Measurement of vibration amplitude - page 8
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
For example, the first range is 10-13 pips, which equals 10+30%. I call it the range with 30% deviation. The maximum percentage (on the chart) in the range 42-54.6 points, it means that out of all single fluctuations (say there are 100) in the range 42-54.6 points, fell 26 pieces, or 26%. It means that there is 26% probability that the price having passed 42-54.6 points will reverse and pass the same amount of points in the opposite direction. Naturally, the wider the range, the more probable it is that a single fluctuation will fall into it.
In a short history (a month), we can see minimums and maximums; if we take the history of 3 years, it becomes almost flat, with a fall in the beginning. Thus, the longer is the history, the more even distribution becomes. It shows how the market changes, and amplitudes distribution differs in every separate time period, so TS optimized for one period will fail in the forefront. Therefore, knowing the distribution of amplitudes, we can adjust parameters of TS, like optimizing in real time.
But this is only for 30% of deviations. This periodicity is right, but moreover, other deviations will also have their own periodicity, but they may overlap each other, resulting in a common dominant how to determine?
Now the graph shows where the main fluctuations occur (dominant) and with what probability, but if you increase the width of the range, the probability of hitting it will increase. For example I have calculated that if you increase the range to 200%, then 70% of all fluctuations fall into the range of 24-72 pips (for the same period). That means that there is a 70% probability that the price will not go beyond 72 points without a pullback but at least 24 points will pass without a pullback. This can be used to build a probabilistic TS.
In the interval, where there won't be a distinct pattern, just don't trade, at that time you can choose (even automatically) another pair with a more distinct pattern.
This I built in Excel, therefore it will take a long time to analyze by history, but if you write an indicator, then by history (when running in the tester) you can examine dependencies in details, perhaps there will be periods with the probability of getting into some narrow range (about 20-40%) that will be even higher than 50%.
Besides, the average number of bars in each range collet is not considered here (I don't know how to implement it in Excel), and it can seriously enlarge the carat.
This should be similar to the calculation of preemptions through the inertia of volatility changes in cluster analysis, only not so directly,
Now the graph shows where the main fluctuations occur (dominant) and with what probability, but if you increase the width of the range, the probability of hitting it will increase. For example I have calculated that if you increase the range to 200%, then 70% of all fluctuations fall into the range of 24-72 pips (for the same period). That means that there is a 70% probability that the price will not go beyond 72 points without a pullback but at least 24 points will pass without a pullback. This can be used to build a probabilistic TS.
In the interval, where there won't be a distinct pattern, just don't trade, at that time you can choose (even automatically) another pair with a more distinct pattern.
This I built in Excel, therefore it will take a long time to analyze by history, but if you write an indicator, then by history (when running in the tester) you can examine dependencies in details, perhaps there will be periods with the probability of getting into some narrow range (about 20-40%) that will be even higher than 50%.
Besides, the average number of bars in each range collet is not considered here (I don't know how to implement it in Excel), and it can seriously enlarge the carat.
Do you have any calculations in excel on this principle, to see what it looks like in an example?
There's something wrong there, it seems to me. Or I don't understand what you're investigating there.
But in general, the frequency of the swing should look something like this:
The shortest appear more often, the longest less often. And the longest ones never.
There is something confusing there, it seems to me. Or I don't understand what exactly you are investigating there.
But in general, the frequency of the sweep should look something like this:
You have probability on the graph, not frequency.
The frequency response can be anything.
You have probability on the graph, not frequency.
The AFC can be anything.
It is not important in this case. This is not a matter of terminology - you can consider probability=frequency=frequency.
As for the AFR, that's not what we're talking about here. As far as I understood, zigzag is being tortured here. And the zigzag has exactly that, for random rambling. It's the same on Finish instruments (many of them), by the way.
ZS, frequency, in the sense of the frequency with which a zigzag knee of a certain size appears.
There's something wrong there, it seems to me. Or I don't understand exactly what you're investigating there.
But in general, the frequency of the swing should look something like this:
The shortest appear more often, the longest less often. And the longest ones never.
Why does it have to look like that? I don't see the logic in it. If it were, it wouldn't be a problem to pips at all. You just wait for a 5 point movement up and then at the peak you go back down. Based on your chart, the possibility is almost 100% that you will make profit on a 5-point pullback. In reality it is quite the opposite, there is a definite distribution at each new intrevale and it is different from the previous one.
Why does it have to look like that? I don't see the logic in it. If it did, it wouldn't be a problem to pips at all. You just wait for a 5 point movement up and then at the peak you go back down. Based on your chart, the possibility is almost 100% that you will make profit on a 5-point pullback. In reality it is the opposite, there is a well-defined distribution on every new intraval and it is different from the previous one.