Tics: amplitude and delay distributions - page 3

New comment
 
How does it look in your humble opinion?
 
Mathemat:
I downloaded the data from http://ratedata.gaincapital.com/ for several different weeks and tried to analyse it. It's an interesting story, though!

Here is the second week of April, 9 to 13 April 2007. Total is 27516 ticks, i.e. slightly less than 4 ticks per minute on average. And here are the statistics (the number stands for the difference between the current tick and the previous one):


It's an interesting study. Can you look at more sequences. Continuous pluses, continuous minuses, etc.?
 
I'm most interested in the second graph, drawn on the first page of the thread. The most interesting thing about ticks is that they themselves are the key to the probability distributions associated with bars. And the amplitude of ticks seems to have little effect here: it is almost always about the same. What matters is the variation in the latency of their arrival and the small skewness in the frequency of ticks of different signs.
 
Mathemat has delved into research:) I'm probably addicted to multicurrency and multi-instrumental analysis:) Yeah, I guess everyone gets there in his own way:) Work on it, it's the right thing to do:) At least there's probably not a single indicator in such a scale, the existing multicurrency ones are based on simple methods, and one can go even further ... One can only guess:) Simple maths will not do, or rather simple algorithms. At least I always get very complicated ones, no matter how I approach them, they just can't get any simpler:). In my researches I can say that so far I haven't got what I wanted, and in fact I've just started, as one task has been cancelled, there is a tendency to develop this task:)
 
You canwork with ticks, for example, by setting such tick control blocks, as a tick tick passed in all currency pairs, no matter how many ticks passed, you need at least one tick simultaneously in all currency pairs, the beginning of this tick is the same as the end, at least one tick passed, this is the control point, let's say, a multi-currency tick. One tick has started to cancel the block and the last tick has finished it. The beginning tick may be the only tick for one currency pair; the last tick must be the only tick for another currency pair; during this time, from the beginning tick to the end tick, many ticks may pass for other currency pairs. It's exactly in this multicurrency tick that we should look for some side facts. Besides, it is possible to take as any other instrument and attach it to the multi-currency tick, for example, by considering it in the range of the multi-currency tick or by extending the range of the multi-currency tick.
 
Sharp increase in tick volume =News release.
 
Mathemat:
I'm more interested in the second graph, drawn on the first page of the thread. The interesting thing about ticks is that they themselves are the key to the probability distributions associated with bars. And the amplitude of ticks seems to have little effect here: it is almost always about the same. What matters is the variation in the latency of their arrival and the small skewness in the frequency of ticks of different signs.

Be so kind as not to quit halfway through. I've been meaning to dig into tics myself for a long time (I've had the feeling for a long time that's where the dog was burrowing), but illness got in the way.
 
The pictures from the first posts are clear and legitimate,
but it is really interesting to see with your own eyes what you assume and what seems obvious.
So it's an interesting topic, and worth digging into - it will be of benefit to everyone.
Although there will be no material way out of it, I'm sure.

About the multi-currency analysis.
IMHO(DrawDaun - in my humble opinion naturally ... :)),
there is little point in considering different pairs.

Almost always, at least in all cases when I tried to analyze it
cross was obtained from other pairs - for example EURJPY was equal to EURUSD * USDJPY
Differences of +-1 pips could be explained by time differences of quotes.

It is more correct to consider currencies rather than pairs.
That is, it is more correct to consider the values of currencies rather than the ratios of their values.
How to calculate the values of currencies and what it gives you can be found in the MIndex indicator
. It should be available somewhere in the libraries, I have posted it here.
IMHO, very interesting for multicurrency analysis ... :))
 
rebus:
Would you be so kind as not to leave it halfway through?
I'm not going to give it up: it's part of a wider project. It's just that this second timetable is a bit of a problem, and I don't have any real thoughts yet. You just need to wait a little - then the thoughts will come ...

P.S. They have. One only for now. I did so: on the second chart from the first page of the branch, to somehow smooth out the frantic differences in tick delays, I simply calculated their logarithms. Here's the pseudo-random delay logarithm process for a couple of weeks in April (1st and 2nd):



Both processes became more "homogeneous" compared to the delay times themselves. The logarithms of the lags are now numbers in the intervals from about 0 (lag = 1 second) to 7 (lag greater than 1000 seconds). In fact, in both graphs the "lags" are deeper and much more frequent and in both cases reach a natural minimum (1 second). Lags of 0 seconds cannot be displayed here, although they are not insignificant - on the order of 2-3% of the total number of ticks. On the other hand, if we could measure lags with an accuracy better than 1 second, we would never have lags exactly equal to zero. And, of course, the periodicity associated with the Asian sessions can be seen here as well, although on the right-hand chart (trend week) not so clearly.

I suspect that "quasi-stationarity" of the logarithm of the lag from time appeared here not by accident. It suggests an analogy with the antipersistence of the lag process itself, similar to that for the standard deviation of Returns and described by Peters.
 
Mathemat:
I'm more interested in the second graph, drawn on the first page of the thread. The interesting thing about ticks is that they themselves are the key to the probability distributions associated with bars. And the amplitude of ticks seems to have little effect here: it is almost always about the same. What matters is the variation in the latency of their arrival and the small skewness in the frequency of ticks of different signs.

The first figure at the beginning of the branch shows a typical noise exponent. Exactly the same
exponent will be obtained if you calculate, for example, the number of points which the rate will pass in
5 minutes and then plot a histogram N from the number of points. The second figure shows
volatility change during the week - its variability is not evident and its variation
are also of random nature.

Tick characteristics among other things strongly depend on the "iron" used by brokerage companies.
uses and its software therefore they are different for different brokerage companies. The volume characteristic
actually shows the number of ticks in a bar varies for different brokerage companies, but they correlate
between each other due to orientation to indicative quotes. Because of the "primary source" in the form of
indicative quotes, the volumes also correlate with the volatility level of the market. When
there are many buyers and sellers in the market, the price is often jumping around
The indicative quotes and the ticks move together with them. During the holidays and between the sessions the sellers-buyers
correspondingly, indicative quotes stand and ticks move seldom.

It is useless to analyze the noise, of course. Even if there is something
and you find something, you won't be able to use it. It's more profitable to look for long-run patterns.
12345678
New comment