OSDev.org

I'm not really into currency trading, but I was wondering about a differential equations based method of predicting. Would it be possible, and would it be practical?

Forex release a free .Net web-service/API (registration required for a unique key), so you can easily set up a simulator and try your idea out. There are also a lot of (mostly commercial and expensive) programs that let you use your own trading logic.

I once stumbled on a forum earlier this year made up of a community of wannabe-rich-quick'ers, mathematicians, and professional traders that were trying everything (including large scale neural network simulation). IRC the simplest methods were the ones that had the most return.

Yeah, the idea is not to make a one and only solution for predicting, but to find something that takes into account a limited number of variables.

How familiar are you with differential equations?

Well with mathematics in general, I started reading calculus books when I was 12.

Well if you've read a good DE book then you should have a good idea of how to use differential equations in modeling. If you're going for modeling with more than 2 variables.

Yes I have a good book. The only real weak area for me is partial differentials as I don't have my desired level of information available to me.

Off topic: I remember in 6th grade I got pissed off because I regarded their method of "prediction" for the path of a hurricane as too simple. I completely confused everybody with my logic. Ahh the fun of geekhood!

I happen to have two copies of Kreyzig's Advanced Engineering Mathematics... PDEs are featured...

It's possible, that said I doubt forex would be the best place to start. Trying to value an entire nation's economy relative to another one isn't an easy process. MessiahAndrw had it right too, usually using a simple model to screen tends to be much more effective. Another thing to keep in mind is that data services for market information usually aren't free and often are not cheap. Even if you had a perfect model many of the measurements would be entirely subjective in nature. If you're serious about it though I'd suggest reading up on macroeconomics and reviewing other investment options first. Grab a copy of the Intelligent Investor and read through it too. There's a lot of good simulators out there for stocks/options as well should you choose to go that route.

I think it sounds like an interesting project.