Is the advisor suitable for real life? - page 5
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Тут всё как обычно:бектестинг на истории, форвард вне зоны оптимизации.
Иначе не честно, зачем себя обманывать? :-)
And what do you do when the forward is unsatisfactory?
In other words, what goes to the furnace(?): the unsatisfied forward test parameter values, or the EA?
:-)
And what do you do when the forward is unsuccessful?
In other words, what goes to the furnace (?): the unsatisfied forward test parameters, or the EA?
I usually optimize to win, until a forward is successful.By the way, it is not thinking of the periods, by which the balance is filtered, so I can come up with a rationale for them.
Обычно оптимизирую до победного, пока форвард не заработает.That's the thing, it's not 'forward earning', it's called.
And it's called - "optimisation with cheating skill has tacked on a forward period of time". :-)That's the thing, it's not "forward earned", it's called.
And it's called - "optimisation with cheating skill has tacked on a forward period of time". :-)well said
In general, the idea of trading the balance curve is not new - it is described by Vince.
Who wants a tester grail on parabolic?
Optimisation for 2000-2006, successfully passes the forward for 2006-2010!
Test for 2000-2010 EURUSD H4:
Interesting?
Or do you think that such a forward can't be trusted anyway?
Forward for 2006-2010 (optimisation was for 2000-2006):
Interesting?
Or do you think that such a forward cannot be trusted anyway?
Optimise at once from 2000 to 2010 according to the min drotown criterion and discover your "good-forward-test" parameters. It's much faster than handcuffing the forward tests.
Зачем так усложнять процесс? :)
Оптимизируйте сразу с 2000 по 2010 по критерию мин дродауна и обнаружьте свои "гуд-форвард-тестовые" параметры. Это гораздо быстрее, чем наручнике отсеивать форвардные тесты.
I understand what you are saying. I.e. if I sifted the parameters manually, then implicitly the optimization was done on the whole sample. The problem is that I took the topmost set of parameters - and got what I got. Perhaps I got lucky.The question is, then, how can we even judge that the EA is not fitted to the story?
I don't need over-optimised "grails" myself, I've been there, I know.
Я понимаю, о чём вы говорите. Т.е. если бы я отсеивал параметры вручную, то неявно оптимизация проводилась по всей выборке. Проблема в том, что я взял самый верхний набор параметров - и получил то, что получил. Возможно, повезло.Вопрос в том, как тогда вообще судить о том, что советник не подогнан под историю?
Переоптимизированные "граали" мне и самому не нужны, плавал, знаю.