The most trivial implementation of a technical analysis trading robot did not work. I'm abandoning that track. It's been nine months or so of experimentation; probably not useful but somewhat interesting.
Before giving up entirely I'll also try applying convolutional artificial neural networks to the problem, but then using ETF's instead of bull/bear contracts. I just need to learn the basics first. This is going to be a one year and three months project at the very least, pretty far from where I envisioned it when starting to hack away at first.
Apart from that a colleague had an interesting business idea we might follow up on. My most interesting spare time projects are being pushed further into the future. Well. Always been a slow starter.
