03 / PROFESSIONAL CASE STUDY · TENNIS ANALYTICS
Using Behavioral Analytics to Inform Tennis Trading Strategy
SQL · Apache Superset · Excel · Commercial Strategy
01 / The Challenge
When the results raised questions
During the Australian Open, the company's ATP tennis trading results revealed significant losses in ace markets, particularly ace parlays. Although the losses were concentrated around betting activity associated with one client, the underlying pricing concerns had implications across the business.
The company offered ATP tennis markets exclusively during Grand Slam tournaments, limiting opportunities to develop familiarity with individual players and the conditions associated with best of five set matches.
Ahead of the French Open, I undertook an investigation to understand the patterns behind the losses and identify whether the existing pricing approach warranted closer attention.
02 / The Investigation
Looking beyond the losses
Using SQL queries in Apache Superset, I extracted the relevant betting data and established a sample of more than 110 selections based on betting activity, turnover, and financial outcomes.
I then used Excel, including PivotTables and charts, to examine recurring bet selections and how betting activity was distributed across tournament rounds.
The investigation explored player characteristics and matchups, playing styles, match duration, and differences between offered ace lines and actual results.
I also examined profitable selections that met the same inclusion criteria, comparing contrasting outcomes rather than focusing exclusively on losses.
03 / The Findings
When playing styles met pricing decisions
The most revealing patterns emerged from the relationship between player characteristics, match conditions, and the ace lines being offered.
Serving ability and pricing expectations
Players known for powerful serving presented a particular challenge. When ace lines remained close to model or market expectations, selections on higher ace totals frequently contributed to losses. When lines had been raised to account for anticipated betting behaviour, outcomes were often more favourable.
Playing styles and match dynamics
The interaction between players, including serving tempo, point duration, rally length, and the frequency of deuce games, helped distinguish conditions associated with higher ace totals from those where fewer aces occurred. Contrasting patterns were evident across profitable and unprofitable selections.
Match duration was another consistent theme, with longer matches generally producing more aces. Losses were also concentrated in particular tournament rounds.
Together, these findings suggested that effective pricing required consideration of playing styles, match conditions, and anticipated betting activity alongside individual player expectations.
04 / The Application
Turning analysis into trading decisions
The findings were documented and shared with the trading manager and wider team, helping inform changes to trading practices across clients ahead of the French Open.
A more cautious approach
The team adopted a more conservative trading strategy during tournament rounds where the analysis had identified greater historical exposure.
More deliberate pricing adjustments
Player-specific ace lines were adjusted to better reflect serving characteristics, matchups, and anticipated betting behaviour, rather than relying exclusively on model expectations.
Additional changes included withdrawing overnight ace markets when they could not be actively monitored and providing live traders with greater awareness of players and match conditions requiring closer attention.
05 / The Outcome
From investigation to improvement
The analysis helped inform changes to trading practices introduced across the business ahead of the French Open.
Australian Open
−2.07%
Company tennis trading margin
French Open
+2.24%
Company tennis trading margin
The French Open became the most profitable Grand Slam tournament in the company's history at that time.
This case study presents a summary of the analytical approach and business outcomes. Proprietary data and specific trading recommendations have been omitted.