From theory to practice - page 374
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
So he may as well be Shredeng's cat next time....
It's not him, probably another take on him. Now he has also changed his gender.
No, on the contrary, this message https://www.mql5.com/ru/forum/86386/page945#comment_7539752 reveals access to the previously hidden signal. As Alexander said, it is registered in his wife's name, now it is not hard to find the signal itself. Assuming that Olga Shelemey is her forum name, we will find the only signal https://www.mql5.com/ru/signals/412532, which reads :
"Author of the trading system - Alexander_K2(https://www.mql5.com/ru/forum/221552)".
Yes, the outcome was not promising, a loss of 328 USD on the real, including about zero for the two months of 2018. However, this does not mean that the ideas behind the proven system are flawed, even if not all of them are known to us. I think if Alexander had not hidden the signal so carefully, but had discussed the results, something positive would still have come out in practice. After all, he would attract attention, and many would give their opinion about the reasons and results of each transaction. Why not bring out the grain. But... he closed his trading to the public early on. And it would have been possible to get, for example, a comparison with Bollinger Bands. And their modifications. И ... how many posts there were with the results of various analyses, which seemed to be close to Alexander's ideas. But it's done. I, for example, no longer have interest in the results of trading using Alexander's (system ?). Probably, it will appear again, if he begins to consider other people's words and stops hiding his failures. They are inevitable, but for him they are painful because of his stated goal - to show that for physicists the task is easy. Had the goal been simpler, he might have moved on.
At the same time, I would like to express my gratitude to Alexander, who has encouraged me to learn so much that I did not know anything about before. In fact, I even went to find out what people are doing now in probability theory and matstatistics, looked through the composition of courses and downloaded many sources studied in recent years at Moscow State University and MIPT. Most valuable for me from this thread I think is the brilliant reminder that studying distributions is very far from studying sequences, or rather that different sequences of events can have the same distributions. And it is the sequence that matters to us who want to profit from rate movements.
No, on the contrary, this message https://www.mql5.com/ru/forum/86386/page945#comment_7539752 reveals access to the previously hidden signal. As Alexander said, it is registered in his wife's name, now it is not hard to find the signal itself. Assuming that Olga Shelemey is her forum name, we will find the only signal https://www.mql5.com/ru/signals/412532, which reads :
"Author of the trading system - Alexander_K2(https://www.mql5.com/ru/forum/221552)".
Yes, the outcome was not promising, a loss of 328 USD on the real, including about zero for the two months of 2018. However, this does not mean that the ideas behind the proven system are flawed, even if not all of them are known to us. I think if Alexander had not hidden the signal so carefully, but had discussed the results, something positive would still have come out in practice. After all, he would attract attention, and many would give their opinion about the reasons and results of each transaction. Why not bring out the grain. But... he closed his trading to the public early on. And it would have been possible to get, for example, a comparison with Bollinger Bands. And their modifications. И ... how many posts there were with the results of various analyses, which seemed to be close to Alexander's ideas. But it's done. I, for example, no longer have interest in the results of trading using Alexander's (system ?). Probably, it will appear again, if he begins to consider other people's words and stops hiding his failures. They are inevitable, but for him they are painful because of his stated goal - to show that for physicists the task is easy. Had the goal been simpler, he might have moved on.
At the same time, I would like to thank Alexander for encouraging me to learn so much that I knew nothing about before. In fact, I even went to find out what people are doing now in probability theory and matstatistics, looked through the composition of courses and downloaded many sources studied in recent years at Moscow State University and MIPT. Most valuable for me from this thread I think is the brilliant reminder that studying distributions is very far from studying sequences, or rather that different sequences of events can have the same distributions. And it is the sequence that matters to us who want to profit from rate movements.
Do you know any more promising directions?
Let's leave aside the word "you know", it is addressed to those who have found the grail and actually means "share, eh? I merely "consider this and that".
As I said, it is more promising in my view to analyse sequences rather than distributions of course values or their increments. By definition, a distribution is a property of the general population and is integral in nature. Lack of memory and the square root law are also integral. What is needed is a model of course behaviour on a local area. Where, in particular, both of these integral properties can be violated.
P.S. Yes, sorry. The model itself is not required at all, any of its properties, regularities will suffice. There are no words, it is more fun with a model, but it is still optional. You can see it in the sequence of Alexander's search - he started with absolute belief in t2-distribution ofStudent's t3, which means a statistically significant relationship between two values of the series, the current and the next, then he began to look for "some index instead of drift ratio", "at least something convergent to the limit of Paulson" and others. He moved on to looking for at least some trade-appropriate properties, with a willingness to change both the distribution and its type; the model became unnecessary.Uncle A_K2!
I've thought of a trick about long tails of increment distribution in order to tweak them,
read up on the Nyquist frequency, see if you can't figure it out....
Uncle A_K2!
I've been thinking about long tails of incremental distributions to tweak them,
read up on the Nyquist frequency, see if you can't figure it out....
It won't work. You crossed a hedgehog with a hedgehog and think you'll get 30 metres of barbed wire.
There are no frequencies there, it's all leopards in the bush, silhouettes of which some, especially frightened citizens, see in the divisions of the mountain.
As I said, I think it is more promising to analyse sequences rather than distributions of course values or their increments.
It is more promising to analyse both sequences and their distributions at the same time. A distribution is one of the properties of a sequence, however.
It is more promising to analyse both sequences and their distributions simultaneously. The distribution is one of the properties of a sequence, however.
In my opinion, it is promising to study different properties of graphs. in econometrics we study many of them.
In my opinion, it is promising to study the different properties of graphs. in econometrics, there are many such studies.
The apparatus of BP processing is the same everywhere, in econometrics it is the same. But the purpose of the processing may be different. I do not think that econometrics can help in matters of short-term trading up to 1-2 months. I don't think econometrics can do anything for short-term trading, say, up to 1-2 months.
It is more promising to analyse both sequences and their distributions simultaneously. Distribution is one of the properties of a sequence, however.
Sometimes yes, but rarely. We are talking about a probability distribution, but who has shown that there is statistical stability in a sequence of courses? I remember that you yourself at least twice cited the figure https://www.mql5.com/ru/forum/221552/page162#comment_6399653 ("I.e., a distribution is formed, then a "jump" of price occurs outside this distribution, and a new distribution is formed around another centre."), where you can see how the statistical characteristics of the series change. They don't have that stability. In particular, there is usually no distribution either. It is another matter to find local properties of a sequence that will make sense at some point, and use probability theory methods for the parameters in those properties at that point.
By the way, after reading bas's remark about difference between distributions and sequences, I got interested and calculated one ratio. I took data from Alexander(https://www.mql5.com/ru/forum/221552/page349#comment_7348507) and calculated sums of positive and negative increments considered separately for the purpose of trading on the movement of the rate in the range of 0.01464 (from 1.20982 to 1.22446). The properties of the distribution are not related to the order of the increments. In other words, the mean, variance and in general all moments will be the same if first all the positive rate increases and then all the negative ones happen. In this case the rate will rise by 0.8 first and then fall by almost 0.8. By methods of distribution analysis we will find fluctuations of interest with the amplitude of 0.015 against the background of fluctuations with the amplitude of 0.8 that are 55 times larger. I will not name this ratio, but I am impressed with it.