You have written some nonsense, excuse me. There is nothing similar in the posted algorithm. I do not understand what you are moving there and why. What two exponents? New ticks? A new sample? What are you talking about?
Nikolai, some know-it-alls here are saying that this method is nonsense and has been known for 100 years. Do you know whether it is called an indicator or an adviser?
As for the time, it is a matter of principle, and I'll never tire of repeating it.
In my opinion - working with ticks indiscriminately is the worst mistake in time series analysis. The notion of time itself is lost; for one and the same amount of ticks at different stages you have different time and vice versa. It is sheer nonsense and as a consequence, the impoverishment and shame of the individual.
This leaves us with two ways:
1. To read data in equal intervals of time, and take the value of a guaranteed arrival of the quote as a discrete time.
2. Through exponential intervals - read about reducing a non-Markovian process to a Markovian one. This is exactly the trick.
This is a kind of oscillator from the bolinger or similar systems
Screenshots from MetaTrader platform
EURUSD, H1, 2018.01.22
FIBO Group Holdings Limited., MetaTrader 5, Demo
BB oscilator
Alexander, imho, you have chosen the wrong methodology to test the strategy. When it [strategy] already exists, it is easier to run it on different parts of history in Tester. As far as I understand, you went straight from primary code to online verification. It is very long and VERY expensive at today's speed. Time is expensive... That's why I somewhat understand the hustlers on the branch, who want their cakes ready hot here and now...
And all because there's probably no implemented strategy in pure MQL4\5.
You've written some nonsense, I'm sorry. There is nothing like that in the algorithm you posted. I do not understand what you are moving there and why. What two exponents? A new sample? What are you talking about?
Read my post again and if you want to answer, then answer in essence.
It's like an oscillator from Bolinger or similar systems
https://www.mql5.com/ru/charts/8199991/eurusd-h1-fibo-group-holdings-bb-oscilator
You subtract the waving from the quotient, and count the STO, and put aside both sides of the detrended quotient.
In the meantime, I'm about to come up with the most ingenious idea.
This value - the sum of the increments - is the amplitude of the probabilities of the wave price function. In fact, this value is the solution to the Fokker-Planck equation with boundary conditions.
After statements like that, you should take your feet off the forum... ^)))))
Alexander, imho, you have chosen the wrong methodology to test the strategy. When it [strategy] already exists, it is easier to run it on different parts of history in Tester. As far as I understand, you went straight from primary code to online verification. It is very long and VERY expensive at today's speed. Time is expensive... That's why I somewhat understand the hustlers on the branch, who want their cakes ready hot here and now...
And all because there's probably no implemented strategy in pure MQL4\5.
In the meantime, I'm about to come up with the most ingenious idea.
This value - the sum of the increments - is the amplitude of the probabilities of the wave price function. In fact, this value is the solution to the Fokker-Planck equation with boundary conditions.
After statements like that, you should take your feet off the forum... ^)))))
Who is this addressed to? A spherical horse in a vacuum?