Discussion of article "Brute force approach to pattern search" - page 3
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
the daily cycle at the majors draws an almost stable curve. Its shape is constant and all extremums and bends occur at the same moments of time. On majors you can (and should) trade orientated by the alarm clock :-)
you must have overdone some research, it happens.
Everything seemed to be simple and transparent. I used a zigzag because it does not change with time change and therefore ignores volatility fluctuations. Two zigzags were taken and the frequency of the fact that the tops of the smaller zigzag can become the tops of the larger one are actually them was calculated.
There is some dependence of this frequency on the time of day, but it is weak and floating with time.
Everything seemed to be simple and transparent. I used a zigzag because it does not change with time change and therefore ignores volatility fluctuations. I took two zigzags and considered the frequency that the tops of the smaller zigzag could become the tops of the larger one.
There is some dependence of this frequency on the time of day, but it is weak and floating with time.
but that's because it's a zigzag. Like any instrument, it has an AFC of sensitivity. Do you know how to determine the natural frequency of a zigzag from its parameters? It's a challenge, so it's best not to touch it. That's where it makes irreparable errors.
In simpler words: when analysed, the zigzag produces X,Y vertices with a significant error in both axes. The error depends on the absolute value of the price and its volatility.
but because it's a zigzag. Like any instrument, it has an AFC of sensitivity. Do you know how to determine the natural frequency of the zigzag from its parameters? It is a challenge, so it is better not to touch it. That's when it makes irreparable mistakes.
In simpler words: when analysed, the zigzag produces X,Y vertices with a significant error in both axes. The error depends on the absolute value of the price and its volatility.
Frequency does not mean "frequency of fluctuations", but "frequency of an event". There are two zigzags on the same asset - one with a small threshold and one with a large threshold. Some tops (not all) of the smaller zigzag could potentially become tops of the larger one, but not all possible tops became so, of course.
We can also say about "movements" that form the larger zigzag and "corrections" belonging to the smaller zigzag and potentially being the beginning of a "movement". The frequency we are interested in is the ratio of the number of "corrections" that became "movements" to the total number of "corrections". This frequency was studied as a function of the time of "corrections". This dependence seemed weak and floating.
Frequency not in the sense of "frequency of oscillation", but in the sense of "frequency of event". There are two zigzags on the same asset - with a small threshold and with a large one. Some tops (not all) of the smaller zigzag could potentially become tops of the larger one, but they became so, of course, not all possible ones.
We can also say about "movements" that form the larger zigzag and "corrections" belonging to the smaller zigzag and potentially being the beginning of a "movement". The frequency we are interested in is the ratio of the number of "corrections" that became "movements" to the total number of "corrections". This frequency was studied as a function of the time of "corrections". This dependence seemed weak and floating.
Dependencies can be found anywhere, and most likely we will only think we have found one. A zigzag is generally just one candle bulging relative to the neighbouring two. It is stronger when two more candles on the sides are taken into account and so on. To analyse the market correctly, you need to know the physics of the pricing process, and these zigzags are in most cases just a visible order in the endless chaos of pricing. You can spend your whole life digging in these indicators and looking for dependencies, but in the end you only waste your time. My opinion is that 90 per cent of indicators can be thrown into the furnace and let the zigzag be the first to fly).
Frequency not in the sense of "frequency of oscillation", but in the sense of "frequency of event". There are two zigzags on the same asset - with a small threshold and with a large one. Some tops (not all) of the smaller zigzag could potentially become tops of the larger one, but they became so, of course, not all possible ones.
We can also say about "movements" that form the larger zigzag and "corrections" belonging to the smaller zigzag and potentially being the beginning of a "movement". The frequency we are interested in is the ratio of the number of "corrections" that became "movements" to the total number of "corrections". This frequency was studied as a function of the time of "corrections". This dependence seemed weak and floating.
one can see a well-skilled theorist (in a good sense) :-)
when processing experimental data, the peculiarities of measuring instruments should be taken into account.
The zigzag is not quite what you expect it to measure. Like any other instrument it has measured signal+noise. And it triggers (shows a kink) not at the exact moment, but when the measured value exceeds the threshold value.
That is, for example, the market has already reoriented to short, and the high surges are just volatility, a high on this marvellous event. But the zigzag will persistently show the shoulder up, and the top will come out later than the trend change. And this "later" is quite significant and has a probabilistic nature, depends on volatility.
Therefore, when you took the exact values of the zigzag tops, you did not catch the seasonality.
Dependencies can be found anywhere, and most likely we will only think we have found them. A zigzag is just one candle rising relative to the neighbouring two. It is stronger when there are two more candles on the sides and so on. To analyse the market correctly, you need to know the physics of the pricing process, and these zigzags are in most cases just a visible order in the endless chaos of pricing. You can spend your whole life digging in these indicators and looking for dependencies, but in the end you only waste your time. My opinion is that 90 per cent of indicators can be thrown into the furnace and let the zigzag be the first to fly).
I think it is called a fractal in our trading science).
Anyway, I mentioned the zigzag there only to explain my idea.
You can see a well-skilled theorist (in a good way) :-)
When processing experimental data, the peculiarities of measuring instruments should be taken into account.
The zigzag is not quite what you expect it to measure. Like any other instrument it has measured signal+noise. And it triggers (shows a kink) not at the exact moment, but when the measured value exceeded the threshold value.
That is, for example, the market has already reoriented to short, and the high surges are just volatility, a high on this marvellous event. But the zigzag will persistently show the shoulder up, and the top will come out later than the trend change. And this "later" is quite significant and has a probabilistic nature, depending on the volatility.
Therefore, when you took the exact values of the zigzag tops, you did not catch the seasonality.
I guess so, my zen is not strong enough to determine whether a tick is real or a HYIP immediately at the moment of its arrival) I guess it's time to go back to Shaolin).
I guess so, my zen is not strong enough to determine immediately at the moment of tick arrival whether it is real or hype) I guess it's time to go back to Shaolin).
Welcome to 少林
Strengthen your zen with the cycles of waves in the rainy season :-)
These are diurnal seasonals (the ones that supposedly don't exist), but which are known in advance to say the least.
It's not that you can draw them 24 hours in advance like in the screenshot, you can draw them a week/month in advance
Welcome to 少林.
Strengthen your zen with cycles of waves in the rainy season :-)
These are diurnal seasonals (the ones that supposedly don't exist), but which are known in advance to say the least.
They are not just a day in advance as on the screenshot, they can be drawn a week/month in advance
It seems to be missing the character "temple" 寺 at the end).
No argument, the picture is beautiful, but the problem is that such pictures are sometimes interspersed with ugly ones.
Nevertheless, thanks - I will meditate on the picture, maybe it will help in my search for enlightenment).
such pictures are sometimes interspersed with ugly ones.
A small study shows that ugly pictures are quite enough. For this purpose it is enough to look at the histogram constructed by sampling the time of day for the tops of the zigzag. The pretty pictures should give troughs with high banks and bottoms at zero on this histogram.
The histogram below is plotted for EURUSD for the year 2020 for a zigzag with a 0.1% knee. There are troughs, but their bottoms are well above zero, suggesting quite a few ugly pictures when reversals happen "off the clock". If such bad days are also badly mixed with good days (which is quite possible), the result can be quite sad.