Reading a tennis match market
Tennis prices are built from serve and return points, not from results. Once you know that, the whole board becomes readable.
A tennis match is thousands of points, and the market prices it as exactly that. Two inputs do almost all the work: the probability each player wins a point on their own serve, and the probability each wins one on return. Run those forward through the scoring system and every market in the fixture falls out of the same distribution.
Why the scoring system amplifies everything
Move a player's serve-point probability from 65% to 68% and their game-hold probability moves much further, because a game is a race that the server keeps restarting from an advantage. Move the hold probability and the set moves further still. By the time it reaches the match price, a three point input change has become a large price change.
This is why tennis prices look volatile relative to the information moving them. They are not overreacting. The scoring system really does turn small edges into large ones, and it does so in both directions.
Where the board goes loose
The match odds at a major book are priced properly. The derivatives underneath often are not: set betting, total games, and handicaps are frequently derived from the headline price with a rule of thumb rather than modelled from the same point distribution. That inconsistency is where a market disagrees with itself, and a market disagreeing with itself is more interesting than a market disagreeing with an opinion.
What moves the inputs
Surface is the biggest single factor, and it is worth its own piece. After that: recent injury, altitude, ball type, and whether a best of five is being played by somebody whose fitness is in question.
What this does not tell you
A point model assumes points are independent and identically distributed, and they are neither. Players tighten at 4-5, serve differently at 30-40, and collapse in ways no serve percentage predicts. The model is a good approximation that is systematically wrong at exactly the moments that decide matches.
Common questions
How are tennis match odds calculated?
From point level inputs. The probability each player wins a point on serve is run forward through the scoring system, which produces game, set and match probabilities from the same two numbers.
Why do small changes in serve produce big changes in the price?
Because the scoring system compounds them. A three point change in serve probability is amplified across dozens of games and can move a match price substantially.
Are set betting markets priced off the match odds?
Usually the other way round. Both come from the same point model, but at many books the derivatives are priced off the headline number rather than modelled directly, which is where they get loose.