A winner percentage appears to offer a quick answer to a difficult breeding question: which sire has produced stock suited to an Arabian racing programme? Yet a percentage cannot explain itself. It might describe two winners among three runners or many winners from a much larger group. It might cover one domestic season or careers spread across countries. Before admiring the number, establish who entered the calculation, who remained outside it and which racing opportunities the groups actually received. An arithmetically correct table can otherwise support a decision that its underlying information does not justify.
The Arabian Jockey Club’s report descriptions distinguish foals, starters, winners and stakes winners, with breakdowns by foal crop and sex. Racing bodies publish results and championship tables within particular periods and jurisdictions. What the reader needs is therefore not another magical index but disciplined attention to definition, provenance and coverage. This guide stays within purebred Arabian racing. It neither compares the breed with another racing population nor turns a sire’s observed progeny record into a direct measurement of genes or a probability that a future foal will succeed.
Every numerical example below is wholly hypothetical. It illustrates arithmetic and is not a live ranking or a record attributed to any real stallion. The analytical framework can be applied to verified official records while preserving separate currencies and acknowledging missing information. Its purpose is to sharpen the breeder’s question: what does this measure establish about this particular group, and which additional evidence is needed before it influences a breeding choice?
1. Establish what is counted before reading success
Every ratio needs a numerator and denominator defined consistently throughout the comparison. For winners to runners, count distinct offspring that won at least once within the scope over distinct offspring that actually raced within that scope.
A horse winning five races remains one winner in that measure; its five successes belong in the wins column. If a provider changes what qualifies as a runner or changes result coverage, two versions of its table cannot automatically be combined into one comparable record simply because their column headings look familiar.
Do not confuse runners with starts. One horse contesting six races represents one individual and six starts. Ten offspring racing three times each produce ten runners and thirty starts. The first number describes the breadth of the tested group; the second describes its exposure to racing.
A useful reading preserves both rather than describing thirty starts as thirty progeny. Read the table heading and notes carefully. Different services can use a translated word for different units, making an important arithmetic distinction look like a minor wording issue.
Registered foals are also different from offspring that have reached the relevant age and actually raced. Runners to foals asks about reaching the racecourse within a defined group. Winners to runners asks what happened among those that did so. Neither measure replaces the other.
An unraced foal is not automatically a sporting failure: it may not have reached comparable age, may lie outside the data coverage or may have followed another path. If the eligible foal population is unavailable, do not invent a denominator to complete the report.
Reliable counting begins with stable horse identity. Preserve a registry identifier, name, parents and verified country and birth year, so a spelling variant does not become an extra offspring in the statistics. Conversely, two animals sharing a name must not be merged.
Associate each result with its race, date and approving authority, distinguishing official, provisional and corrected outcomes. These are background record requirements, not a reason to expose a public horse directory. They prevent duplicated horses or results and let an aggregate be traced back to identifiable evidence.
2. A worked example: the question changes the measure
Consider two entirely hypothetical groups of Arabian racing progeny in the same country and period. Sire A has ten runners, four individual winners and eight wins from thirty starts. Sire B has forty runners, twelve individual winners and eighteen wins from one hundred starts.
Winners to runners is therefore forty percent for A and thirty percent for B. The higher percentage is real within the example, but it does not erase the fact that B supplied more distinct winners and was represented by more horses and racing opportunities.
Change the question to wins per start and the calculation changes: eight divided by thirty is approximately twenty-six point seven percent, while eighteen divided by one hundred is eighteen percent. This is a different measure, not an automatically more precise version of winners to runners.
Successful offspring can return repeatedly, while others may need several attempts before winning. Display the full measure name and the underlying counts. Calling both “win rate” and assuming that every reader or provider uses the same numerator and denominator hides the most important distinction in the comparison.
Next ask how wins are distributed among offspring. A group whose successes mostly come from one horse tells a different story from one whose victories are spread more widely, even when total wins match. Concentration does not diminish the exceptional horse’s achievement.
It does mean that the record answers the question of producing a standout more directly than the question of breadth of success. Preserve distinct winners beside total wins and, where available, show the leading offspring’s contribution. This describes the composition of the record; it is neither a penalty for having a champion nor an invented ranking.
This reading prevents an automatic verdict that A must be the better sire. One breeder may value a larger record to investigate; another may need evidence at a particular distance. Either choice requires information relevant to the stated objective.
The example estimates no future foal’s chance and recommends neither A nor B, which are not real stallions. Its purpose is to turn a percentage into answerable questions: how many offspring were tested, how often did they race, how was success distributed, and does the group actually resemble the ages and races targeted by the programme?
3. Period, age and country: align opportunity
Define the time window before downloading a table, including its start, end and last verification date. A current season is not a sire’s lifetime record. Combining three seasons is not achieved simply by adding each season’s individual winners, because the same offspring may win in all three and should count once when the question concerns distinct winners across the combined period.
Its separate race wins remain separate results. A figure labelled “current season” cannot become a lifetime study without an explicit account of its coverage and any changes to that coverage.
Compare foal crops with attention to age and opportunity. A stallion whose first runners have only recently reached the racecourse has not had the same observation window as one represented by older horses across many seasons. Show ages and crops when the source permits, without assuming that absence of early wins settles later development.
Narrowing a comparison to similar ages and scope can help, but does not remove differences in dams or training. If filtering leaves only two or three horses, make that plain. A precise-looking percentage cannot compensate for a very small observed group.
Country is not a decorative field. It defines the programme, available opportunities, race levels and earnings represented in the record. An offspring can race in two countries during one period; a global count of distinct runners should not double it because it travelled.
A domestic comparison must state that performances elsewhere are excluded. Do not assume one database covers every jurisdiction equally well. Reading IFAHR’s scope and checking local authority results helps identify limitations, but displaying a national flag does not prove that every result from that jurisdiction is present.
Opportunity also needs a sporting description, not only a count. Retain verified distance, surface, race category and conditions where available, and check what a classification means under the relevant authority. Success in one collection of races establishes achievement there, not universal superiority across Arabian racing distances.
Frequent participation is neither good nor bad by itself. Review official results and individual progression before assigning a pattern to the sire. Official performance tools also have context; figures produced under different systems cannot simply be added together without a justified interpretation.
4. Small samples and missing records: precision can mislead
A second hypothetical sire has three runners and two winners, giving approximately sixty-six point seven percent. Add a fourth runner who does not win and the percentage becomes fifty; with five runners and the same two winners it becomes forty. The two winners have not lost their achievements.
The denominator has changed. This illustrates sensitivity in a small group, not proof that the initial calculation was false or that the decline was genetic. Displaying count and percentage together lets the reader see how much one additional individual can change the description.
No magical threshold makes every comparison reliable once the count passes it. Size matters, but it does not repair wrong identity, selective coverage or major differences in age and opportunity. A study reporting uncertainty intervals must explain its method, assumptions and observation unit.
Repeated starts from one horse are not automatically independent observations, nor are horses under one management randomly assigned circumstances. This simple example therefore carries no invented confidence bands. It teaches arithmetic rather than pretending to estimate statistical certainty from a study design that has not been carried out.
Separate zero from unknown. If a complete source confirms that ten offspring raced and none won in the defined period, zero winners is information. If some race records or fields are unavailable, replacing the gap with zero is not legitimate.
Preserve the missing-data state and last verification date, and do not build sporting or earnings rankings on that substitution. Sometimes the right response is to defer comparison or show coverage alone. Filling a cell to satisfy the layout adds no knowledge; it hides the reason that a judgement remains incomplete.
Watch how the group was selected. Advertising that displays the best offspring does not describe all runners, and a record limited to an export destination may omit animals that stayed elsewhere. Absence is not evidence of inability or non-participation, and a selected showcase cannot be compared with a comprehensive record.
Ask for inclusion rules, original sources and a dated version. Recalculate when late information arrives while retaining the earlier snapshot, so the team can distinguish new results from identity corrections or wider coverage. These reasons for a changed ranking have different meanings.
5. Earnings and achievement: separate money from breadth
A sire may lead total earnings because one offspring won a particularly valuable race. That documented achievement matters, but it does not necessarily show a broad winner base or an elevated opportunity for every offspring. Separate total earnings, the number of earners and the leading horse’s contribution.
If complete individual amounts are available, a median or another defined descriptive measure may be useful. It cannot be reconstructed from one aggregate or only the top five earners. Information unsupported by the source remains unknown even when an attractive chart would be easy to fabricate.
In a wholly hypothetical example using one currency, two groups each earn one hundred thousand monetary units. One horse contributes sixty thousand in the first group, while the largest contribution in the second is twenty thousand. Totals match, but concentration differs: sixty percent against twenty percent for the largest contributor.
Neither distribution is declared universally better; the point is that a total does not describe its distribution. These are not net profits. Breeding, training, transport and other costs are absent from the example and are not automatically deducted by a race-result amount.
Preserve every amount with its currency code, scope, verification date and source. Euros, riyals and pounds must not become one total or a ranking of bare numbers treated as equivalent units. An official converted series has a method, rate, date and scope that need disclosure; this guide creates no conversion of its own.
Even within one currency, check that period earnings are not mixed with lifetime totals and that premiums or breeder bonuses follow a consistent inclusion rule. Similar-looking money field names do not guarantee that they contain the same thing.
Describe the level of achievement alongside money. Distinct winners of classified races differ from the number of classified-race wins, and one horse may recur across seasons. Verify the official category of the exact edition rather than granting international status because the purse is large.
Preserve distance, age and surface where available so the group’s sporting context remains visible. Combining unlike classifications into invented points and calling the result “genetic value” exceeds the evidence. A useful reading respects achievement and its scope instead of manufacturing an unsupported universal scale.
6. From table to choice: a bounded decision protocol
Do not separate the sire from dams, training and opportunity. The observed record concerns offspring from particular mares, managed in particular places and entered in selected races. It is not an experiment randomly distributing dams and resources between stallions.
Say that the measure describes the sire’s progeny in the observed group, not that it has isolated his genetic effect. Genetics recognises interaction between genes and environment, but that general principle supplies no ready-made coefficient of Arabian racing success. Winners to runners is neither a “winning gene transmission rate” nor an individual guarantee.
Begin with a written objective before opening the ranking: target racing age, distances, surfaces and country, within Arabian racing only. Then assemble verified sire, mare and family records. Do not choose the highest figure and redefine the objective to fit it afterwards.
Include non-statistical requirements such as identity, reproductive health and practical suitability, using the appropriate professionals. An excellent statistical record can still be insufficient when the mare’s information, access arrangements, transport requirements or costs are unclear. The table supports a breeding decision; it does not replace the whole decision process.
At the second stage, align period, unit and scope, presenting distinct runners, winners, wins and starts separately. Add age crops, achievement levels and currency-specific earnings while retaining missing-data states. Then examine sensitivity: does the impression change when looking at winner count instead of percentage, isolating the leading offspring’s contribution or restricting to an appropriate age group?
These are review questions, not permission to try dozens of filters until a favoured stallion leads. Keep the original comparison plan so the conclusion remains faithful to the question the team actually set out to answer.
Finally write a bounded, revisable conclusion: why does this candidate appear relevant to this programme, what remains uncertain, and which document or professional assessment is needed next? Preserve date, sources and the compared version, then revisit it when official results arrive or records are corrected.
Do not silently overwrite history or present hindsight as a prior prediction. The protocol does not promise a champion. Its immediate value is preventing identifiable errors: duplicated counts, mismatched denominators, pooled currencies or a descriptive percentage presented as genetic fact. Avoiding those errors is useful before any future foal races.
Aletaq Golden AdviceNever retain the percentage alone. Keep its numerator, denominator, period, country and verification date, then ask about age, opportunity and the leading progeny’s contribution. If that information does not let you reconstruct the measure, set it aside until its definition is clear. An elegant table is not sufficient grounds for a breeding decision.
Closing perspective
Sire figures become useful when they describe a defined group and preserve its limits. Winner count shows breadth, winner percentage needs a clear denominator, wins describe repeated achievement and starts describe racing exposure. Earnings add a financial dimension within a documented currency and scope without becoming net profit or a direct genetic measurement. Reading these layers together is more valuable than asking one number to make the entire decision.
Choose one table this week and check every field definition. Can you distinguish runners from starts, establish age and period, identify the source and currency of earnings and state what is missing? Where a gap remains, make a request for evidence the next step. Which measure changed your impression most once its denominator was visible? Discuss that question with the breeding team and subscribe to Aletaq’s newsletter for practical readings of records and pedigrees that respect achievement while separating verified fact from analysis. Stallion choice still combines mare, programme, welfare and resources. Better questions improve it more than a larger collection of unexplained percentages.
Subscribe to Aletaq’s newsletter for new reference guides.Sources and references
- Arabian Jockey Club — Sire statistical reports and record definitions
- IFAHR — Arabian racing statistics and date scope
- IFAHR — Official Arabian racing results
- ARO — Championship tables and leading sires
- ARO — Official results archive
- ARO — Arabian race categories and racing framework
- AFAC — French Arabian racing programme
- BHA — Performance figures and race context
- NHGRI — Gene–environment interaction: general concept
- France Galop — Statistical themes in the Baromètre du Galop