Corn Yield Prediction Calculator
Forecast yield weeks before dent. Use plant stand plus ear-size projection to get an early number for marketing and input decisions — with honest caveats about accuracy.
Corn Yield Calculator
Use projected ear counts and kernel counts for an early prediction
Harvestable ear count
Kernel rows around
Length-wise count
Standard: 15.5%
Use "Average of multiple" for accuracy
Ear count is set by R1 (silking). Kernel count is set by R3 (milk). Predictions before R3 are educated guesses; predictions at R4 (dough) tighten to about plus or minus 15%%; at R5 (dent) to about plus or minus 10%%.
Predicting before R4 means kernels can still abort under stress. Build in a 15-20%% downside cushion on any pre-dent prediction, especially in a stressful pollination window.
Predicting Yield Before Dent
The standard ear-count method works best at R5 (dent) or later, when kernel size is set and only drydown remains. But growers often need a yield number earlier — in July, when forward contracting decisions are on the table, or in early August, when input orders for next season get locked in. This page adapts the same USDA method to earlier growth stages by substituting projected values for the inputs that aren't yet finalized, and by being honest about the wider error band that comes with predicting further from harvest.
The key insight is that the three components of yield become measurable at different times. Ears per acre is effectively set by R1 (silking) — once pollination is over, the ear count won't change much. Kernel rows per ear is set even earlier, during ear development around V5-V12, and is locked in by R1. Kernels per row is the wildcard: it depends on successful pollination of each ovule and on grain fill, both of which can still go wrong through R4 (dough stage). The earlier you predict, the more you're guessing at kernels per row.
The Early-Season Prediction Protocol
At R1-R2 (silking to blister): Count ears per acre and kernel rows per ear. These numbers are reliable. For kernels per row, use a projected value based on your hybrid's typical kernel count (check the seed catalog — most hybrids run 30 to 40 kernels per row at maturity). Apply a stress discount: if pollination weather was hot or dry, knock 10 to 20 percent off the projected kernel count. Run the calculator with these projected values. Treat the result as a best-case scenario with a wide error band of plus or minus 20 percent.
At R3 (milk): Kernels are now visible and you can count them directly, though some may still abort under stress. Count actual kernels per row and run the calculator. Error band tightens to about plus or minus 15 percent. This is the first stage where the prediction is genuinely useful for marketing decisions.
At R4 (dough): Kernel abortion is largely over. Count actual ears, rows, and kernels. Error band is about plus or minus 12 percent. The number is now good enough to size storage and trucking.
At R5 (dent): Kernel weight is set. Standard ear-count accuracy of plus or minus 10 percent applies. This is the stage the calculator was designed for.
Stress Discounts: Reading the Crop
The biggest source of error in early prediction is stress that happens after you sample. A heat wave during silking can cut kernel set by 20 percent in a week. A dry August can shrink kernels and reduce test weight. To account for this, discount your prediction based on current conditions. If pollination was clean and the forecast is benign, use the full projected number. If pollination was stressed, discount kernels per row by 10-20 percent. If drought is ongoing, discount the whole yield by 10-15 percent on top of the kernel adjustment.
The honest framing for an early-season prediction is not "the crop will make X bushels" but "the crop is on track for X bushels if the rest of the season cooperates." That framing keeps the prediction useful without overpromising. For marketing, use the low end of your range; for input planning, use the midpoint.
Frequently Asked Questions
Everything you need to know
You can make a rough prediction as early as R1 (silking) by using projected kernel counts, but the error band is wide (plus or minus 20%). By R3 (milk) you can count actual kernels and tighten error to about 15%. By R4 (dough) error is about 12%. The standard plus or minus 10% accuracy of the ear count method applies at R5 (dent) and later.
Before R3, use your hybrid's typical kernel count from the seed catalog (usually 30-40 per row) and apply a stress discount. If pollination was clean, use the full projected count. If pollination weather was hot or dry, knock 10-20% off. Once kernels are visible at R3, count them directly instead of projecting.
A July prediction (at R1-R2) has an error band of about plus or minus 20%, which is too wide for aggressive forward contracting but useful for sizing up the crop's general trajectory. By R3 (early August) the error tightens to about 15%, which is good enough for moderate marketing decisions. For high-confidence contracting, wait until R5.
Heat (above 95F) or drought during silking can cut kernel set by 10-30%, primarily by reducing kernels per row. If you're predicting yield shortly after a stressful pollination, discount your projected kernel count by the amount of stress observed. Walk the field and look for bare cob tips or missing kernels — a visible sign of pollination problems that should directly reduce your estimate.
Editorial Standards & Sources
All formulas and benchmarks on this page are based on the USDA Yield Component Method (Nielsen, R.L., Purdue University Extension) and USDA NASS crop production data. Our calculators are reviewed annually against the latest extension publications from Iowa State, University of Illinois, University of Nebraska, and Ohio State.
Methodology: The default 90,000-kernels-per-bushel divisor and the 15.5% moisture standard used throughout this site come directly from published USDA and land-grant extension guidance rather than an internal estimate. Where a figure is our own approximation — such as the ±10% accuracy range for the ear count method — we say so explicitly rather than presenting it as an official statistic.
Known limitations: Pre-harvest estimates from any ear-count-based calculator are a planning tool, not a substitute for actual combine yield-monitor data or elevator scale tickets. Accuracy depends heavily on how representative the field samples are; we recommend a minimum of five sample locations per field, as outlined in the sampling guidance on this page.
Content last reviewed for accuracy in August 2025. Have a correction or suggestion? Contact our editorial team — we respond to all substantive feedback within 5 business days.