I am interested in the way you have build the algo (I am a data scientist). Without revealing your formula, what kind of parameters does your algorithm take. Do you base it on volume, TA thresholds/levels etc? Just to get an idea, as I think it is very interesting to hear what kind of parameters your prediction model are using :-)
It's a system based only on historical prices, which is the only input fed to it,
the output is one of three: LONG/SHORT/UNCERTAIN + additional data that I can't discuss.
I will quote only bitfinex price here, but I feed it historical data from the api of multiple exchanges.
Sometimes the system will signal based only on one/two exchanges and not others, and sometimes they're all in sync (which is a better signal).
I don't use volume, except for threshold tests, meaning if an asset does not have
enough volume in the give time window the algorithm will not run,
so if it's a cryptocurrency or a stock with a very low volume there's no way to reliably predict its movement.
Bitcoin is great for forecasting since it's one of the assets with the highest volume in the world.