I'm working on a project. I try to beat the Dow. Every week I select stocks from the Dow (usually 1-10 stocks) that I think will beat the Dow. Last week was AXP, GE, IBM, CSCO, and DIS, a few weeks ago was XOM and NKE, you get the idea. I select stocks using a machine learning algorithm I created (engineering undergrad, compsci masters, interned as a quantitative analyst). It can make predictions for any arbitrary portfolio, but people care about the Dow - it's a reasonable benchmark. There's a more in depth explanation on my blog, please check it out:
It's a time-stamped log of all my stock predictions, and a plot of my performance compared to the Dow. It should give you some idea of what I do. You can use the blog to copy my trades in your own portfolio (hypothetically! don't do this! entertainment only etc.) because predictions for Monday mornings are always posted Sunday nights. It's only 9 months of data, but it's all verifiable in an interactive brokers account. I have another 6 months of paper trading data (unverifiable unfortunately) that can be prepended to the graphs.
I've put a lot of work into developing the algorithm. Let me know what you think. I'd love to hear about how to improve things: what kinds of features and sections people want to see. Would people like to be able to log in and be on an email list? Perhaps you like the simplicity? Please, let me know. That's what this post is about. I want feedback on how to improve this.
Heads up iOS users, the graphs don't seem to display properly in iOS safari (OS X safari is fine).
Submitted June 17, 2017 at 03:44PM by AeroZach http://ift.tt/2rIrcoU
No comments:
Post a Comment