Time machine. Your actions. - page 13

 
Mischek2:

_
Basically an interesting task, there's a lot to come up with.

This is only part of the problem. When solving a complex problem, the latter is broken down into parts and each part is solved separately.
 
Mischek2:


Well, it's about finding the single optimal algorithm

For example, if you enter today at a deviation of so many points from the price on August 1, you should close, even if the price reaches the level of August 1, for example, tomorrow, with subsequent opening again, of course, when the same deviation is reached. Basically, it is an interesting task, and we can come up with a lot of other ideas.

What if the price reaches the level of August 1, will it go higher tomorrow? And will it continue to grow? Only on July 31 will it start to correct to the known price
 
In December 2002, FBI agents arrested a 44-year-old man in New York on suspicion of fraud.
He was allegedly using insider information to play the stock market. That is, by colluding with the managers of companies trading in stocks, he received commercial information from them. Which was the reason he was so successful financially.
 
As I recall, criminal conspiracy was not proven and the dude disappeared
 
alexx_v:
What if the price reached the level of Aug 1 will go higher tomorrow?


The problem implies free-riding. It means that the player refuses from forecasting as a process and uses only the deviation.

In this case we should close at every price touch on Aug 1 on the days before Aug 1.

 
alexx_v:
as I recall, criminal conspiracy was never proven, and the dude's missing.

yep, back in 2036...

that's about it:

 
Mischek2:


The problem is a freeloading one. That is, the player refuses to forecast as a process and only uses deviation.

In this case, you have to close at every price touch on Aug 1 in the days up to Aug 1.

the task implies efficiency, and to close every time the price drops on Aug. 1 before Aug. 1 is inefficient.
 
moskitman:

yep, back in 2036...

that's about it:

I don't know, maybe the guy's still in the loony bin proving he's from the future and they're treating him.
 
Mischek2:


Well, it's about finding the single best algorithm

For example, if you enter today at a deviation of so many points from the price on August 1, you should close, even if the price reaches the level of August 1, for example, tomorrow, with subsequent opening again, of course, when the same deviation is reached. Basically, it is an interesting task, and there is a lot to think of.


If we take the wave model as the one based on the random walk, then the price spread around the forecast price will decrease in direct proportion to the root of the remaining time. You can open, for example, at X sigma deviation from the forecast price. And calculate for each moment the equity share and how X changes. But still there is no optimality purely in terms of income - only some functional of income and risk.

A more competent volatility model than the SB model will give a better result. I.e. it all depends on the volatility model and what functionality we want to maximize. There are risk-free options, but not with maximum returns.

 
alexx_v:
I don't know, maybe the guy's still in the asylum proving he's from the future and they're treating him.
I don't think so. Knowing his success in the market and his short life he should have been listened to, imho.
Although you may listen to him in an asylum to prove his point more forcefully... ;)