OPTIMISATION as the basis of a trading system. - page 4

[[Deleted]]

Demi:

_ nothing prevents you from making money from them.

Why not? There are a lot of factors which interfere. For example, peculiarities of trader's psychology. A trader craves for stable profits. And the word "stability" is somewhat synonymous to the word "stationarity".
There cannot be stable profitability if the process is non-stationary.

[Sceptic Philozoff]

DmitriyN: Why doesn't it interfere? There are many factors that interfere. For example, the peculiarities of trader's psychology.

Usually they speak about robots here, so psychology has nothing to do with it.

There can be no stable profitability with an unsteady process.

Categorical, but wrong. You have no proof.

[Леонид]

DmitriyN: А слово "стабильность" в какой-то степени синоним слову "стационарность".

A professional trader's approach to optimising their TS or selecting the parameters of their TS allows them to obtain stability of profit at a certain drawdown in an unsteady market.

Mathemat : Categorical, but wrong. You have no proof.

+1

[Леонид]

Moreover, some thoughts on the topic of non-stationarity lead to the conclusion that if the market were stationary, the equity line would be flat. It is precisely the non-stationarity that leads to drawdowns and further depo drains on the real, in the future on unknown data with straight equity going into the sky, on the optimization (training) section - the data that we knew.....)))).

[[Deleted]]

Lately, all talk of optimisation (although I prefer the word 'learning') has been bogged down in the theorising debate about stationarity/non-stationarity... In my opinion, it is quite obvious that the market has both of these components. Our task is to train TS so that it earns using stationarity and loses less than it earns on non-stationary component.

For these purposes it seems to me that practical experiments are much more productive, and we have a tool (tester) for that. You can have a bunch of babies without reading a book on physiology, sexology, and holding the Kamasutra, or vice versa, you can read, watch, talk about and remain a virgin.

Topekstarter how do you practically imagine predicting TC parameters in isolation from time series, trading, profits, etc.?

[Alexander]

Figar0:

Topekstarter how do you practically envisage predicting the parameters of the TS in isolation from time series, trading, profits, etc.?

For example, if the optimization parameter is one, and it somehow miraculously fluctuates, for example, along a sine wave with a more or less constant period. In this case the market behaves as a usual market. I don't know whether such a situation is possible or not. I only thought about it some time ago but didn't investigate it. But in theory it is possible.

[Леонид]

4x-online: Но в теории возможно.

Forex market is not mostly theory, but practice - practically, you have to open a real account, put your hard-earned 100 G's and try to make money. Theoretically - we are all millionaires, but practically - not everyone succeeds ))))

[Alexander]

LeoV:
Forex market is not mostly theory, but practice - practically, you have to open a real account, put your hard-earned 100 G's and try to make money. Theoretically - we are all millionaires, but practically - not everyone succeeds ))))

You do not have to bet immediately. It is also possible to explore this hypothesis in a tester to begin with.

[Леонид]

4x-online: Не обязательно сразу ставить. Можно и в тестере для начала эту гипотезу исследовать.

You have to keep in mind that a tester is one thing, a demo is another and the real is a whole separate story....)))) The gap between the tester and the real, and even more so the real above 100k is huge ))))

[Alexander]

LeoV:
You have to keep in mind that a tester is one thing, a demo is another and the real is a whole separate story....)))) The gap between the tester and the real, and even more so the real above 100k is huge ))))

In this case it's just a matter of investigating the ability to predict on fluctuating EA parameter(s). And as for the differences, that's all clear, but if you take the matter so literally, then you don't need a tester with optimization. And there is no point in trading on the real account. :)