Method

For each active stallion we project the average price of his yearlings at the sale two seasons after breeding, together with an 80% range. Net Profit compares that projection with the stud fee for the season.

The 80% range is a calibrated range from the model, not a statistical confidence interval. It is meant to say how wide the plausible outcomes are, and we report how often past outcomes landed inside it.

How we check it

Projections are backtested point-in-time: each past season is projected using only what was known then, and compared with what the yearlings actually sold for. On the 2023 to 2025 sales, measured on stallions with five or more yearlings sold, the projections rank stallions with a Spearman correlation of 0.92 and a mean absolute error of 0.383 in log price.

Refresh cadence

Projections are refreshed when the Report of Mares Bred lands in the autumn and again after the September sales. Stud fees update through the breeding season, and the season shown switches on September 1. Yearling results update after each sale.

Projections as of October 2026 · model forward-v1.2e · Yearling sales through Oct 22, 2026 (The October Sale)