Volumes, volatility and Hearst index - page 4
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
In this case no tick volumes are used. Only price data from lower timeframe is taken (PeriodData parameter).
All the same cyclicities are visible.
I did that yesterday. Only I didn't calibrate, but looked at what the indicator showed on a clean SB. The result was unexpected for me. The average value on M10, H1 and H4 is around 0.54. Now I wonder why ?
It's not about cyclicality after all. It is obvious in itself. It is about how to either take it into account or introduce a measure that does not depend on it. In both cases it means to exclude cyclicality from the readings. Then market conditions can be compared more or less reasonably.
I don't understand what you are looking for? What do you need a measure for that is independent of cyclicality? You can compare market conditions a year or two or more ago with the given indicator. It is not tied to an ethereal tick volume. And one can reasonably estimate volatility.
Or am I misunderstanding something, please explain.
Bernoulli process type, p - probability of continuation q - probability of reversal, p > q - trend, p < q - reversal, p = q - random walk. That is, the important thing is not to work with probabilities +1 and -1, but with probabilities of coincidence of the sign and its change.
This is a process that I have been unwinding for two years now. I got the distribution density for it in analytical form. Very much wanted to link it to Hirst. I even wrote about it in that forum. However everything ended there.
In general my results seem to me more interesting than Hirst's. Especially since they are not taken from the ceiling (like Hurst's formula), but were obtained quite rigorously. However, Hurst's figure would also be interesting. A single scalar is a very handy and simple thing to characterise market conditions.
So, can you help me get the formula for SB in analytical form ? The SB distribution is known. It is a function of binomial coefficients. You can use Strling's formula to simplify it. The only thing to do is to figure out the spread.
I don't understand what you are looking for? What do you need a measure for that does not depend on cyclicality?
For example, there is the Hurst index, which is not just a measure of volatility, but also an indicator of market conditions: random walk (H=0.5), a return process (H<0.5) or trend (H>0.5). It means that depending on one absolute value it is already possible to decide which strategy to apply. There is no need to compare anything with anything, which means there is no need to store or swap past data, solve the problem of scaling values, etc.
Well there you go, that confirms my view that it's not Hearst but a different metric.
That's not a fact. I still think it is a Hearst figure. However, the algorithm I suggested seems to have some flaws. Formally it turns out that I first calculated the averaged range for the hour of the week (for simplicity, suppose, that it is H1), and also calculated the average number of ticks for this hour, and then divided their logarithms. Maybe, this averaging is wrong. As you know, the logarithm of a sum is not equal to the sum of logarithms. :-)
That's why it's important to work out what the spread is.
Peters, as far as I remember, split the series into identical intervals, i.e. he dealt with the same number of ticks. And under those conditions he averaged the spread.
So, can you help me get a formula for SB in analytical form ?
Yurixx:
Peters, as I recall, split the series into identical intervals, i.e. dealing with the same number of ticks. And under those conditions averaged the spread.
For example, there is the Hearst index, which is not just a measure of volatility, but also an indicator of market conditions: random walk (H=0.5), return process (H<0.5) or trend (H>0.5). It means that depending on one absolute value it is already possible to decide which strategy to apply. You don't need to compare anything with anything, which means you don't need to store or swap past data, solve the problem of scaling values, etc.
So it says that it is trand. What to do? What trend, its characteristics? From where and how to look at it.
Why this indicator should not depend on the cyclicality, which is a property of the market?
This thread is not about the definition of the flats and the trends, and it started with some volatility analysis based on the tick volume.
A working method for measuring volatility, moving away from tick volume, has been suggested.
It could also be further improved by showing how the width of the channels varies with time. And compare this.
Let's say you find an indicator that allows you to tell whether it's a trand or a flat by a single number. So?
Well, it says it's a trend. What do we do?
That's not the point. If you have R = AN^h, then you have to have A = 1 to draw rays from the origin for Hurst. Otherwise it's not a Hurst.
Correct. I determined that for the series built by the described technology, the coefficient is 1. Since for SB 0.5 does not work, then either I defined incorrectly and it is not equal to 1 (then I need to figure out what error), or the computational algorithm is incorrect (then I need to figure out what error), or something else (then I need to retrain as a manager).
You see, even if I set the coefficient not equal to 1 and determine it in some way for the EUR in TF = H1, it does not mean that for the pound and the other TF it will be the same. And it is not interesting. It is like dealing with the scale for each pair separately. If so, we may work with volumes.
No, I haven't done this kind of thing for a long time, it's not the right shape, it'll take a long time to practice :). Try throwing it in here:).
Not funny.