The method of optimizing the parameter of EA is written. The person who knows thinks it might be unexpectedly few.
This is how to use “Optimization”

I think that I adjust the value of the place of “Value” displayed with a red frame usually first pushing “Expert properties”, push “Start”, and do the backing test.

However, when it is this, the backing test is done only by this value. I want to know what result to become at a dash by the value to 100~500. Then, the function of “Optimization” is used.
The check is put in the place in which Optimization first of all. Afterwards, the value is put in the start, the step, and the stop on the right side. This becomes a thing of doing ten backing tests from 100 to 500.

The check is put in Optimization of the image below, and the start is pushed.
This is simulated in 41 patterns while testing backing now.

The result of the backing test can be confirmed at a place called “Optimization Results” and “Optimization Graph”.
Stoploss is displayed from 480 to 500 as for the result, and PF1.0 or less is not displayed here.
When nothing is displayed here, it becomes a thing that there was none in PF1.0 or more in the value. (There is a method of displaying the result of 1.0 or less, too. )

The following test back from 480 to 1500.

The result has expanded to PF1.30 in stoploss810. (It is possible to expand by clicking the image.)

When the item of 810 on is double-clicked most, the value is reflected in the “Value” part.
Many items to be simulated can be done at the same time in Optimization. It can be possible to do stoploss and takeprofit at the same time, and do the thing done by all parameters. (When all parameters are simulated, it takes considerable time. )

Please wanting you to note it must not judge with PF alone and judge the numerical value while seeing Total trades, Expected Payoff, and Drawdown %, etc.

Finally, it is even if PF of a good numerical value goes out and there is no security to be going to obtain a similar result from tomorrow as it is. The backing test is a past result to the end.

*Take care about the curve fitting.



▐ 最真· 我▐

evilrainz 發表在 痞客邦 PIXNET 留言(0) 人氣()