Sports intelligence · Tennis · Model governance

When the market wins: governing a WTA model

A model can be technically sound and still not deserve release. This is the story of building a leak-safe tennis evaluation, learning what it could honestly support, and choosing evidence over launch theatre.

Sports modellingLeakage controlEvaluationData leadership
Read the decisionInspect the evidence
19,780Matches
61Chronological folds
0Odds inputs

· Updated with the June 2026 evaluation.

Decision

The right call: do not ship yet.

The research question was simple in public language: could match history produce a useful pre-match ranking without borrowing information from the market or the future? The evaluation answered a narrower question well. It did not justify pretending that two different evaluation populations were directly comparable.

Protect time

Every prediction was generated after training only on earlier information. Random splits would have made the model look at a world it could never have known.

Protect meaning

The market baseline and model card covered different populations. Presenting them as a clean head-to-head result would have overstated the evidence.

Protect the decision

The evidence supported further research and a shared-cohort test—not a public performance claim or an operating launch.

Chronological evaluation

Recreate the decision in time.

The June evaluation expanded each match into orientation examples, kept time order intact and produced out-of-fold predictions. “Out of fold” matters because no row helps fit its own prediction; “chronological” matters because later matches cannot teach an earlier decision.

Flow chart: 19,780 matches become 158,240 orientation examples and 156,008 out-of-fold predictions across 61 chronological folds.
The evaluation is large enough to test behaviour across time while keeping the information boundary explicit.
Data behind the evaluation-scale chart
StagePublic countWhy it matters
Source matches19,780The event-time unit used to order the evaluation.
Orientation examples158,240Both player orderings support symmetry checks and paired handling.
Out-of-fold predictions156,008Predictions generated without fitting on their own rows.
Chronological folds61Repeated forward-looking windows rather than a random split.

Information boundary

Show what the model could—and could not—know.

The model card covered 275 engineered features and explicitly recorded zero odds-derived inputs. Reversing player orientation produced a complement error of roughly 1.2 × 10−7, a useful integrity check that the representation behaved consistently.

Model governance chart showing 275 engineered features, zero odds inputs, an orientation complement error near 1.2 times ten to the minus seven, and separate baseline and model cards.
Good governance separates input integrity from performance claims, then labels comparisons that do not share a cohort.
Long description of the information-boundary chart
CardEvidenceInterpretation
Model inputs275 engineered features; 0 odds inputsThe research signal is separated from the market comparison.
SymmetryComplement error ≈ 1.2 × 10−7Paired orientations behave consistently to numerical precision.
Honest baselineBroader comparison populationUseful context, but not a like-for-like scorecard.
Model cardChronological out-of-fold populationNeeds a shared-cohort rerun before a direct ranking.

Leadership takeaway

Stopping is a delivery milestone.

Data leadership is not measured by how many models cross a production line. It is measured by whether the organisation can distinguish promising research from dependable evidence before exposure grows.

  • Write the comparison contract first. Decide the shared cohort, timestamp and metric before seeing results.
  • Make leakage controls part of the product. Time ordering, paired orientations and input provenance should be repeatable checks, not notebook footnotes.
  • Keep “interesting” separate from “actionable”. The model earned another evaluation cycle; it did not earn an operational release.
  • Reward truthful stopping. A team that can surface an evidence gap early saves time, reputation and downstream risk.

Limits

What this evidence cannot claim.

The aggregates describe evaluation design and integrity checks. They do not establish current predictive quality, profitability, suitability for wagering, or superiority over a market. The honest next test is a prospectively defined shared cohort, followed—only if warranted—by shadow operation and explicit promotion criteria.