About the data.
How every number is made.

The importer requests completed Australian gallops meetings from FormFav. For each accepted race, it stores the winning horse, winning margin, sire, dam, broodmare sire, track, state or territory, distance, race class, trainer and jockey when those fields are supplied.

Each race uses a stable key made from its date, track and race number. A later import updates that record rather than adding a duplicate. The most recent seven days remain visible in the detailed results table, while all successfully stored winners continue contributing to the season rankings.

Distance profiles group winning races into sprint (up to 1200m), mile (1201–1600m), middle distance (1601–2200m) and staying (2201m+). Race classes are grouped from the class description supplied with each result.

Winning margin is taken from the result's explicit winning-margin field when present, otherwise from the margin attached to the runner-up. It is displayed as supplied by FormFav and is not estimated when unavailable.

Runner statistics are calculated only from races where the complete field was successfully captured. This coverage can be shorter than the winner-results coverage, so the dashboard displays its runner coverage dates separately.

The efficiency table begins with the current Top 10 sires by season wins and then compares their conversion rates. A missing percentage means sufficient runner fields have not yet been captured for that sire.

The Value Index is a comparative indicator, not a valuation or breeding recommendation. It combines five performance measures for eligible leading sires.

A sire needs at least 10 completely captured runners to receive a ranked score. Only verified active fees expressed in Australian dollars are eligible; deceased, retired, unverified, unlisted and non-AUD fees remain visible but are not ranked.

Each stored win is credited to the jockey and trainer supplied with that result. National tables combine all eight Australian states and territories; jurisdiction tables count races staged in that location.

The leading partnership is the jockey–trainer combination with the most stored wins nationally during the season. Equal totals are ordered alphabetically. Records without a supplied jockey or trainer are excluded from the relevant table.

Stud fees are the advertised 2026 service fees published by the named farms. The dashboard preserves the farm's stated currency and GST treatment. Links open the source farm page so readers can verify the current listing.

When a current fee cannot be verified, the site labels it as unavailable, retired, deceased or not listed rather than estimating a value. Fees and standing statuses are scheduled for review each April, and should still be confirmed directly with the stud before any decision.

The prediction page scores the runners returned in FormFav's current race form and ranks the best three in every race. The controls are grouped into four main indicators: form and class 30, race suitability 30, speed and finishing ability 30, and connections and breeding 10. Readers can adjust these four weights and recalculate a race or full meeting instantly.

Form and class combines recent results, race-level performance and career win/place rates. Race suitability combines track, distance, going, class fit, expected pace and barrier. Speed and finish combines comparable-distance times with shared closing sectionals. Connections and breeding combines jockey, trainer and sire performance.

Race-level performance ranks Group 1 above Group 2, Group 3, Listed, Open and benchmark/class races. A Group win therefore contributes more than a BM90 win; strong placings receive a reduced share of the relevant class value. The score blends the best class performance with up to three leading class performances and applies a modest recency adjustment.

Race suitability blends the horse's supplied track-and-distance, distance, going and track records, while reducing the effect of very small samples. When the provider supplies a class-fit profile, a suitable level or drop in class rates above a major class rise. Pace and draw fit compare the horse's usual running style with the expected race tempo and apply a modest wide-barrier penalty to leaders and on-pace runners.

Comparable-distance time converts supplied recorded times to metres per second, uses starts within 15% or 150 metres of today's distance, and ranks that evidence relative to the current field. Up to three leading comparable performances are blended, with the best given the most weight. Times are not used when the provider does not supply both a valid distance and recorded time.

Closing sectionals use supplied final 800m, 600m, 400m or 200m splits from starts near today's distance. The model selects the split with the strongest coverage across the field, preferring the final 600m when coverage is equal, and requires at least two runners with the same split before scoring it. Up to three leading splits are blended with a small recency adjustment; missing sectional data is excluded rather than scored as poor.

The optional market indicator compares each runner's median opening and current fresh fixed-odds prices across available bookmakers. A runner shortening in the market rates above one holding steady or drifting. Stale quotes, tote markets and runners without both prices are excluded. Market movement can reflect public sentiment or liquidity and is supporting evidence, not proof of winning ability.

The full-meeting view identifies the leading three runners in each race and highlights the meeting's strongest overall chance. That comparison combines the top runner's model score, separation from the second-ranked runner and evidence coverage, which reduces the tendency to favour a runner merely because it is in a smaller field.

Scores are converted into relative field probabilities that total 100%. They are experimental estimates, not bookmaker odds or guarantees. Track condition, course layout, rail position, race tempo, late scratchings, limited samples and unavailable form can materially change the ranking.