Machine learning in trading: theory, models, practice and algo-trading - page 1313

 
Aleksey Vyazmikin:

Even the models are almost ignoring the first half of 2014... Yes everything changes, and no one knows when it will happen next time.

2014? Why are you digging so deep? The market is changing fast, and last year's story is already almost nothing.
I have it all together - training and test, no more than 9 months. I don't see the point in doing more than that. Imho, even that is a bit much.
 
Yuriy Asaulenko:
2014? Why do you dig so deep? The market is changing rapidly, and last year's story is already almost nothing.
I have everything together - training and test, no more than 9 months. I don't see the point. Imho, even that is a bit much.

I'm looking for stable patterns, because I expect them to have a better chance of working long enough after they are identified. And my models show that such regularities exist.

 
Aleksey Vyazmikin:

I look for stable patterns, because I expect them to have a better chance of working long enough after they are identified. And my models show that such patterns exist.

Me too, but on short intervals. I have a suspicion, that the long ones,

patterns are unlikely to exist or are easily destroyed by external factors.

PS As an example, you found a pattern lasting 24 hours, the pattern is great!!!, entered into a trade on it. How many external and other events can happen in a day, such that from your pattern will not remain a trace? And in training, you won't learn anything about these broken patterns, only later, when you're already in the real game. At least because of the multitude of possible outcomes.

 
I think he has a lot to say:

No one was going anywhere.

All strategies that this "physicist" tried to build cannot work because of the non-stationarity of quotes, because at the moment when the flat changes into a trend, the spread between the quotes and the average will decrease, but the position will be unprofitable - the mathematical expectation of the process is not a constant. And some mythical indicators of these two states will not work, because

1. they do not and cannot exist;

2. If they existed, there would be no need to calculate any distributions - you could earn money with a simple handwheel.

They tried to explain it to him several times, but he's insane - in the end he took the same dead-end road that millions went before him and with the same result.

What's the point of trying to explain obvious things to a man who won't listen to them...

Look, Dimitry, in my opinion Alexander realized everything on his last deal on EUR/JPY, let's move on.The strategy of returning to the average is quite controversial, the price goes to the average, not the price goes to the average, the result is a loss (to be more precise - they are close to each other out of proportion), let's go backwards (if the mountain doesn't go to Mohammed then Mohammed goes to the mountain) - if the price meets the close (read: crossed), it will go to one of the channels, to the top or to the bottom - I don't know whether it will reach it, or not. As for the targets, let's ask Chegevara, he has something to say.

I myself in the trade I use so clumsy methods, that after reading here science, I'm even ashamed to voice them, sometimes reading it all, I remember Lavrov - "morons, fuck.
 
Aleksey Vyazmikin:

I look for stable patterns, because I expect them to have a better chance of working long enough after they are identified. And my models show that such patterns exist.

Yuriy Asaulenko:
Me too, but on short intervals. I suspect that they are long-term ones,

patterns are unlikely to exist or are easily destroyed by external factors.

PS As an example, you found a pattern lasting 24 hours, the pattern is great!!!, entered into a trade on it. How many external and other events can happen in a day, such that from your pattern will not remain a trace? And in training, you won't learn anything about these broken patterns, only later, when you're already in the real game. At least because of the many variants of events.

I wrote that there is a pattern and it has not changed since the first day of the hand.

This is very surprising, but it is true.

//Not about the tick volumes, of course.

There is no cycle, it is not about this either.

There is a pattern to the price movement to the balance of supply and demand, not to any average.
 
Renat Akhtyamov:

I wrote that there is a pattern, and it hasn't changed since the first day of the handicap.

This is very surprising, but it's true.

//Not about tick volumes, of course.

There's no cyclicality, that's not what it's about either.

Renat, we're talking about MYH, you said yourself it's a different song and it's the other way around).

 
Yuriy Asaulenko:

Renat, we are talking about MOEX, you yourself said that this is a different song and everything is the opposite in general).

MOEH - it's elementary there.

The balance is constantly maintained through clearing.

 
Renat Akhtyamov:

MOEX - it's elementary there.

The balance is constantly maintained through clearing.

Clearing? Gee, I wish I had known that before.

 
Yuriy Asaulenko:

Clearing? Gee, I wish I had known sooner.

They don't like it there. They just shave.
 
Renat Akhtyamov:

I wrote that there is a pattern, and it hasn't changed since the first day of the handicap.

This is very surprising, but it's true.

//definitely not talking about tick volumes.

There is no cycle, it's not about that either.

There is a pattern - the movement of the price to the balance of supply and demand, not to any average.

Apparently you don't understand what cyclicality means.