Most algorithmic trading products are built on a single bet.
The bet is that if the system finds one or two market patterns that have worked before, returns will compound and investors will get rich. The bet usually fails because real markets contain patterns that did not exist in the training data, and the system gets caught in a drawdown it never recovers from.
We took a different approach when we founded Vincere six years ago.
Our Chicago-based team includes quantitative researchers with institutional trading-system development experience and set out to build a portfolio of algorithms where the design priority was stability first. The thesis was simple. If we could keep the worst losing periods small, the compounding math would take care of itself.
That work produced 12 separate algorithms operating on U.S. regulated futures markets. Each one was built to capitalize on a different type of market behavior. Each one was tested across years of historical data before going live. And each one runs entirely automatically.
The record across six years shows what the design philosophy produced.
