From theory to practice - page 607
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
is that my solution to the branch's topic is still the best.
but the fish aren't there and never will be.
There are fish, but not here. And I even feared at first that A_K would find and foul up the fish spots. The first 10 pages)). Turns out he won't.)
Undoubtedly the best option is to use tick quantisation, as it is primary.
I don't agree with that either.
There was the most fundamental research on tick intervals by Doc (where is he now?? apparently having read this thread, now he's just working as a janitor...) and Nova:
Forum on trading, automated trading systems and trading strategy testing
From Theory to Practice
Dr. Trader, 2018.03.27 23:15
Me and Novaja did a little research.
Ticks come to the terminal at some random intervals. Alexander wrote that it is important to find out what this distribution is. I took the tick history from mt5, so I can work with millisecond accuracy.
The distribution of pauses between ticks itself looks like this
"Approximately" is because the graph is averaged. Dilling distorts the real time of ticks. Either round them up to ~10 milliseconds, or cyclically change their generation rate. Before averaging, this graph (0-100ms window) looks like this
I saw these same peaks on Alexander's graphs, now it is clearer where they are coming from.
As a result it turned out that the averaged frequency graph can be described using the sum of Gamma and Cauchy distributions. The first parameter of gamma distribution for some dealings can be selected as an integer number and Gamma distribution becomes its special case - Erlang distribution.
This is a chart from another dealing:
black line - distribution as it is received from the tick history
red - average
the purple one is what is obtained according to the gamma+coshi formula.
It doesn't look perfect, but it also looks good on a logarithmic scale, and the top and tail generally coincide.
The distribution tail coincides well with tens of seconds
The formula:
gamma(4e+00, 2.8e+01) * 5e-01 + cauchy(0e+00, 7.37e+02) * 2e+00 * 5e-01
This is a dealer-specific formula, all have slightly different parameters and coefficients.
In general, the function looks something like this
function(k, Θ, γ, c){
gamma(k, Θ) * c + cauchy(x0=0, γ) * 2 * (1-c)
}
parameter c is from 0 to 1
So, this Cauchy from gamma distribution has to be bluntly "sifted out".
The result is a perfect sparse tick flow.
I don't agree with that either.
There was a most fundamental study of tick intervals by Doc (where is he now?? apparently, having read this thread, now works as a janitor...) and Nova:
So this one Cauchy from the gamma distribution has to be bluntly "sifted out".
The result is a perfect sparse tick flow.
Try to think about the fact that price movement events - ticks for the most part are not time-synchronized.
Decisions to make trades are made out of sync.
And you and the above researchers always have time on the horizontal axis.
And time, if we understand it as a measured order of events, on intervals of ticks is nearly always nonlinear.
Therefore all your statistics attached to linear time scale lame on both legs.
But on the intervals of a day or so the statistics will probably give results, because there is a rhythm there: the change of day, market opening time, the typical time of news release.
Figure out how to measure the horizontal dimension and you will be happy.
:-)
that I have the best solution for the topic of the branch anyway
but it still won't make you happy.
Late.
Late.
no
only the last tick is analysed, not even an indicator is needed
and still the chargeIn the intervals of a day or so, the statistics will probably give results, because there is a rhythm there: change of day, time of market opening, typical time of news release.
Figure out how to measure the horizontal dimension and you will be happy.
:-)
So we are talking about a moving time window dimension?
Bas sure it is a day. Most likely it is, because the flow of tick quotes in this case is almost Poisson.
Or are you talking about throwing the time out altogether and having something else horizontally?
Maybe so...
Frankly, I'm confused by the notion of a sliding time window in the market at all.
Von Koldun, for example, takes quotes and time and somehow gets an almost stationary series of something. He does not need any window at all. He simply puts this artificial series into the neuronet and replenishes his pockets.
And time, if you understand it as a measured order of events, is almost always non-linear at intervals of the order of ticks.
So all your statistics tied to a linear timeline are lame on both legs.
Think of something to measure the horizontal dimension and you'll be happy.
I offered it to Alexander 100 pages ago and 200 pages ago, and equal-typed bars, and adaptive-typed and equal-height bars, and even equal "path" price, but he didn't hear anything)
He will not hear it now)
and the physicist also confuses "size" and "dimension" )why are you doing this? was there any intrigue in the thread ))))
if the gsb is going around the price.
ok
more info
enter buy, price reacts downwards
or vice versa
enter in the sell, the price reaction is up
this never-ending faith in conspiracy theories.
I fundamentally disagree with this point.
There is no Brownian motion model in the market, where particle collisions are infinitely frequent and time intervals are distributed according to an exponential law. In this case it is legitimate to use a uniform time scale.
In the market there is no uniform timescale - there are Erlang flows of different orders. But, many people do not understand this and never will. The result is ripped, empty pockets.
You have learned that there are erlang flows in the market. the result is ripped pockets.