I should. Yet what I'm trying to express in my comment is that the part of the system that would analyze the data and then make trades is not only unfinished, is still unstarted. So I have continually polished the ingest portion but felt "stuck" when contemplating how to get the system as-a-whole into a testable state.
The broker API I plan to use features the ability to "paper trade" in a simulated account. So I'm not so worried about switching the thing on one day and immediately blowing my account.
I do trade manually in a broker account but as fun and instructive as that is, it doesn't represent the kinds of high-speed analysis and execution my system would make once it was completed.
Seems like you would greatly benefit from just coding something that trades, even if it sucks, and give it a little money. It's not likely to be that much worse than 50/50 anyway.
