A 1–1 draw rarely attracts attention beyond the two clubs involved. In several large professional-football datasets, however, it ranks among the most frequent individual final scores. Published analyses of large match datasets have recorded 1–1 in roughly 12% of fixtures, a proportion that recurs across independent studies. These figures reflect a mathematical characteristic of football: when both teams have modest scoring expectations, one goal apiece often carries a higher probability than any other exact result.
Which football scores occur most frequently?
In these datasets, 1–0 is typically the second most frequent result, at around 10% of matches, followed by 2–1, 0–1 and 0–0, each accounting for between roughly 7% and 9% of fixtures.
The rankings depend on how scores are classified. Combining 1–0 and 0–1 produces more single-goal victories than 1–1 draws. Yet the two winning scores remain separate outcomes because the home team’s goals are recorded first.
This classification also appears in contexts where a correct-score bet is presented as a distinct final-result category. That classification does not establish the historical frequency or likelihood of any score in a particular match.
The published datasets cover different competitions and periods. Their percentages cannot be extrapolated directly to every league, where scoring rates, competitive balance and match characteristics vary.
Why scoring once is often more probable than scoring twice
The mathematical explanation begins with the Poisson distribution, a probability model used to describe the number of events occurring within a defined interval. In football analysis, the event is a goal, while the model’s central parameter, lambda, represents a team’s expected number of goals.
Suppose a team has a scoring expectation of 1.5 goals. Applying the Poisson formula gives a 22.31% probability of scoring none, 33.47% of scoring once and 25.10% of scoring twice. Three goals account for another 12.55%.
One goal is therefore the most probable individual total, despite the expectation being 1.5. The average describes the centre of a probability distribution; it is not a prediction of the actual score.
The relationship has a precise mathematical basis. When the expected number of goals falls strictly between one and two, scoring once has a higher probability than any other individual total under Poisson. Below an expectation of one, a goalless performance becomes more probable.
This concentration of probability around small goal totals helps account for the prominence of low scores in historical football records.
How two scoring distributions produce 1–1
To estimate an exact score, the simplest football model assigns separate Poisson distributions to each team. It assumes their goal totals are independent and multiplies the relevant probabilities.
Consider an illustrative match in which the home team’s expected goals are 1.5 and the visitors’ are 1.2. The resulting distribution places several familiar scores close together:
Both teams have individual scoring distributions that peak at one goal. Their combination makes 1–1 the highest-probability exact score, narrowly ahead of a home victory by 1–0.
A different calculation produces a 44.15% probability of a home win and 25.48% for a draw. There is no inconsistency: a home victory includes every score in which the hosts finish ahead, whereas 1–1 represents only one possibility.
An individual score can lead the rankings without its broader result category being the most probable.
Why the most common score changes between fixtures
Poisson does not identify 1–1 as the leading score under every set of conditions.
With expected goals of 0.7 for each team, 0–0 leads the distribution. Expectations of 1.5 for the hosts and 0.7 for the visitors favour 1–0. When each team has an expected-goals value of 2.2, 2–2 becomes the most probable individual result under the independent-Poisson model.
The relationship between the two scoring rates is as consequential as their combined total. A balanced fixture can produce a different probability distribution from one dominated by a considerably stronger attacking side.
Real matches also challenge the assumption that teams score independently. The Dixon–Coles model, introduced in 1997, incorporates adjustments for four low-scoring outcomes: 0–0, 1–0, 0–1 and 1–1. Its purpose is to address patterns that the basic independent-Poisson framework may not represent adequately.
These corrections improve the treatment of statistical dependence but do not identify the tactical causes of any particular draw.
What historical scores cannot establish
Even when 1–1 occupies the leading position, its probability remains relatively small. In the illustrative model, it accounts for 12.10% of possible outcomes, leaving 87.90% distributed among other scores.
Historical frequency and match-specific probability measure different things. The former summarises recorded results across competitions; the latter depends on the scoring expectations and assumptions assigned to a particular fixture.
That uncertainty also applies to gambling markets. Frequent historical outcomes are not guaranteed to recur, and betting, which in Great Britain is restricted to adults aged 18 and over, involves financial risk regardless of the statistical models used to describe them.
The prevalence of 1–1 reflects the way two relatively low scoring rates interact. Its position in the historical rankings is mathematically understandable, but never inevitable.
