Break Points Saved and Holds: Different Denominators

Break points saved and holds are not the same statistic

Break points saved asks what share of an opponent’s break-point chances did not end with the server losing the game. Its denominator is the break points faced by the server. Hold rate asks what share of the player’s service games they won. Its denominator is service games played. Both describe serve defence, but they track different events and different numbers of opportunities.

An example with visible denominators

A player faces ten break points and saves seven: 7 / 10 = 70% saved. Across those matches the player holds twelve of fifteen service games: 12 / 15 = 80%; they lose the other three. The example is consistent: several break points can occur in one lost game, while some service games contain none. You cannot conclude that they “saved 80% of break points”: one service game may contain several break points, and some service games contain none. Several break points can also occur in the same game. If a server saves two break points but loses the game on a third, that is two saved chances out of three and zero holds out of one service game. The numerators describe different events.

In a second hypothetical example, a player holds all eight service games but faces only one break point and saves it: 100% holds, 100% break points saved. The percentages look identical, yet one is based on eight games and the other on a single point under threat of a break. Show the denominator, not just a rounded percentage.

Keep server and returner perspectives separate. “Break points saved” is the server-side metric: the server prevents the opponent from taking the game. “Break-point conversion” is the returner’s metric: how many of their opportunities they turn into a game win. The existing guide to break-point conversion on return covers the opposite side of this pair. Our serve and return overview treats hold and break separately.

Why percentages can move

The number of break points depends on the opponent’s return quality, surface, format, and individual games. One long game can produce several break points and affect a small sample. Do not add an ordinary tie-break as an extra break point or ordinary service game in these metrics; check the statistic definition and treat tie-breaks separately. When comparing tournaments, check how the source records a break point and handles retirements. Use the same time window and event rules before drawing a conclusion.

How to use the pair

For service hold, read both metrics together: the overall result of service games and how often the player preserved serve under pressure. This describes a pattern; it does not make one percentage a reliable forecast of the next match. Before considering a games or breaks market, also check the expected number of opportunities, line-up, player condition, and format. A high rate in a small sample does not promise continuation.

The calculation protects against a common mistake: first ask what the numerator counts, then identify the opportunities in its denominator. Only then compare the context and sample size.

Read also

• Match Tie-Breaks: Checking Games, Sets, and Handicap Settlement