From theory to practice - page 258
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
My God, what is this?
Gentlemen, look at the bizarre distribution of trading intensity with an exponential reading time in the sliding exponential window = 10.000 (about 4.5 hours)
This was nowhere near to being the case when the quotes were read evenly.
I have no idea what it is - let it remain a mystery.
By the way, it reminds one of the brightness histograms of a photograph in which there are several predominant colours (grayscale).
Now it's time to paint a picture of the market literally. :)
Read, read... After all, I know Doc is a genius, he doesn't write anything for nothing. Then I reread it again.
I.e. knowing that we have p=0.7 and using discrete LF generatorshttps://habrahabr.ru/post/265321/ we should set exponential time intervals not randomly, as I now TimeInterval = INT(- Ln(U))+1, where U is uniform CB from the range [0;1], but exactly TimeInterval = INT(- Ln(degree(0.3;U)) +1?
It seems that in this particular case, we are going to destroy "memory" almost completely... If so, Doc gets a Nobel Prize!!!!!!!!!!!!Since morning I've tried to pick up parameters for exponential distribution, as on last pages have advised, nothing worked, and even have picked up other constant instead of e, and picked up on what to multiply result, in general it was somehow complicated, and the curve built still badly coincided with yours.
And then I guessed to subtract one, and distribution parameters immediately found (though worse than in your excel, but through the exponent). I then deleted the first message with the curve formula, and it was still cumbersome and worse.
Conclusion - for an exponential distribution, it is better to start with zero. I haven't experimented with discrete distributions, I don't know.
If you saw something more in my words - I was glad to help, but I did not mean anything like that :)
Since morning I tried to pick up parameters for exponential distribution, as I was advised on previous pages, nothing worked, I even picked up another constant instead of e and chose how to multiply the result, in general it was somehow complicated and the curve obtained did not coincide with yours anyway.
And then I guessed to subtract one, and distribution parameters immediately found (though worse than in your excel, but through the exponent). I then deleted the first message with the curve formula, and it was still cumbersome and worse.
Conclusion - for exponential distribution it is better to start from zero. I haven't experimented with discrete distributions, I don't know.
If you saw something more in my words - I was glad to help, but I didn't mean anything like that :)
Thank you very much. Still, our tails are exponential to a greater extent( after 15 ) than logarithmic. Still not enough data 200,000 to study the distribution of tails.
Alexander, I have a request, can you make the data bigger, at least 2-3 times bigger? Is it technically possible?
Sure. On Saturday there will be about 300,000 ticks with time stamps.
Thank you very much. Still, our tails are exponential to a greater extent ( after 15 ) than logarithmic. Still not enough data 200.000 to study the distribution of tails.
And why do we need to know the tails of the time interval distribution? Do you want to find an exact formula? That is, if it turns out to be the product of some function by an exponent, then work exactly on the frequency of the exponent?
Sure. There will be about 300,000 ticks with time stamps on Saturday.
Thank you very much, you are a true gentleman!))
Thank you very much, you are a true gentleman!)))
I would do anything to turn my 80% success rate into 100% (in fact, to find the living Grail. I can already see his ears ;))))
Even with complex arctangens I'm willing to bend the time series:))))My God, what is this?
Gentlemen, look at the bizarre distribution of trading intensity with an exponential reading time in the sliding exponential window = 10.000 (approximately 4.5 hours)
This was nowhere near to being the case when the quotes were read evenly.
I have no idea what it is - let it remain a mystery.
And this is something that goes against
We need practical conclusions from what we have seen. If this bimodal distribution of trading intensity says to trade in a sliding window = 12 hours, then I will do so immediately.