Gamestudio Links
Zorro Links
Newest Posts
Blobsculptor tools and objects download here
by NeoDumont. 03/28/24 03:01
Issue with Multi-Core WFO Training
by aliswee. 03/24/24 20:20
Why Zorro supports up to 72 cores?
by Edgar_Herrera. 03/23/24 21:41
Zorro Trader GPT
by TipmyPip. 03/06/24 09:27
VSCode instead of SED
by 3run. 03/01/24 19:06
AUM Magazine
Latest Screens
The Bible Game
A psychological thriller game
SHADOW (2014)
DEAD TASTE
Who's Online Now
0 registered members (), 863 guests, and 4 spiders.
Key: Admin, Global Mod, Mod
Newest Members
sakolin, rajesh7827, juergen_wue, NITRO_FOREVER, jack0roses
19043 Registered Users
Previous Thread
Next Thread
Print Thread
Rate Thread
Montecarlo statistical mode parameter optimization #475467
12/12/18 11:40
12/12/18 11:40
Joined: Aug 2017
Posts: 102
Spain
B
Brax Offline OP
Member
Brax  Offline OP
Member
B

Joined: Aug 2017
Posts: 102
Spain
I was wondering if i can access to the montecarlo generated curves of a strategy test and all the parameter values for each one of them.

The idea is take the best 'n' montecarlo curves, extract the parameter values associated and then compute the statistical mode for each parameter.

We would take this parameter modes as the supposedly most robust ones.

¿Has anyone tried this? ¿Is it worth doing it? ¿How could i do it with zorro and other tools?

Thanks.

Re: Montecarlo statistical mode parameter optimization [Re: Brax] #475468
12/12/18 12:09
12/12/18 12:09
Joined: Feb 2017
Posts: 1,718
Chicago
AndrewAMD Offline
Serious User
AndrewAMD  Offline
Serious User

Joined: Feb 2017
Posts: 1,718
Chicago
Monte Carlo just randomly shuffles the order of trades from a backtest to see what effect it has on drawdown and profit. It’s an evaluation tool.

Re: Montecarlo statistical mode parameter optimization [Re: AndrewAMD] #475479
12/13/18 09:52
12/13/18 09:52
Joined: Aug 2017
Posts: 102
Spain
B
Brax Offline OP
Member
Brax  Offline OP
Member
B

Joined: Aug 2017
Posts: 102
Spain
I understand that, but i´ve seen this idea mentioned on an article on a blog and i wanted to know if it´s feasible in zorro.

Anyway, when the author of the article releases the second part, i´ll see how exactly this idea works and if it has any sense.

AndrewAMD, in your opinion, ¿What´s the best approach for selecting good robust parameters? ¿Is it enough with the zorro approach?.

Thanks.

Re: Montecarlo statistical mode parameter optimization [Re: Brax] #475481
12/13/18 12:36
12/13/18 12:36
Joined: Feb 2017
Posts: 1,718
Chicago
AndrewAMD Offline
Serious User
AndrewAMD  Offline
Serious User

Joined: Feb 2017
Posts: 1,718
Chicago
May I have a link to the article?

Re: Montecarlo statistical mode parameter optimization [Re: AndrewAMD] #475489
12/14/18 13:14
12/14/18 13:14
Joined: Aug 2017
Posts: 102
Spain
B
Brax Offline OP
Member
Brax  Offline OP
Member
B

Joined: Aug 2017
Posts: 102
Spain
Of course.

This is where i first saw this idea mentioned at the end of the post (spanish article):
https://www.rankia.com/blog/oscar-cagigas/4101414-cual-mejor-ratio-optimizacion

In short, the author performs a comparison of different ratios for obtaining system parameters, and he concludes that in the second part, he will apply the statistical mode of a given ratio over the whole montecarlo generated curves.

This is another article with a more detailed explanation, adding some MQ5 code:
https://www.mql5.com/en/articles/4347

Looking forward for your reply.

Bye.

Re: Montecarlo statistical mode parameter optimization [Re: Brax] #475495
12/15/18 14:18
12/15/18 14:18
Joined: Feb 2017
Posts: 1,718
Chicago
AndrewAMD Offline
Serious User
AndrewAMD  Offline
Serious User

Joined: Feb 2017
Posts: 1,718
Chicago
I have some thoughts, in no particular order:

* The first article does not talk about monte carlo at all, from what I can tell.
* I wouldn't know what the "best" optimization method is, I simply use the most practical ones. Currently, this would be the default Zorro optimization method.
* Perhaps the method described in the second article can be a useful technique, it might be worthy of experimentation.
* One other optimization method I like is described in Trading Systems by Jaekle and Tomasini. When they initially design a trading system, they use a dual-parameter brute force optimization to produce various 3D plots, in order to identify stable profitable parameter regions. This helps with the final design of the script. I have been able to produce such 3D plots by exporting trade data CSV's and making 3d plots with the plotly R package. (Anyways, this has nothing to do with statistics.)

Re: Montecarlo statistical mode parameter optimization [Re: AndrewAMD] #475496
12/15/18 18:50
12/15/18 18:50
Joined: Aug 2017
Posts: 102
Spain
B
Brax Offline OP
Member
Brax  Offline OP
Member
B

Joined: Aug 2017
Posts: 102
Spain
Hello.

* About the first article, when the author says he will consider the statistical mode of the best 100 equity curves, i interpret he is obtaining these curves from montecarlo analysis. Anyway, when he publishes the second part, i will tell you.

* I also think that Zorro style is the simplest and most practical method for choosing parameters, but one always like to experiment new things.

* I will have a look to the book you have told me.

Thank you for your reply.


Moderated by  Petra 

Gamestudio download | chip programmers | Zorro platform | shop | Data Protection Policy

oP group Germany GmbH | Birkenstr. 25-27 | 63549 Ronneburg / Germany | info (at) opgroup.de

Powered by UBB.threads™ PHP Forum Software 7.7.1