SuperQuant runs on the same engine institutions use — XTech, the data infrastructure trusted by leading hedge funds and investment banks. Nothing is downsampled for individual users. Here is what that engine has proven, in plain English, with the published research behind every number.
Flows 101 — how money moves marketsMacro 101 — seeing inflation before it's printed
Chapter 1
“Flows” are the dollars that institutions and retail traders push into and out of stocks. XTech has tracked them minute-by-minute since 2007. Here is what the research found.
Institutional buying and selling alone explains most of the market's moves: about 63% of the S&P 500's quarterly swings, and up to 75% day-to-day.
When institutions buy today, roughly a third of the price bump fades over the following weeks. The other ~70% sticks for good.
Every $1 of net institutional buying lifts a stock's market value by about $7–$10. Markets are far less elastic than the textbooks say — small imbalances have outsized effects.
Institutional flow predicts where money is really going — up to 71% accuracy against official 13F filings that arrive months later. Retail flow predicts essentially nothing. Separating the two is what turns raw volume into a usable input.
In the most volatile stocks, the predictive power of flow roughly doubles — the signal is strongest exactly when markets are hardest to read.
Energy is the most flow-sensitive sector (~10x impact); Materials and Tech the least. The harder a stock is to replace in a portfolio, the more flows move it.
Chapter 2
Once a month the government prints the CPI — the headline inflation number that moves trillions in bonds, currencies, and stocks. XTech built forecasts that see it coming. Here is what the research found.
XTech's CPI forecasts land about 20 days ahead of the government's release — and get the direction right 81.9% of the time, versus 75.5% for the market consensus (November 2017 – September 2025).
Fewer than 1 in 5 professional economists reliably beat the consensus. XTech does, consistently, with lower errors across both headline and core inflation.
In 2025, the forecasts called the direction of inflation correctly 100% of the time (consensus: 78%), with roughly half the error.
Instead of guessing the headline, the models forecast the pieces — gasoline, used cars, shelter, and the rest — then add them up using official CPI weights. Detail beats extrapolation.
In backtests, simple systematic rules built on the forecasts won 65–80% of trades in currency and bond futures around CPI releases.
Those backtested strategies showed Sharpe ratios near 1, drawdowns under 3%, and a positive statistical edge that held across markets — not a handful of lucky bets.
Every finding above comes from the engine that powers SuperQuant's dashboards, live in OpenBB. Pick your data packages to get started.
Choose your tier →SuperQuant is a software and data product built on Exponential Technology. It does not provide brokerage, investment advisory, financial planning, accounting, legal, or tax services. All statistics on this page are drawn from XTech published research over the periods stated; backtested results are hypothetical and do not reflect actual trading. Past performance is not indicative of future results.