Machine learning in trading: theory, models, practice and algo-trading - page 3045

 

Installed, but it still doesn't work.

> abline(h = 0,col=2,lty=2)
Warning messages:
1: replacing previous import ‘lifecycle::last_warnings’ by ‘rlang::last_warnings’ when loading ‘tibble’ 
2: replacing previous import ‘lifecycle::last_warnings’ by ‘rlang::last_warnings’ when loading ‘pillar’ 

> abline(h = 0,col=2,lty=2)
Warning messages:
1: replacing previous import ‘lifecycle::last_warnings’ by ‘rlang::last_warnings’ when loading ‘tibble’ 
2: replacing previous import ‘lifecycle::last_warnings’ by ‘rlang::last_warnings’ when loading ‘pillar’ 
 
Aleksey Vyazmikin #:

It installed, but it still doesn't work.

No, I drew a picture


 
Forester #:
Can you elaborate? Or an example. What values, what FF?

individual for a strategy.

For example, roughly, we choose a FF - balance, make tests on different parts of history with different parameters. we measure the standard deviation of the FF value for each set of parameters on separate parts of history.

example, balance by months: 10, 1000, 200, 8000, 9000, -5000: bad

example, balance by month: 1000, 1100, 1500, 800, 900, 950: good

 
Aleksey Vyazmikin #:

No, I drew a picture


I've almost got all of them in the plus side - strange randomness :)

 
Aleksey Vyazmikin #:

No, I drew a picture

Clearly something is wrong, everything is in warning... update all packages or something, I don't know.

 
mytarmailS #:

Clearly something is wrong, everything is in warrnings... update all packages or something, huzzah

Is there a button for that?

 
Aleksey Vyazmikin #:

Is there a button for that?

I found the button, but it writes questions - is there any way to answer all of them in the affirmative with one button?

 
Aleksey Vyazmikin #:

I found the button, but it writes questions - is there any way to answer all of them in the affirmative with one button?

I think so.

 
mytarmailS #:

In simple words, if I evaluate one TC by statistical significance, it's good,

if I have 100 TS and I choose the best one according to the same criterion, it's bad?


I must have misunderstood something? It can't be right either?

You have understood everything correctly, compare two trading stats for clarity:

1) It is obtained simply as a result of trading one TS on one trading account.

2) Many trading accounts are opened. Simultaneously trades are opened on all of them - on half of them up and on half of them down. The losing accounts are closed, and the remaining accounts are repeated until there is only one account without losses, and the steith of which is taken for comparison.

The second state will almost always be better than the first one (by any indicator).

It is clear that this is an extremely bad variant, but the selection deception can be clearly seen here. More often this deception is covered up and rather becomes self-deception.

 
Aleksey Vyazmikin #:

I found the button, but it writes questions - is there any way to answer all of them in the affirmative with one button?

Autoclicker should be written
Reason: