CONSIGNER RATINGS
Consigner ratings measure something most buyers care about but rarely quantify: do the horses this operation puts through the ring go on to run? Our algorithm tracks post-sale racing outcomes for every horse breezed, applies statistical correction for sample size, and produces a single composite score.
The foundation of every consigner rating is racing performance. We track four metrics for every horse a consigner breezes at a sale, whether the horse sells or not. This is an important distinction — RNA horses still reflect on the consigner’s program.
% Started — What percentage of horses breezed went on to make at least one start on the racetrack. This is the most fundamental measure. A consigner whose horses consistently make it to the track is doing something right in selection and preparation.
% Won — What percentage went on to win at least one race. Getting to the track matters, but winning matters more. This is weighted equally with starting rate because both are reliable indicators of consigner quality at reasonable sample sizes.
% Stakes Winners — What percentage produced at least one stakes victory. This separates good programs from elite ones, but it’s inherently rare — even the best consigners typically produce SW rates in the low teens.
% Graded Stakes Winners — The top of the pyramid. Graded stakes winners are so rare that this metric carries the least weight, but when a consigner consistently produces them, it’s a powerful signal.
A consigner who breezes 2 horses and both win has a 100% win rate. A consigner who breezes 60 horses and 35 win has a 58% win rate. Which one is actually better at producing winners?
You don’t know, and that’s the point. Raw percentages are unreliable at small sample sizes. A single lucky horse can make a two-horse consigner look like Godolphin. We solve this using Bayesian shrinkage — a statistical technique that pulls small-sample rates toward the population average. The smaller the sample, the stronger the pull.
In practice: a consigner with 2 horses and a 50% graded stakes winner rate gets adjusted down significantly (toward something like 8%), while a consigner with 60+ horses barely moves. This means you can trust the ratings for large and small consigners alike — the math accounts for the uncertainty.
Each consigner’s final score (0–100) combines three components.
Racing Outcome Score — The dominant factor. This blends the four Bayesian-adjusted racing metrics (started, won, stakes winner, graded stakes winner) with heavier weight on starting and winning, lighter weight on stakes performance. The rationale: getting horses to the track and winning are the most reliable and repeatable indicators of consigner quality. Stakes results are too rare and variance-prone to carry heavy weight.
Volume Reliability — A smaller component that rewards consigners with larger sample sizes, capped at a threshold. This isn’t about rewarding size for its own sake — it’s about statistical confidence. A consigner with 40 horses has a more trustworthy track record than one with 4, all else equal.
Market Validation — Consigners who breeze horses but sell none of them receive a penalty. If the market won’t buy your horses at any price, that’s a meaningful signal. Consigners with at least one sale are unpenalized.
Tiers are assigned by percentile rank of the composite score across all consigners:
| Tier | What It Means |
|---|---|
| ELITE | The best consigning operations in the sale. Consistently high start rates, win rates, and stakes production across meaningful sample sizes. |
| STRONG | Above-average programs with solid racing outcomes. These consigners reliably produce runners and winners. |
| ABOVE AVG | Better than the median consigner. Horses from these operations are more likely than not to make it to the track. |
| AVERAGE | Middle of the pack. Outcomes are in line with the sale-wide averages. |
| BELOW AVG | Below-average racing outcomes. Often small sample sizes with limited track record. |
| WEAK | Bottom of the rankings. Typically zero-sold consigners or operations with very few horses that rarely reach the track. |
In the consigner ratings table, you’ll see both raw and adjusted columns. The “Adj” columns are the Bayesian-adjusted values that feed into the composite score. For large consigners (30+ horses), raw and adjusted values will be similar. For small consigners, the adjusted values may differ significantly from raw — that’s the correction working as intended.
When comparing consigners, always use the adjusted columns. They’re the apples-to-apples numbers.
The current consigner ratings cover OBS 2024 March and April combined sales (1,474 horses, 170 unique consigners). Ratings are based on racing outcomes through the data collection date and will be updated as more race results become available. Consigner names are normalized to merge agent variants (e.g., “Eddie Woods, Agent VIII” rolls up under “Eddie Woods”).