Skip to content
Stadiuminsight
Stadiuminsight

Stadium capacity explained

A stadium capacity looks like a fact and behaves like a measurement. Two sources can print different numbers for the same ground and both be correct, because they are counting different things — and the commonest mistake in reading them is assuming there is one right answer to find.

A ground does not have one capacity

It has several, and which one gets published depends on who is publishing and why. A club selling advertising quotes the audience. A safety authority licenses a maximum. An architect states a design figure. A ticketing system knows how many seats exist. These are different quantities about the same building, and they diverge by hundreds or thousands of places.

Wolverhampton Wanderers state "31,700 Capacity Stadium" on an advertising facts-and-figures page — a rate card selling an audience, not a licensed total. Brentford describe the Gtech Community Stadium as holding 17,250 fans rather than seats. Manchester United publish four different figures across their own material, and the one this site uses is the one filed with a financial regulator. None of those clubs is being dishonest. They are answering different questions.

Seats are not a licensed capacity

The most useful distinction to carry is between how many seats a ground contains and how many people may lawfully be admitted to it. The second is almost always lower, and the gap is made of things a seat count does not know about: segregation lines between home and away supporters, camera positions, media facilities, wheelchair bays and their companion seats, and areas closed for sightlines.

Everton published an unusually clear worked example when the Hill Dickinson Stadium opened. The design capacity was 52,888. After a post-construction audit the official capacity for Premier League fixtures became 52,769 — 119 seats lower, 0.2%, and the club named where they went: segregation lines, and compliance provision for media and camera positions. Most grounds have the same arithmetic and simply never show it.

Shared grounds sharpen the point. The MKM Stadium is described as a 25,586-seat stadium by Hull FC, who share it with Hull City. That is a count of seats. A segregated football fixture at the same ground may be licensed lower, and nobody publishes the licensed figure — so 25,586 is true and is not a matchday maximum.

Changed, or disputed?

When you find two figures, the first question is not which is right. It is whether they describe the same moment.

St James' Park was published at 52,305 for 2025/26 and 52,719 for 2026/27, because Newcastle United added around 500 seats between the two. That is a timeline. Presenting it as a disagreement would invent a controversy out of some building work. The same applies to Craven Cottage, where 25,700 appears on a club history page describing the ground before the Riverside Stand rebuild, and 30,000 is the stated figure for after it — neither competes with the 27,000 the ground currently operates at during the works.

A real conflict looks different: two present-tense claims about the same ground that cannot both be true. Brighton and Hove Albion state that one project — an additional 1901 Club area at the Amex — will take capacity to 32,000 in one club article and to 32,500 in another. One project, two totals, no measurement distinction that explains it. That is a contradiction, and the honest treatment is to show both with their publishers rather than pick one.

How to read a published figure

Ask four things. Who published it — the club, the venue operator, a league, or an aggregator repeating someone else? What did they measure — seats, licensed capacity, matchday capacity, an expanded configuration for one event? When — is it a season figure that has since moved? And does anything on the same page contradict it, which is more common than it sounds.

This site records the concept alongside the number for exactly that reason, and marks whether anyone has opened the cited page and read the figure off it. A correct number with no idea what it measured is still capable of misleading, and a figure nobody has checked is not improved by being repeated confidently.