MARKET REPORT · TENNIS
Tennis Margins: One Sample and the Cleanest De-Vig Case
We have exactly one tennis row so far, and we're telling you that up front. What tennis lacks in sample size it makes up for as a teaching market: a 2-way match-winner is the simplest place to learn how margin actually works.
Our only tennis sample to date: Cloudbet, match-winner, in-play, priced 2.08 / 1.74 = 5.55% overround LIVE (8 July 2026). n=1 — far below our n≥6 bar. No median, no ranking, no verdict. Just the first datapoint.
The sample, flaws included
| Book | Market | Odds | Overround |
|---|---|---|---|
| Cloudbet | Match-winner LIVE | 2.08 / 1.74 | 5.55% |
Captured 8 July 2026. The player names were not recorded in round 1 — a logged flaw, and our methodology now requires full event identification on every row. We publish the mistake because pretending the row is cleaner than it is would defeat the point of this site.
Tennis as the cleanest de-vig case
A tennis match-winner has two outcomes and no draw. That makes it the one market where the multiplicative method — divide each implied probability by the total — is both the default and usually the right call. Here is the canonical worked example, using the standard US -110/-110 line (BettingUSA benchmark, updated June 2026):
-110 implies 52.38% per side.
52.38% + 52.38% = 104.76% total book.
Overround = 104.76% − 100% = 4.76%.
Fair probability per side = 52.38 / 104.76 = 50%.
Fair price = even money (+100).
That's the whole trick. The book quotes both players as if each wins 52.38% of the time; probabilities can't sum past 100%, so the extra 4.76% is the fee. Divide out the total and the fair coin flip reappears. On our Cloudbet row the same arithmetic applies — 1/2.08 plus 1/1.74 sums to 105.55%, and dividing each side by that total gives the fair probabilities. For 3-way markets like football the even-split assumption breaks down and you need Shin or the power method; the de-vig guide covers all five methods with when-to-use rules.
Live vs prematch: read the badge
Our one row is in-play, and that matters. Live prices structurally carry more margin than prematch — the book is defending itself against score latency and fast-moving match states, and the US benchmark treats up to about 7% as typical for in-play and props versus 4.76% for standard prematch pricing (BettingUSA, June 2026). So a live 5.55% is not the same signal as a prematch 5.55% would be, and we will never mix the two states in one median. Every live row on this site wears the badge.
When does tennis get real data? Medians unlock at n≥6 per book, per market state. Tennis rounds are accumulating now; watch the counters fill on the history page.
Where to go from here
If you want data with actual sample depth today, the football 1X2 comparison has same-fixture rows across three books. For the full sampling rules — basket criteria, logged-out capture, live badging, the n≥6 bar — read how we measure. And if you're new to the site, the front page explains the premise in one line: margins are measured, not promised.