How Strong Is Home Advantage? Research

Of the 2,481 Tests played with a home side on the card, the hosts won 1,046, the visitors won 659 and 776 ended drawn or tied, which is 31.3% of them. Set the draws aside and the 1,705 decided Tests split 61.3% to the home side and 38.7% to the tourists.

The one-day game tells the same story with almost none of the draws. Of 3,485 ODIs with a home side, 1,959 went to the hosts and 1,356 to the visitors, with 170 tied or abandoned, only 4.9% of the total. The decided 3,315 split 59.1% to 40.9%.

Two formats, two very different eras of scheduling, and a gap of roughly 23 points in Tests and 18 in ODIs. That is a different order of thing from the coin. When we looked at whether winning the toss wins Tests, calling correctly turned out to be worth a few points on a decided match. Home is worth several times that, and it applies to every ball.

With samples this large the arithmetic is not in doubt. A plausible range for the true Test figure runs from 59.0% to 63.7%, and for ODIs from 57.4% to 60.8%. Neither range comes close to an even 50%, and in both formats the chance of a split this uneven arising from a balanced process is below one in ten thousand.

What the arithmetic cannot do is tell you why. A coin toss is random, so the sides that win it are, as a group, no better than the sides that lose it. Which team is at home is not random at all. A host country appears in its own home column with whatever squad it happened to have in that decade, so a strong home record mixes two things that this data cannot pull apart: the effect of the venue, and how good that team was at the time. Read every host figure below as the combination, never as the venue alone.

One check is worth running before trusting any of it. Some matches have no home side at all: a World Cup fixture between two visiting nations, or a series relocated to a third country. The archive still lists one team first on the card, and if nothing but the venue were driving our home number, that side should win about half of the decided ones.

In ODIs it does. Of 1,496 neutral-venue matches, 1,447 were decided and the side listed first won 741 of them, 51.2%. That is close enough to half to say the card order carries next to no signal, which in turn says the 59.1% home figure is not an artefact of how matches are recorded. The Test check is far weaker: only 56 Tests in the archive count as neutral by this rule, among them Pakistan's home Tests at Dacca before 1971, a ground now in Bangladesh, and 42 of them were decided, which produced 57.1% for the side listed first. On 42 matches that number would swing several points if two results flipped, so it neither supports nor undermines anything.

Be clear about what the check does not do. It shows the bookkeeping is sound; it does not control for team strength, and neutral fixtures are not a random sample of cricket, since they cluster in tournaments and in the years one board could not stage matches at home.

Most people expect home advantage to be shrinking. Pitches are better prepared, squads fly in earlier, video of every ground is a click away, and neutral umpires have stood in Tests for decades. The record says the opposite.

The last row is the strongest in the table: 2010 onwards has the highest home share in the whole record, at 64.3%. The first row, 1877 to 1914, sits at 54.6%, low enough that on its sample a split that lopsided would turn up by chance roughly once in 3 tries. In the earliest years the home edge is not distinguishable from noise. The modern edge is unmistakable.

By decade the picture is the same. Among decades with at least 150 Tests, the 2010s are the peak at 65.8% and the 1950s the trough at 57.7%. The chart below also shows why the older figures were hard to read: in the 1960s, 47% of Tests with a home side were drawn, so a home edge had far fewer finishes in which to show itself.

The ODI arc is less tidy but ends in the same place. Home sides took 58.0% of decided matches in 1971 to 1989, rose to 60.2% across 1990 to 1999, dipped to 56.3% in 2000 to 2009, and have run at 60.9% since 2010. The 2020s are the strongest decade in the one-day record at 63.3%, on 538 matches. Whatever has flattened out around world cricket, it has not flattened this.

The average hides an enormous spread. Australia is the hardest place to tour in the Test record: home sides there have won 71.7% of decided Tests across 445 matches. At the other end, home sides in Zimbabwe have won 19.6% of theirs. In ODIs the toughest destination is South Africa, where the hosts have taken 70.7% of 292 matches.

That last pair of columns is where the story is. India has gone from 54.5% at home before 2000 to 82.4% since, a rise of 27.9 points on 114 modern Tests. New Zealand has climbed from 38.6% to 64.9%. Of the 9 hosts with Tests on both sides of the line, 6 are higher now than they were and 3 are lower, with almost all of the downward movement belonging to two of them.

Keep the warning in view while reading it: India winning 82.4% at home since 2000 is partly about conditions and partly about how strong India has been in those years, and nothing here separates the two.

West Indies is the clearest reversal in the record. Before 2000 the home side won 70.0% of decided Tests there across 157 matches, well clear of the 61.3% all-time average for home sides. Since 2000, across 106 Tests, that has fallen to 42.5%, a drop of 27.5 points and the largest reversal of any host in the table. The one-day record moves the same way: West Indies hosts have gone from 73.4% to 47.0% across the same divide.

Pakistan is the second faller, from 75.9% to 60.0%, and that one comes with a caveat built into the method. Matches are assigned to the country of the ground, so Tests a board stages away from its own soil count as neutral and never enter its home row at all. Pakistan has only 45 Tests in its modern home sample for that reason, which is thin enough that a couple of series moves the figure noticeably.

A collapsing home record is at least as likely to be a story about a declining side as about a changing pitch. A team that stops winning anywhere stops winning at home too.

Turn the question around and ask it of the teams rather than the countries. The gap between a side's win share at home and its win share away is the cleanest single measure of how much it depends on its own conditions.

India sits at the top of the Test list: 68.2% at home across 289 Tests against 33.1% across 287 away, a gap of 35.1 points. In ODIs the widest gap belongs to New Zealand, 63.4% at home against 33.9% away. The narrowest Test gap is Bangladesh's 10.8, but that is the flat kind of narrow: Bangladesh has won 20.6% at home and 9.8% away, so there is simply less room between the two numbers.

The rarest line in the table is a side that wins more than it loses on the road. Only one team manages it in Tests: Australia, on 53.8% from 407 away matches, and even they win 71.7% at home. In ODIs, 2 sides clear half away from home: South Africa and Australia. For everyone else, touring means losing more often than winning, whatever the quality of the squad.

Three things come out of the numbers. First, home advantage is the largest single fact anybody can measure from a scorecard: 61.3% in Tests and 59.1% in ODIs, on samples that leave no room for doubt. Second, it has not been worn away by modern preparation: the most recent Test era in our record, 2010 onwards, is also the strongest for the home side at 64.3%. Third, the spread between hosts is wider than the spread between eras, so where a series is played tells you more than when.

The caution stays attached to all of it. Home and away records describe teams as much as they describe grounds, and this data cannot say which part is the venue and which part is the side. A touring captain does not much care about the distinction, but a reader should.

You feel the same pressure in The Follow-On, where a series abroad is a different job from one at home, and in The Chase, its one-day game. Our guide to how a Test pitch plays covers what changes under a touring batter, and the eras of Test cricket puts the decades above in context.

• The Test archive ends on 2024-03-30 and the ODI archive on 2026-08-13; later matches are not in any figure here, and the two formats therefore do not cover the same window.
• Men's international cricket only. Nothing on this page says anything about the women's game, or about domestic and franchise cricket.
• "Home" is decided by the venue country matching a team name, so a board staging its home series abroad has those matches counted as neutral rather than home. That removes them from the host row entirely and thins the modern sample for the affected countries. Pakistan's Tests at Dacca, played before Bangladesh existed, count as neutral for the same reason.
• Draws, ties and no-results are excluded from every win share. In Test cricket that is 31.3% of matches, and a home edge that converts a likely defeat into a draw does not appear anywhere in these numbers.
• Some samples are small: only 42 decided Tests at neutral venues, and hosts near the bottom of the table have few enough matches that a single series moves their percentage by several points.
• Comparing groups is not the same as explaining them. A host record blends the venue with the strength of the side that happened to be playing there in that period, and no split in this dataset separates the two.
• Each p-value on this page is read on its own. 71 subgroups are examined here, by era, by decade, by host and by team, with no adjustment for how many looks that is, so at the usual one-in-twenty bar about 4 of them would be expected to look convincing by luck alone. The two headline figures rest on the full archive and are far past that bar; the borderline ones, such as the earliest Test era, are exactly what this caveat is about.
• The eras are contiguous, so every match with a home side falls in exactly one of them and the rows add to 2,481 Tests and 3,485 ODIs. An earlier version left a gap between the 1877 to 1914 row and the one after it, which quietly dropped the three Tests of the 1913-14 tour of South Africa that began in 1914. The script now fails rather than publish era rows that do not add up.

Figures computed from CricArcade's match archive with our own analysis scripts. Drafted with AI assistance, checked against the scorecards and edited by Vishal Jaiswal. Found a number that looks wrong? Tell us and we will check it against the scorecard.

• CricArcade Test match archive: 2,537 Tests, 1877 to 2024
• CricArcade ODI match archive: 4,981 ODIs, 1971 to 2026

More research: Does Winning the Toss Win Tests?, Batting First or Second?, How Often Is 300 Chased Down?, The Anatomy of a Collapse.