[Archive] Learn how to make money villagers! - page 733

New comment
 
baykanur:

It's the opposite of pure martin and as you know martin likes to double at the most inopportune moment so you can't win in the casino.

Here's the same one, only in reverse.

I cannot test it now, the terminal is glitchy, I'll test it tomorrow!

 

New creation, once asked for help to refine the owl on the delays

I got it from someone who helped me out, thanks but it worked :)

now there are 2 grails :D


 
OnGoing:

See the graph given by Alexey (bottommost post)https://www.mql5.com/ru/forum/136747/page679.

The main thing is that the systems in the portfolio are statistically independent.

Thanks, I've seen it. But Alexey there also says that the drawdown grows as the square root of the number of independent systems in the portfolio. So if profit of one system is 10, another 15, and drawdowns are the same and equal to 5, then the total profit of systems will be 25 and total drawdown will be 7.07.

And this is a proven statistical truth. If the systems are dependent, the drawdown will either be higher or lower, but as you understand it cannot be calculated analytically.

 
7Konstantin7:

New creation, once asked for help to refine the owl on the delays

I got it from someone who helped me out, thanks but it worked :)

now there are 2 grails :D

Cool. And over a longer period of time, does the opener pass as well?
 
OnGoing:
Cool. And for a longer period of time and on the option minutes also passes?

opening prices don't work owl only needs H1

I will use it on a demo in addition to the one I have...

i have changed the order setting conditions, they got better. in general both owls are based on relatively the same logic)


I thought the owl would work as a portfolio but it did not because of the spread I had to put it on a new demo only on the Euro

for a week

Too little of course, need to force something like mm to stick but i wish i knew how...(

 
alexeymosc:

Thank you, I have seen it. But Alexey there also says that the drawdown grows according to: the square root of the number of independent systems in the portfolio. So if profit of one system is 10, another 15, and drawdowns are the same and equal to 5, then total profit of systems will be 25, and total drawdown will be 7.07.

It is possible, although the FS will still increase in the end)

Although, this is something to sort out. As I understand it, it is only in the case when volumes of deals of each system are not reduced proportionally to number of systems, and saved.

I.e. if each system had 0.1 lot, they also have 0.1 lot in the portfolio, the sum is 0.5.

For true smoothing of indicators we need to decrease the lot of each system to 0.02 in this case. In order to get the initial value of 0.1 in total.

 
OnGoing:


For real smoothing of indicators it is necessary to reduce the lot of each system to 0.02 in this case. To get the initial value of 0.1 in total.

If to do so, and if TC are independent, then it will turn out that drawdowns will decrease proportionally to the lot, but when adding them to the portfolio, they still will not decrease, but only increase.

As I said, we should try to make systems mutually compensate for drawdowns. They should be dependent systems with a negative correlation on drawdowns. )

 
OnGoing:

Possibly, although the FV ends up going up anyway)


Yes, yes, you're right. FS (profit/loss) will increase. But in absolute terms, the drawdown will increase. Therefore, if the systems individually are super risky, their sum will become even riskier.
 

Strangely, the number of trades is almost the same, parameters are always the same (but above are tests from Dukas)

old version

new version

I've always wanted to try it on mt5... if you're up for it let's try it :)

I've got to look for a 5 mark on futures that i don't have a spread.

 
alexeymosc:

If you do so, and if TCs are independent, then it will turn out that drawdowns will decrease proportionally to the lot, but when you add them to the portfolio, they will still not decrease, but only increase.

They will increase in absolute terms, but in the end they will still amount to less than one system with 0.1 lot

alexeymosc:

As I said, you should try to have systems mutually compensate for drawdowns. They should be dependent systems with negative drawdown correlation. )

Now, this seems to me to be a task that is not really achievable. That is why at some time the drawdowns of different systems will coincide and create a kind of resonance. But we will just have to accept it as a temporary phenomenon.

1...726727728729730731732733734735736737738739740...882
New comment