Quantopian's community platform is shutting down. Please read this post for more information and download your code.
Back to Community
Combining Algorithms / Predicting stocks with Black-Scholes

I was just wondering if anyone has had a go at combining a finite number of different trading algorithms through Fourier analysis whereby a weighting can be assigned to each algo (based on previous success for the specific stock, market conditions ect.) and the sum of all taken and used to trade with.
I also wanted to see if anyone has had a go at propagating the Black–Scholes equation and traded based on this prediction. I was thinking of just using a finite element or difference method for this.
If anyone is interested in having a go at either of these with me let me know. Advice would also be great.

2 responses

Hello Steven,

This article is not at all helpful to you but it is quite amusing:

Scholes knows, can’t help but know, that the model has failed. He’s stonewalling for a very simple, sleazy reason: his whole life depends on it. It’s his whole schtick, the foundation of not just of his wealth but his whole idea of himself. For quasi-celebrities like Scholes, even sexual success is bound up with the model. The model made him rich, famous and adored. He can’t afford to entertain any doubts about it. So he insists, in the face of total disaster, that the model is perfect. It can’t be the model’s fault, so it must be the world’s fault for messing up the model.

See: http://exiledonline.com/myron-scholes-nobel-prize-winner-con-man-and-high-priest-of-derivative-divinity/all/1/

Regards,

Peter

I believe the terminology is Committees/Ensembles where you add an extra layer that controls how multiple algorithms are applied to an overall regime.