Machine learning in trading: theory, models, practice and algo-trading - page 3252

[Maxim Dmitrievsky]

Flat pair, sentinels

pattern found in 15 seconds :) From 2020, everything before is OOS

slightly optimised the parameters from 2019

upd

The same pattern works on another character as well

It is faster than MO

[Valeriy Yastremskiy]

Maxim Dmitrievsky #:

Flat pair, sentinels

pattern found in 15 seconds :) From 2020, everything before is OOS

slightly optimised the parameters from 2019

upd

The same pattern works on another character as well

It is faster than MO.

But you yourself realise it is a choice among SBs)))) but it lasts much longer on the history, and we can assume that it will not end immediately. But that's not a fact. The task is of course much more difficult, to assess the state of these repetitions/well, like patterns, oh, pseudo-laws)))) The pattern as it is, in almost SB, is of course already something, even to find it, you need to make an effort, but for forecasting it is nothing. Unfortunately... or I don't know, but that's the way it seems to be))

The beginning of the pattern and the end of the pattern are interesting.

[Maxim Dmitrievsky]

Valeriy Yastremskiy #:

But you realise it's a choice among the SBs)))) but much longer lasting in history, and it can be assumed that it won't end straight away. But that's not a fact. The task is of course much more difficult, to assess the state of these repetitions/well, like patterns, oh, pseudo-laws)))) The pattern as it is, in almost SB, is of course already something, even to find it, it takes effort, but for forecasting it is nothing. Unfortunately... or I don't know, but that's the way it seems))

The beginning of the pattern emergence and the end of the pattern are interesting.

SB + inefficiencies

Not SB should be predicted almost all, with minor adjustments - totally inefficient market, 10 out of 10. No SB to be predicted, completely efficient. Acts as null hypothesis.

SB + inefficiencies: need to separate the predictable from the unpredictable. There may be degrees of inefficiency, e.g. pattern to noise ratio, in quantitative or other terms.

This is roughly the classification to avoid confusion.

Of course, you should do stat tests if possible, but you are usually lazy.

[fxsaber]

Maxim Dmitrievsky #:

the pattern was found in about 15 seconds.

Since the speed allows, it is logical to look at smaller TFs. If I understand correctly, the pattern length is only about ten values.

[fxsaber]

Rorschach #:

The situation is like in modern physics, do you want to ride or drive? Earlier physics tried to understand how the world works, but now they just stretch formulas over data, invent virtual entities, nobody understands anything, everything is very complicated.

In data processing, it's the same situation. In the past, we took a problem, tried to understand it, then wrote an algorithm by hand, optimised the calculations. To simplify the task, some relationships were neglected, others were reduced to a linear form. When there was enough power and data, the solution of the problem was shifted to an optimiser (roughly speaking, as in MT tester), which selects coefficients of some polynomial. Nobody understands how what is calculated, there is no full confidence in the result, but this approach is able to take into account non-linear and non-obvious relationships, accelerate some scientific calculations by orders of magnitude.

When the solution is obvious, one should use the classical approach. But in conditions of great uncertainty, MO is not a panacea (that's why they add noise to pictures in captcha).

Clear explanation, thank you.

[Maxim Dmitrievsky]

fxsaber #:
Since the speed allows, it is logical to look at smaller TFs. If I understand correctly, the pattern length is only about ten values.

We will look at it one of these days

[fxsaber]

Maxim Dmitrievsky #:
We'll look at it one of these days.

Would go the other way to change the scale.

[Maxim Dmitrievsky]

Antimoshnik in his thread continues to blow his mind 😀😀😀😀 Go to the algotraders forum and deny the very basis of why everyone is here.

I mean like what alternative is offered in return? Alcotrading? 😀

[Maxim Dmitrievsky]

fxsaber #:

Would go in the other direction of changing the scale.

The other day means soon )

[fxsaber]

Maxim Dmitrievsky #:
How's that?

https://www.mql5.com/ru/docs/constants/chartconstants/enum_timeframes

Identifier	Description
PERIOD_CURRENT	Current period
PERIOD_M1	1 minute
PERIOD_M2	2 minutes
PERIOD_M3	3 minutes
PERIOD_M4	4 minutes
PERIOD_M5	5 minutes
PERIOD_M6	6 minutes
PERIOD_M10	10 minutes
PERIOD_M12	12 minutes
PERIOD_M15	15 minutes
PERIOD_M20	20 minutes
PERIOD_M30	30 minutes
PERIOD_H1	1 hour
PERIOD_H2	2 hours
PERIOD_H3	3 hours
PERIOD_H4	4 hours
PERIOD_H6	6 hours
PERIOD_H8	8 hours
PERIOD_H12	12 hours
PERIOD_D1	1 day
PERIOD_W1	1 week
PERIOD_MN1	1 month