ScraperCaper — A machine-learning trader for the five-minute market

Markets move in five-minute heartbeats. ScraperCaper listens to every one of them — six of the largest digital-asset markets, around the clock — and turns each closed window into evidence. It does not guess. It measures, re-measures, and only commits capital where the numbers keep agreeing.

ScraperCaper is the house's high-frequency mind: a machine-learning trading system that reads six of the largest digital-asset markets in five-minute windows — 1,728 of them a day — and only commits capital when its own evidence says the edge is real.

Every window is recorded, scored and re-tested against real outcomes, fees included. The rules that steer it are rebuilt every five minutes from what actually happened — and when the evidence is not there, it stands aside.