A probabilistic neural network - page 7

 
goldtrader писал(а) >>

I'm not much versed in applying NS to the financial markets, but I can see that even super-specialists like Better fail to make consistent profits using their proprietary tools.

However, that does not mean that NS are useless. They are at least very difficult to use and the result is poorly predictable. IMHO.

A year without profit is an ordeal as long as financial markets are the main source of income.

It is useful to do NS at least because in the process of formalizing price movement you begin to understand much better all subtleties of TA and find out that in many cases you can do without it.

 
goldtrader писал(а) >>

I'm not much versed in applying NS to the financial markets, but I can see that even super-specialists like Better fail to make consistent profits using their proprietary tools.

However, that does not mean that NS are useless. They are at least very difficult to use and the result is poorly predictable. IMHO.

If you do not have a profit for a year, it is hard to believe that financial markets are your main source of income.

No profit, but no loss, right? The year was hard in itself, and very hard for TA. After all, NS, like any kind of TA, uses repetition of patterns of the past in the present, and here there is a crisis and it does not care about the patterns. I do not agree that neural networks are necessarily complex, or difficult to apply. Practice shows that using complex algorithms doesn't particularly do anything...

Reason: