From theory to practice - page 1119

New comment
 
Renat Akhtyamov:

He's broken a lot of people.

And they still come.

By the way, the most rose-coloured glasses are those who cannot write software and do not want to learn.

Because they cannot test their idea, they have no choice but to blindly believe in it.

Exactly, the clear formalisation of a strategy almost always changes the colour of the points!
 
Renat Akhtyamov:

He's broken a lot of people.

And they still come.

By the way, the most rose-coloured glasses are those who cannot write software and do not want to learn.

Because they cannot test their idea and have no choice but to blindly believe in it.

Yeah. You're right.
Especially, you need to test it for ten years at a time, and you cannot do it except with a program.
Even by programming robots, it only takes a few years to get the desired result. Having written more than a thousand robots of various implementations.


 
Martin Cheguevara:
Yeah You're right
Especially, you have to test ten years of history at a time, and it is impossible to do it apart from software.
Even by programming robots, it only takes a few years to get the desired result. Having written more than a thousand robots of various implementations.


If you write 1000 robots, then, according to probability theory, one of them will show a good result by chance.
 
multiplicator:
if you write 1000 robots, then according to probability theory, one of them will randomly produce a good result.

without the theory of probability, you can write 1 robot, one of 1024 copies of which will produce a series of 10 successful trades in a row :-)

PS/ maybe even one in 21 :-)

PPS/ and ter.ver - instances of no more than 15

 
Maxim Kuznetsov:


PS/ even perhaps one in 21 :-)


no. one in 1024

 
multiplicator:

not. one in 1024

1024 is guaranteed, even for a random walk, if the trades are opened in a single moment and the series should be obtained right away.

However, if the goal is to obtain the series within a foreseeable time period, and the trades are traded on EURUSD for example, this number considerably decreases.

The total profit is of course negative, but the series will be obtained and 1 copy will be profitable.

Not less than half of the stats are started that way.

 
multiplicator:
If you write 1000 robots, then, according to probability theory, one of them will show a good result by chance.
Controlled linearly dependent on input parameters, not even one out of 100,000 may show a stable profitable result.
 
Renat Akhtyamov:

By the way, the most rose-coloured glasses are those who do not know how to write software and do not want to learn it.

Because they cannot test their idea, they have no choice but to blindly believe in it.

The downside of manual trading is that "in order to test a strategy, you need to test it over a period of a year or three, and no one will test it manually for that long, they'll get sick of it. "
The disadvantage of automated trading is that you can't always write what you see with your eyes into a trading programme.

 
multiplicator:

The disadvantage of manual trading is that "in order to test a strategy, you need to test it over a period of a year or three, but nobody will test it manually for that long, they get bored. "
The disadvantage of automated trading is that it is not always possible to write what you see with your eyes into a trading programme.

What you see with your eyes can be an illusion, so you have to formalise your ideas.

You can also test it manually, but you may not have the time of your life.

 
apr73:

What you see with your eyes can be an illusion, so you have to formalise your ideas.

That's what I'm saying: a trader can't always transfer everything to machine code.

Maybe someone else could do it, but who would waste their idea?

apr73:

It is possible to test by hand, but life may not be enough.

If you spend 3 years on testing each system, it's really not enough of a life.

1...111211131114111511161117111811191120112111221123112411251126...1981
New comment