My pet project, Tetsuo, is finally rewritten. The Vulcan build will be called MK-IV.
Day one looked great once it got trading with live money. Then it went down some. It’s behaving like simulations, and that’s fine.
I guess its time to just watch it for a month or so and see how it stands up.
Unfortunately, while the new metrics are immensely helpful for analyzing its performance, there just seems to be no end in sight to getting to a tolerable level of reliability, particularly with directional accuracy. Not that there were expectations — nobody has the stock market figured out. In short runs, I can reliably get it to produce impressive profits, but, the fully automated approach’s lulls are about as significant as its peaks, so, it averages out to about 0 when pulled over, say, a month.
Still, it was a fun project, though I don’t know that I would have spent two years on it if I’d have known what the outcome would be by this point.
In any case, it’s time to let it cook while I pivot to other needs.