Optimization

Tagged: 

This topic contains 2 replies, has 2 voices, and was last updated by  waterstreet 8 years, 11 months ago.

Viewing 3 posts - 1 through 3 (of 3 total)
  • Author
    Posts
  • #3234

    waterstreet
    Member

    is optimization only brute force ?

    #3235
    Seer
    Seer
    Keymaster

    Currently, yes, only brute force.

    We’ve looked into other forms of optimization (genetic algo’s/monte carlo/goal seek etc) but the benefits of using these approaches is really dependent on the trading system being modeled and the number of possible variations involved.

    Depending on the size of the variation set that’s involved it can be quicker (for example) to do a brute force backtest (than say a genetic algorithm) due to the speed of Seer and the ‘cost’ of the GA. Optimizations under Seer will scale with the number of CPU cores in your box (so almost 4X faster with 4 cores) so the time taken by the brute force approach is less than it used to be.

    That said, we’re always open to new use cases.

    #3237

    waterstreet
    Member

    I agree with you in terms of CPU cost however when you have 8 & 9 variables even with 64 cores brute force could take weeks. The market that this platform is aiming at I would highly suggest putting one in. Seer really has a very clear advantage over other programs like Multicharts and Tradestation. To get clients from those platforms you’re going to want a (genetic algo’s/monte carlo/goal seek). Iv used both those platforms for 5+ years to manage my fund and am extremely excited to see a new contender that actually provides a coding language that is not restrictive. there are 100′s of small shops like me that us MC as their platform and many will flock this way once you have a solid foundation.

Viewing 3 posts - 1 through 3 (of 3 total)

You must be logged in to reply to this topic.