## How Tennis Totals Work
Tennis totals markets are offered on:
- **Total games:** Combined games played across all sets (e.g. Over/Under 22.5 games)
- **Total sets:** Whether the match goes to 2 sets or 3 (best of 3) / 3, 4, or 5 sets (best of 5)
- **Set score:** Exact set count (e.g. 2-0, 2-1 in a best-of-3)
## The Modelling Approach
Tennis can be modelled from the point level up using a Markov chain:
1. Estimate P(server wins point) for each player on each surface
2. Derive P(server wins game) from the point probability (using game-level Markov chain)
3. Derive P(player wins set) from game probabilities
4. Derive P(match outcome) from set probabilities
5. Derive expected total games from the probability distribution over all possible score lines
## Key Drivers of Tennis Totals
- **P(server wins point):** The dominant driver. High server dominance → more tiebreaks, more games, higher totals.
- **Surface:** Clay courts produce more baseline rallies (longer games, more games per set) than grass courts (more serve dominance, shorter rallies, fewer total games).
- **Player style:** Aggressive baseliners vs net rushers vs defensive counterpunchers all produce different game counts.
## The Bagel and Breadstick Effect
A "bagel" (6-0 set) is the minimum possible set score — 6 games. A "breadstick" (6-1) is 7 games. Dominant players produce more of these compressed sets, lowering the total games. Models that do not account for dominant player service games will overestimate total games.
Create a free account to track your progress and save bookmarks.
