Retro: April 2026 Forecast & Watchlist
Grading: GRADED Jul 11, 2026 — original report: April 2026 Forecast & Watchlist
Graded July 11, 2026, against USDA NASS state prices, AMS trade prints, and the water record through early July.
The one-paragraph verdict
Right about direction, roughly right about magnitude, and right that $225/ton was a defensible price — the market literally printed a 2,000-ton SE Colorado alfalfa contract at exactly $225.00/ton (delivered-feedlot) in June. Colorado alfalfa (state average) rose $165 → $170 → $190/ton from March to May, the fastest two-month climb since 2022. What we got wrong was the demand side: the Texas/Oklahoma drought collapsed instead of persisting, diesel fell 19% when we'd flagged the risk in the other direction, and El Niño came in far stronger than our "weak, mixed signal" read.
Price bands, month by month
| Month | April forecast (low/avg/high) | What actually happened | Verdict |
|---|---|---|---|
| May 2026 | 175 / 210 / 240 | State avg $190; implied SE Premium ≈ $205–225 | ✅ Inside band |
| Jun 2026 | 185 / 225 / 260 | 2,000-ton contract at $225.00 — dead on the avg | ✅ Bullseye |
| Jul 2026 | 195 / 235 / 275 | First-cut asks $250; trades landing $225–250 | ✅ Inside band |
Honest caveat: AMS coverage of SE Colorado is thin (the region printed zero alfalfa quotes in April), so grading leans on a handful of real trades plus the NASS state series. Even so — every observed price sat inside the April bands.
The best call in the report
The statistical model forecast May at $163. We published $210, saying in print that "the model is $30–50 too low because 2026 snowpack is outside its training distribution." Actual: $189–190 state avg — the model missed by $26, and the hand adjustment was almost exactly the miss. Trusting the reasoning over the regression was the single best methodological decision in the April report.
The misses
- Texas/Oklahoma drought collapsed. We leaned on regional drought to keep outside demand strong; June rains and a tropical storm cut the TX+OK drought index from 282 to 127. Biggest demand-side miss.
- Diesel went the wrong way. We flagged Iran-war upside risk; the war ended June 18 and diesel fell 19%.
- El Niño magnitude. We called it "weak, mixed signal." By July it was an official Advisory heading toward a very strong event — which matters enormously for the spring 2027 snowpack, and forced us to cut the spring-2027 price blowout scenario.
What wasn't on our radar at all
- Kansas hay stocks up 70% year-over-year — a substitution cap on feedlot-grade hay two hours east.
- Colorado May 1 hay stocks down 32% — quantified the scarcity we'd only reasoned about.
- Statewide Drought Emergency declared June 4 — first since the 2002/2012-class years.
- NE Bent County went D4 (exceptional drought) July 1.
- Zero Fry-Ark Project allocation — second time in 54 years.
The pattern
Everything we said about SE Colorado's water was right or worse-than-right (Fort Lyon Canal at 21% of last June's diversions, John Martin Reservoir in crisis, 38 active river calls). Everything we assumed about the rest of the map broke the other way. That lesson — grade the supply and demand legs separately — reshaped every report since.
Audit note (added August 9, 2026): an independent audit found this band-grading too generous — it compares the forecast against statewide averages plus an assumed local basis and a delivered contract, not a directly comparable ex-stack trade series. Direction and rough magnitude stand; treat the "inside band" verdicts as directional evidence.
Full technical grading with sources: research/09 in the project archive.