Methodology
Where our data comes from, and how we compute it
RetailVest is a data-driven platform: our signals are built from authoritative public sources — the CFTC, EIA, USDA, NOAA-derived weather, and exchange price data — not opinion. This page documents every source, how often it updates, and exactly how we turn it into the numbers you see. If a figure is on the site, it traces back to one of these.
CFTC Commitments of Traders (COT)
Weekly (released Fridays, for the prior Tuesday)Weekly futures positioning by trader category — commercial hedgers, managed money (large speculators), and small traders — per commodity. We derive net positions, a 3-year z-score, and a hedging-pressure signal from it.
Source: U.S. Commodity Futures Trading Commission · www.cftc.gov ↗
EIA Weekly Petroleum Status (crude oil stocks)
Weekly (Wednesdays)U.S. crude oil ending stocks (excl. SPR), 1982–present. We compute the deviation from the prior-5-year average for the same week of year.
Source: U.S. Energy Information Administration · www.eia.gov ↗
EIA Natural Gas Storage (Lower-48 working gas)
Weekly (Thursdays)Weekly working gas in underground storage, 2010–present, with the same seasonal vs-5-year-average deviation.
Source: U.S. Energy Information Administration · ir.eia.gov ↗
USDA Crop Progress & Condition
Weekly in seasonPercent of the corn and winter-wheat crop rated good/excellent across the growing season.
Source: USDA National Agricultural Statistics Service · www.nass.usda.gov ↗
Weather — degree days & regional anomalies
WeeklyU.S. population-weighted heating/cooling degree days (natural-gas demand) and corn-belt / HRW-wheat-belt temperature and precipitation anomalies.
Source: Open-Meteo / NOAA-derived · open-meteo.com ↗
Futures & spot prices
DailyUp to 40 years of daily futures prices for the metals, energy, and grains we cover, used for trend, momentum, and backtesting.
Source: Exchange-sourced via Yahoo Finance · finance.yahoo.com ↗
Macro indicators
Daily / monthlyTreasury yields (2s/10s), the dollar index, CPI/PPI, breakevens, real yields, and jobless claims — used for regime detection.
Source: Federal Reserve Economic Data (FRED) · fred.stlouisfed.org ↗
How the signals are computed
- Positioning z-score: a trader category's net position standardized against its trailing 3-year history. ±2 is a statistical extreme.
- Inventory / storage deviation: the current weekly level minus the average of the same week over the prior five years, expressed as a percentage — the standard "vs 5-year average" market metric.
- Regime: classified from VIX, the 2s10s yield curve, and equity momentum into states like Goldilocks, Inflation Shock, or Crisis.
- Backtests: run on our historical price data with out-of-sample checks and realistic publication lags on fundamental data, so signals can't peek at information that wasn't public yet. Reported returns are gross of transaction costs unless stated.
What we're honest about
- Our backtest engine evaluates one commodity at a time, so cross-sectional and carry strategies are tracked but not yet backtestable here.
- Natural-gas storage history begins in 2010, which limits how far back those tests can run.
- We publish negative results too — when a strategy doesn't work on our data, we say so. See our strategy research.
- Our strategy research is cross-checked against peer-reviewed academic sources (NBER, SSRN, AQR, the Journal of Futures Markets).
Disclaimer.RetailVest is an educational and analytical tool, not investment advice and not a recommendation to buy or sell any security or futures contract. Trading futures involves substantial risk of loss and is not suitable for everyone. Data is provided "as is" from the sources above and may contain errors or delays. Past performance does not guarantee future results.