Comments on: Better Strategies 4: Machine Learning https://financial-hacker.com/build-better-strategies-part-4-machine-learning/ A new view on algorithmic trading Mon, 07 Oct 2024 09:24:17 +0000 hourly 1 By: jcl https://financial-hacker.com/build-better-strategies-part-4-machine-learning/#comment-108010 Mon, 07 Oct 2024 09:24:17 +0000 http://www.financial-hacker.com/?p=931#comment-108010 In reply to Connor.

CAGR is based on profit divided by capital. The more profit, the higher the CAGR. The more capital, the lower the CAGR. Of course only if you forgot to increase the investment accordingly.

]]> By: Connor https://financial-hacker.com/build-better-strategies-part-4-machine-learning/#comment-107948 Sun, 06 Oct 2024 02:34:36 +0000 http://www.financial-hacker.com/?p=931#comment-107948 I have a few questions. I’m wondering why my machine learning system’s CAGR decreases as I increase the capital amount. It shows parallel to Z7 which is labeled as a small-capital forex trading system. My system also focuses on forex pairs but I’m still confused as to why the CAGR decreases as I raise the capital level. Does this mean a system like this is capped out at $10,000 capital or can it still trade profitable at like $50,000 capital? Secondly, I’m having trouble defining position sizing in the code. I want position sizing to be dynamic based upon the capital and profit closed. But, using the margin function risking a certain percentage size of my account isn’t taking trade amounts based upon that. I’m wondering how I would use the margin function or any other function to determine proper position sizing. There aren’t many zorro script examples that show how to correctly code position sizing. Explaining all of this would be paramount. Thanks

]]> By: Kiran Rama https://financial-hacker.com/build-better-strategies-part-4-machine-learning/#comment-92589 Tue, 04 Jul 2023 14:37:25 +0000 http://www.financial-hacker.com/?p=931#comment-92589 Ah – of all the features in the blog, The Ehler’s were massive overfits. Funnily the highest predictive power was for price to 52week low ratio!

]]> By: quan nguyen https://financial-hacker.com/build-better-strategies-part-4-machine-learning/#comment-90946 Thu, 20 Apr 2023 02:39:56 +0000 http://www.financial-hacker.com/?p=931#comment-90946 Thank you very much!

]]> By: jcl https://financial-hacker.com/build-better-strategies-part-4-machine-learning/#comment-90914 Wed, 19 Apr 2023 10:22:48 +0000 http://www.financial-hacker.com/?p=931#comment-90914 I can answer that relatively precisely: By 2022, 66% of data mining systems that we did so far for clients had been successfull, and 34% did not meet the success criteria.

]]> By: quan nguyen https://financial-hacker.com/build-better-strategies-part-4-machine-learning/#comment-90861 Mon, 17 Apr 2023 23:37:51 +0000 http://www.financial-hacker.com/?p=931#comment-90861 Thank you for the very helpful article. It is now 2023, and I wonder if your comment “Compared with model based strategies, I’ve seen not many successful machine learning systems so far” is still true today? Thank you very much.

]]> By: jcl https://financial-hacker.com/build-better-strategies-part-4-machine-learning/#comment-84167 Fri, 04 Nov 2022 14:20:26 +0000 http://www.financial-hacker.com/?p=931#comment-84167 Thanks for the link, I’ll look into that. Though convolution stages are normally used for image recognition, 1-D convolution has been useful for price curve preprocessing.

]]> By: tovim https://financial-hacker.com/build-better-strategies-part-4-machine-learning/#comment-84152 Fri, 04 Nov 2022 07:02:38 +0000 http://www.financial-hacker.com/?p=931#comment-84152 Hi Johann,
There are maybe hundreds of new sophisticated machine algorithm architectures emerging .But this one looks interesting :have you heard of The Convolutional Tsetlin Machines(stochastic learning automata)? Here is the link https://arxiv.org/abs/1905.09688

]]> By: Peter Williams https://financial-hacker.com/build-better-strategies-part-4-machine-learning/#comment-44040 Thu, 20 Dec 2018 10:14:46 +0000 http://www.financial-hacker.com/?p=931#comment-44040 Thank you,
Understand your comments – but sort of came to this site from Zorro and your own Black Book

]]> By: jcl https://financial-hacker.com/build-better-strategies-part-4-machine-learning/#comment-44038 Thu, 20 Dec 2018 10:02:25 +0000 http://www.financial-hacker.com/?p=931#comment-44038 Looks like a typical R version issue, but this blog is not a good place for questions like that – better ask on the Zorro forum, or an R forum. The R code here is meanwhile quite old.

]]>