Trader's self-deception: distrust of forwards. - page 17

[Youri Tarshecki]

Avals:

it's all sophisticated curwafitting. If such a theme was profitable, you'd have to compete with a bunch of quants with super resources from big funds. No fish there)) Unless you create artificial intelligence))

I wonder how you know about quants and super resources from big funds? Do you work in such a fund? With them everything is more prosaic - get a limit, buy and hold. And you do not need artificial intelligence to generate options. We need a quite simple algorithm, especially since the MT solution set is quite limited, although of course not small.

[Youri Tarshecki]

Yuri Evseenkov:
The task seems to me to be very resource intensive. Do you have your own tester or do you use an in-house one?

Yes a normal MT tester, just driven by the autotester. No special resources other than 10 computer cores. You can charge as many variants as you want if you want. Another issue is that you still have to analyse the forward results yourself. But if we want to automate this part, we can write a program that will make Expert Advisors from "bricks" and according to the feedback principle perform evolution of those EAs. The problem is that I cannot write the results of the runs into a separate file so far. To be more exact, I can do it by means of the clicker but it is labor-intensive writing and there is no urgent need in this field.

[Avals]

Youri Tarshecki:
I wonder how you know about quants and superresources from large funds? Do you work in such a fund? Everything is more prosaic with them - get a limit, buy and hold. And you do not need artificial intelligence to generate options. We need a simple algorithm, especially since the MT solution set is quite limited, although of course not small.

I talked to people who work in them)

[Youri Tarshecki]

Avals:
talking to people who work in them)

And how do they use these super resources?

[Avals]

Youri Tarshecki:
And how do they use these super resources?

differently)).

My point is that the idea of autogenerating profitable strategies or even auto-adapting leads to curvafitting monstrosities of varying levels.

Optimisation makes sense as a market research tool, not for autogeneration or autoadaptation. imha

[Youri Tarshecki]

Avals:

differently))

My point is that the idea of autogenerating profitable strategies or even auto-adapting leads to curvafitting monstrosities of different levels.

Optimisation makes sense as a market research tool, not for autogeneration or autoadaptation. imha

This again you are confusing with fitting techniques. Any mechanisation of labour has never been at a disadvantage, it's an objective process and if an idea can be formalised, sooner or later machines will do it.