From theory to practice - page 249

 
Novaja:

The management of randomness. This random random random world.

Can you recommend anything else to read?

He also has a book called "By chance" from '86, which is newer, I don't know what's different.

 
Maxim Dmitrievsky:

He also has a book from '86, which is newer, so I don't know what's different.

Thanks, Maxim, can you drop me a line?

 
Alexander_K2:

Rena, wait!!!!!! Grail give!!!!!!!!!! :))))))))))

Alexander, your graaal is a bluff, I'm tired of repeating it already.
 
Renat Akhtyamov:
Alexander, your graaal is a bluff, I'm tired of repeating it already.

Post yours in the form of a signal + a physical and mathematical justification. I would be happy to subscribe.

For my part, I am tired of saying that I personally have no confidence in any signal, no matter how profitable it may be, if there is no physical sense in the strategy. I will never subscribe to just a pretty picture. Am I wrong?

 
Novaja:

Thank you, Maxim, can you send me a message?

sent to

 
Maxim Dmitrievsky:

sent to

Thank you very much!))) Can't thank you in private, no messages are being sent.

 
Alexander_K2:

Post yours in the form of a signal + a physical and mathematical justification. I would be happy to subscribe.

For my part, I am tired of saying that I personally have no confidence in any signal, no matter how profitable it may be, if there is no physical sense in the strategy. I will never subscribe to just a pretty picture. Am I wrong?

Are you kidding?

There's money in here!

Statistics and other heuristic guessing are not appropriate

Just finance and credit, economics a bit, politics!

And understanding that every cent counts for a brokerage company and it won't give it away so easily, it may trap you, but only to rob your pockets of money.
 
Novaja:

Thank you very much!))) I can't thank you in private, no messages are sent.

I only friendzone can write me, if you add as a friend you can

Another book I was recommended - Sarah Boslough "Statistics for All", they say nice, I have not read yet

 
Alexander_K2:

In fact, we cannot fully convert a non-Markovian process into a Markovian one, no matter how hard we try. But in the exponential time-price scale, most entry points behind support/resistance lines mean a return to the mean.

However, in my case 20% of 100% of crossings of these lines mean, well, let's say, the middle of the trend which is exactly the sign that the process memory cannot be completely "destroyed".

I need an additional parameter to improve the result.

As I said before, I'm using asymmetry coefficient now. But... Not the right one! Not quite right!

Absolutely sure that this parameter is the non-entropy coefficient https://en.wikipedia.org/wiki/Negentropy, i.e. a measure of deviation of a non-random process from a random one.

But I have no time to do it - the housemates are shaking like a pear, demanding that I stop research and immediately replenish the cash bags :)))

I may be writing nonsense now, but I can't help but pour out here what occurred to me during my thoughtful reading of this post of yours.

1. What is a non-Markovian process by definition? It is a process whose evolution depends on its past. As far as I understood your research showed that price series is a NOT Markovian process. It seems to me at this point all the technicians should cry hallelujah! All of them from Charles Dow and the ancient Japanese, to Larry Williams and Herczyk! Because you've given technical analysis a chance, and importantly, you've given it a chance in terms of mathematics!!! Previously they had based it all on postulates: "history repeats itself", tra-la-la, "trends", back and forth. But now it follows from your research that there is some dependence of the market in the future on the market in the past. So it can be attempted to be predicted. So forecasting can make sense at all. That alone is sufficient material for an article, if only you can more or less rigorously prove that the price series of today's markets is aNOT Markovian process.

Well yes I understand, there won't be such an article, because all you're thinking about now is "five golds on the field of miracles"... :)

2. How on earth are you trying to make your billions by making such a discovery? Paradoxical!!! You come up with tricks to turn a non-Markovian process into a Markovian one! That is, not dependent on the past! But to which "your mathematics" is applicable - exactly the mathematical apparatus you know well. Why not? Yes, you can!

Here's what you write: "In the exponential time-price scale, most entry points behind support/resistance lines mean a return to the mean. However, in my 20% of 100%, crossing these lines means, well, let's say mid-trend, which is exactly the sign that the process memory cannot be completely 'destroyed'."

In trading parlance, if I understand you correctly, it translates as follows: you are trying to make a counter-trend trading strategy and according to your tests so far 80% of trades are profitable! That's awesome! Really! In 20% of the trades it gets carried away on an unexpected trend - this is the expected behaviour of counter-trend strategies! If this is true, then the approach looks logical! After all, if you reduced everything to a Markov process, then the link to the past is gone! And that means you can no longer predict movements! So how can there be prediction in a random process? But you just can calculate characteristics of a given random process - first of all mathematical expectation, and possibly the form of distribution, from which you can draw interesting levels - that's valuable.

3. I used to read somewhere in this voluminous thread "let's make money on fat distribution tails", and to be honest at the time I understood it as a "black swan" variant. Now I understand that fat tails for your strategy means just enough trades - if they were thin, the trades would be few. And the target is always a mathematical expectation. Have I understood your idea correctly?

By the way, more than once I've read studies showing that distributions of market price series do have thick tails - certainly much thicker than in a normal distribution. Also, the long-established, plausible opinion is that the currency market is flat 70% of the time and trends 30% of the time. I wish someone would illustrate this statement statistically, with calculations...

4 I do not understand why you think that "the crossing of these lines means, well, say, the middle of a trend, which is just a sign that the process memory cannot be completely "destroyed"."

Why is it that if a trend has started, it is from" processmemory "? Let's think about it, where does the trend lead? Either to the tail of the distribution, farther back than the point where you entered the return to the mean - it's not really a trend, just an anomalous deviation. Or the trend leads to a complete move of the average together with the bell of the distribution to another location by price - this is really a trend and this is the natural behavior of the market. Do you think that if some kind of transformation rips up the "memory" of the process, trends will disappear?????

 
Maxim Dmitrievsky:

Only my friendzone can write to me, if you add me as a friend you can.

Another book I was recommended - Sarah Boslough "Statistics for All", they say nice, myself have not read

Go ahead, Maxim, I have an appetite!)))) I added you))))