Signals. Strategies. Trades.
CubistStart from a published strategy.
Hover any signal to see the pattern it trades. Then prove it in the backtester.
Rides a trend once a fast average crosses above a slow one.
Backtest it →Buys the break above a range it has been coiling inside.
Backtest it →Bets price snaps back toward its mean after an extreme.
Backtest it →Fades the move when price stretches outside its bands.
Backtest it →Trades relative to the volume-weighted average price.
Backtest it →Switches behaviour when the market changes state.
Backtest it →Rides a trend once a fast average crosses above a slow one.
Backtest it →Rides a trend once a fast average crosses above a slow one.
Backtest it →Don't bet on one edge. Build a book.
A single strategy is fragile. Combine your proven survivors into a portfolio that trades as one — and let the risk model do the weighting a desk would.
Your other backtests are lying to you.
Six classic lies. Each one closed by default — watch the numbers come back to earth.
Is the edge real?
One number, plus a pass/fail seal. Where the score grades, the seal decides.
Every test passes all five, or it doesn't earn the seal.
We built Minerva to make institutional-grade quant tools available to everyone.
We believe that most observations are the result of luck, not patterns. Trading strategies are fragile if they're not properly understood. Separating the signal from the noise is a hard problem, whether in physics or finance. The key to doing so is testing. And good testing requires two things: expensive computations, and the right data.
Minerva gives you both.
Minerva was born out of a failed search for a rigorous backtester, and a bigger failure to have even the latest and best AI models try to wire one up. The unfortunate truth: all of them produced garbage disguised as an amazing high-Sharpe trading strategy. Why? Because understanding a trading signal requires computational power that your machine at home doesn't have, and so the AI wouldn't build. Quant funds devote massive resources to computer science; so we did, as well.
Our founding team, including PhDs in physics and math, teamed up with advisors from blue-chip quant funds — Two Sigma, Citadel, Millennium, and Cubist — to build the most statistically rigorous quant testing system that is publicly available. It runs on a purpose-designed computing cluster to deliver state-of-the-art statistics at speeds that allow any quant — aspiring or professional — to test many signals across large universes of stocks.
With Minerva, the world of quant trading is finally at your fingertips.