[Home](/) › Statistics

# CropIntel by the numbers

Headline statistics for diligence, citation, and quick reference. Each figure is grounded in the public [track record](/track-record) and [methodology](/methodology).

- 62.3%Tradeable hit rate, walk-forward 2003-2025: of the 61 medium/high-conviction directional calls the model made, 38 were correct.
- 46.3%Direction-correct across all 253 region-years. The gap to 62.3% is the point: the consensus filter only calls when components agree, turning a coin-flip into an edge.
- 92%Bad-year accuracy: of the 26 below-average calls the model made with conviction, 24 were correct, the strategically valuable subset.
- 72%High-confidence call accuracy (13 of 18).
- pSignificance of predicted-vs-actual yield anomaly across the 231 scored region-years (r=0.316).
- 11DEFRA regions modelled independently (Eastern, East Midlands, Yorkshire & The Humber, etc.).
- 253Region-years in the walk-forward backtest population, 2003-2025 (231 with confirmed outcomes).
- 4 layersIndependent data sources feeding the ensemble: weather, soil, satellite, sentiment.
- 2 layersIndependent scoring methods per practitioner observation, a hand-curated UK-agri lexicon and a frontier LLM, blended toward the LLM.
- DailyRefresh cadence, every morning at 07:00 UTC, full pipeline finishes by 07:25 UTC.
- 8 of 9DEFRA regions called correctly in 2019, the canonical UK wheat disaster year.
- 6 of 6DEFRA regions called correctly in 2023, a wet-harvest year.

## How to cite

CropIntel statistics may be cited freely for research, journalism, and diligence purposes. Suggested citation:

CropIntel (2026). UK wheat yield forecasting, walk-forward backtest 2003-2025. Retrieved from https://cropintel.net/statistics

Related: [Full track record](/track-record) ·
[Methodology](/methodology) ·
[Case studies](/case-studies/) ·
[Yield & price](/yield-and-price) ·
[CropIntel vs alternatives](/comparison)