Hub City Boxing
fight history·Article 7

Reading a fight through its published statistics

Punch stats are widely quoted and often misread. What thrown-versus-landed can show, what one 108-bout study found and how it was coded, and why a landed count is not the scoring-blow count judges use.

By Hub City Boxing Notes Editorial Team·Published 2026-11-22

Punch statistics, “thrown” against “landed”, are among the most quoted numbers in boxing and among the most easily misread. This article does not quote statistics from any particular fight. It uses one openly described research study of amateur bouts, the amateur scoring rules, and an invented example to show what these numbers can and cannot support.

What a thrown and landed count records

Thrown is the number of punches a boxer throws. Landed is the number of those a counter judges to have reached the opponent. Accuracy is landed divided by thrown, expressed as a percentage. Every one of those figures is a human or coded judgement about what happened on screen or at ringside, not a sensor reading. We have not opened the counting method of any broadcast statistics provider, so this article does not describe how those services count.

What one 2026 study did

A 2026 paper in PLOS ONE analysed 108 international elite amateur bouts from 2019 to 2025, giving 216 boxer observations, using official competition videos and official results. The coded indicators were thrown and landed totals, accuracy, counterattacks that succeeded, how often defence worked, control of the centre of the ring, and warnings. Because each bout has one winner and one loser, the authors compared winners with losers inside each bout rather than across all boxers.

Reported results include:

Indicator Winners Losers
Mean punching accuracy 30.41% 26.84%

The difference is 3.57 percentage points (30.41 minus 26.84, which matches the 3.57 in the abstract). In the study’s model, punching accuracy, defensive success rate, ring-centre control and total punches thrown were each independently associated with winning; successful counterattacks were not independently significant after adjustment. The study reported that its model discriminated winners from losers well in its own data.

Limits to keep in mind: this is an association in coded video of elite amateur bouts, not proof that raising accuracy wins bouts, and it does not describe professional bouts or any specific boxer. We read the abstract only, so we do not know how the paper defines each indicator in detail.

Why a landed count is not the judges’ count

Under the World Boxing scoring rules, judges count scoring blows, and a blow only counts if it connects cleanly with the target area, has the weight of the body or shoulder behind it, connects with the knuckle surface of the glove, breaks no rule, and is clearly seen by the judge (Rule 7.2.2). A published landed count is not necessarily built on that definition. A light touch to the gloves, a glancing blow, or a blow to a non-scoring area may or may not appear in a landed figure, depending on how the counter coded it. So a landed total is useful context, but it is not a score.

An invented example

The numbers below are invented for this article and describe no real bout. We computed the percentages with a short script.

Thrown Landed Accuracy Landed per round
Boxer A 67 15 22.4% 5, 6, 4
Boxer B 37 16 43.2% 5, 5, 6

What the table shows:

  • Boxer A threw nearly twice as many punches (67 against 37) but landed one fewer (15 against 16).
  • Accuracy is far apart (22.4% against 43.2%) while the landed counts are almost the same. If a commentator quotes accuracy alone, it suggests a large gap in performance that the landed counts do not show.
  • Two boxers with identical accuracy can look very different: 18 landed from 60 thrown and 9 landed from 30 thrown are both 30.0%.
  • Round by round, boxer B leads only in round three (6 against 4); a whole-fight total hides that the first two rounds were level or slightly in A’s favour.

A short checklist for reading any punch statistic (our suggestion)

  1. Ask who counted, and how a punch is defined as landed.
  2. Compare thrown, landed and accuracy together, not accuracy alone.
  3. Look for round-by-round figures, not just totals.
  4. Treat a statistic as context for what you saw, not as the verdict; the decision in an amateur bout comes from judges’ scorecards under the scoring rules.
  5. Do not use a number you cannot trace to a source.

How to check this article

The accuracy arithmetic is landed divided by thrown; you can repeat every percentage in the table with a calculator. The study figures can be checked against the abstract at the PubMed link below.

Sources


Sources

AI assistance recorded as: AI draft, not edited