Machine learning in trading: theory, models, practice and algo-trading - page 650

 
Ivan Negreshniy:

It's been a long time since anyone posted a report - here's my forward.

It's great! If the EA really was created on the 22nd and has been trading without updates since then, great.

Test on dukascopy ticks from January 1, 2018 to today (10 pips spread). With an arrow marked a noticeable knee, this is just the 22nd day, you can see that profits have stopped growing rapidly, but there is. What will be the next step - time will show.


 

Please attach at least a minimal description of the system to the pictures :) and of course I understand that everything is secret, but just looking at pictures is not very interesting

 
Maxim Dmitrievsky:

I wish MT4 was an anachronism, I don't even have it installed :D Switch to MT5

I think it is better to combine advantages and compensate disadvantages of using both of them, sometimes even several at the same time, including with other platforms, through interprogram interaction, than to use one terminal alternatively.
 
Ivan Negreshniy:
IMHO, instead of alternatively using one terminal, it is better to combine the advantages and compensate for the disadvantages of using them both, sometimes even several at once, including with other platforms, through the interprogram interaction.

that's right... but I haven't seen any other normal terminals, especially testers) the nearest competitors are very far behind

 
Dr. Trader:

That's great! If the EA really was created on the 22nd and has been trading without updates since then, great.

Test on Dukascopy ticks from January 1, 2018 to today (10 pips spread). With an arrow marked a noticeable knee, this is just the 22nd day, you can see that profits have stopped growing rapidly, but there is. What will happen next - time will tell.


There seems to be a need to vary the threshold (BuyLevel, SellLevel) - the default (100%) may be for the training period (up to 22), to assess memorization, and on the OOS may be to decrease.
 
Maxim Dmitrievsky:

that's true... but I haven't seen other normal terminals, especially testers) the nearest competitors are very far behind

If you don't approach it alternatively, you don't need to pull it out, it's enough to complement it)

 
Maxim Dmitrievsky:

the average here is 0, it is a neutral strategy

farther purely on the statistics have to take out) what is there where the shift is already impossible to understand in the black box

In short, almost finished 2 out of 3 strategies planned (the results of the first have already written, did not work)

if this one does not work - the last one will be left )

And this will be the end of the MO

If I've got no results with MO I advice to give up algotrading, it's not your thing, there is nothing in indices.

 
pantural:

If the MO "did not work", then I advise you to give up algotrading altogether, it is not your thing, there is nothing in the turkeys at all.

I'll go teach people how to algo-trade

 
Maxim Dmitrievsky:

I will go teach people how to algorithmic trading

I think that the era of Forex gurus is in the past, modern sucker is more sophisticated, he knows that a small deposit does not disperse, and a large one, you need to spend many years to develop the skills, to the guru now sucker does not go, probably better to look for investors in different asset management schemes, pams, etc.

 

It is now fashionable for dealing companies to call you and, if you open an account, to give you a Personal Tutor. Under his leadership you will open and close trades, he will show you how to make a million. "And right now the euro is in an uptrend, if you open an account right now and deposit 100 dollars - you will earn one thousand by the weekend, but you have to do it today, hurry up or you will miss the big money". That is, the gurus remain, but they are not promoted by themselves, but by the dealing center.

Reason: