Where is the line between fitting and actual patterns? - page 8

New comment
 
Reshetov:
You can make unsupported assumptions, i.e., nonsense, all your life. Unlike nerds, traders need to make money, not make assumptions. And to make sure that the TS is adequate, it is necessary and sufficient to run the forwards on it.


Did you run enough forwards? - If the answer to both questions is yes, then you are a nerd, and I am a trader ;)

I cannot make an algorithm without a mathematical model that does not require optimization/fitting/martin - that's why I asked in the thread

 

IgorM:


Have you run enough forwards?

My problems. I hate those who like to count money in other people's pockets.

IgorM:


I cannot make an algorithm without a mathematical model that does not require optimisation/fitting/martin - that's why I asked in the thread

Your problems.

 
Reshetov:
Thanks for the constructivism in the topic :D, as always, some kind of doom hangs over this forum - as my acquaintance said: they've been searching there for years and can't find it )))
 
IgorM:
Thanks for the constructivism in the thread :D, as usual it's all just banter and a certain doom hangs over this forum.

No doom, apart from some buridan donkey problems.

Go forwards and get answers to the questions posed by the topicstater without all the flubbing. If you don't want to do that and you keep on blathering on without proof, well, as they say in the classics:

Don't blame the mirror when you're wrong (from Kozma Prutkov)

IgorM:
as my acquaintance said: they have been searching there for years and cannot find it ))

It's hard to look for a black cat in a dark room, especially if it is not there (c) Confucius

 
IgorM:

It is easier to look for patterns if the speed of price movement is taken into account, but there are several problems:

- what to take as a reference point (zero/start);

- what to take as a unit of time (second/minute/tick/bar);

if there are methods of calculation of speed and correlation of indicator parameters depending on speed, I would like to know more

So what does speed have to do with anything? In fact, speed has properties that make it fundamentally irrelevant, meaning that at any given moment in time, speed is in the past and has no meaning at all for the present, a hundred percent fit at a price/time combination.
 
Reshetov:

Н


It's hard to look for a black cat in a dark room, especially if it's not there (c) Confucius


It is even harder to call it a cat when you step on it (c)

You can build a system on the speed of price movement.

 
paukas:

It's even harder to call it a cat when you step on it (c)

You can build a system on the speed of price movement.

Going back to print:

Gerasimm:
So what does speed have to do with it anyway?
Let me tell you a secret: TC can be built on anything, even a random number generator. That's not what this is about.
 
Reshetov:

Going back to what was printed:

Let me tell you a secret: TC can be built on anything, even a random number generator. That's not what this is about.
Build it. Am I disturbing you?
 
Reshetov:

Going back to what was printed:

Let me tell you a secret: TC can be built on anything, even a random number generator. That's not what we're talking about.
Come on :o)...Everyone talks in parallel anyway.And a random number generator will be way ahead of systematised approaches in many cases...
 
joo: An idea suggests itself - to select the OOS size as small as possible, thus increasing the relevance of optimization. But there is a problem here as well - the smaller OOS, the greater the probability that the TS will fail in real trading.

It's a double-edged sword, so to speak. Or even three ends.

The issue here is not the size of OOS, but the choice of such a size of an optimization and OOS periods that the patterns that we will find at these periods will continue in the future. What are your thoughts on this?
123456789101112131415...64
New comment
Reason: