Test Cricket Stopped Drawing: 847 Matches of Our Own Data
In 2009, more than a third of all Test matches were drawn. In 2024, one in twenty five. We now hold every ball of 847 Tests played since 2005, and the numbers show the draw did not fade slowly. It fell off a cliff, and the reason is not the one most people reach for.
Three Test series start this month: Australia against Bangladesh in the Top End, India in Sri Lanka, England against Pakistan. That is a lot of five day cricket in a sport that keeps being told it has no time for five day cricket.
So we went to our own data to ask a simple question. Is the long format actually dying, or is it just changing shape?
We now hold every ball of 847 Test matches played between January 2005 and late July 2026. Here is what they say.
The draw has almost disappeared
In 2009, 37 per cent of Tests were drawn. Thirty seven. If you sat down to watch a match that year, better than one chance in three said nobody would win it.
In 2024, out of 50 Tests in our data, two were drawn. Four per cent. In 2025, three of 40. Through July 2026, one of eleven.
That is not a gentle decline. Between 2005 and 2015 the draw share sat in a band from about 17 to 37 per cent and wandered around inside it. From 2016 it settled near 15 per cent. Then, in the last three years, it collapsed to a level Test cricket has never sustained before.
Scoring is the visible cause
The obvious suspect is scoring rate, and the obvious suspect is guilty.
Across our data the game got slower for a decade, then reversed hard. Runs per over by year:
2005: 3.47. 2011: 3.15. 2018: 3.11. 2021: 2.99. Then 2022: 3.31. 2023: 3.50. 2024: 3.66. 2025: 3.58. So far in 2026, 3.74, which is the fastest year in our entire record.
Between 2021 and 2026 the game added roughly 0.75 runs per over. Over a 90 over day that is about 67 extra runs, which is most of a session. Add a session's worth of runs to each day and a match that used to run out of time now finishes.
But the interesting part is where those runs come from
Here is where the data contradicts the usual story. Everyone assumes modern Test batting means more boundaries, more sixes, bigger hitting.
Sixes have gone up, and dramatically. Our 2005 data has 207 sixes hit in Tests all year. In 2024 there were 622.
Yet the share of total runs that came from boundaries has barely moved. In 2005, 51.1 per cent of all Test runs in our data came in fours and sixes. In 2024 it was 49.2 per cent, and in 2025 48.0 per cent. If anything, slightly less.
The number that actually changed is the dot ball. In 2021, 75.6 per cent of legal deliveries in Test cricket produced no run. By 2024 that was 69.9 per cent, and in 2026 so far 69.3 per cent.
Six percentage points sounds small. Across an innings of 90 overs, it is about 32 extra scoring shots. Batters are not mainly hitting more boundaries. They are refusing to let the ball go by.
Wickets fall faster too
The other half of a result is the bowling side taking twenty wickets, and that has got easier.
In 2009, a wicket fell every 67.1 balls in our data. In 2018 that was 53.3. In 2024 it was 46.9, the lowest of any year we hold, and 2025 and 2026 sit near 50.
Roughly three overs faster per wicket, applied twenty times, changes what is possible inside five days. Combine it with a session of extra runs a day and the draw stops being a destination that teams can reach by accident.
Was 2024 a fluke?
Fair question, and the honest answer is that one year proves nothing. Two Tests drawn out of fifty could be a run of results-friendly pitches, an unusually thin set of fixtures, or genuine change.
What makes us think it is more than noise is that the supporting numbers all moved together and all stayed moved. Run rate up, dots down, balls per wicket down, sixes up, and 2025 and 2026 have not gone back. When five independent measures shift in the same direction and hold, that is a trend rather than a bounce.
What this means for August
Two Tests in Galle and Colombo, two in Darwin and Mackay, three in England. In the cricket our data describes, the natural expectation for each of them is a result rather than a stalemate, and a result arriving inside four days more often than not.
That is a genuinely different sport from the one that produced 37 per cent draws in 2009, and it happened inside the career of players still in these squads.
What our data covers, and what it does not
Our Test records begin in January 2005, so nothing here speaks to the Test cricket of the 1990s or earlier, when draws were more common still. Anyone quoting us on Bradman is not quoting us.
Matches involving Afghanistan's men's team are absent from our Test and one day records, because the public ball-by-ball archive we build on withholds them. That is 13 Tests missing, and we would rather say so than let you assume completeness.
The 2026 figure covers 11 Tests to 25 July and will move as the year finishes. We publish it labelled rather than waiting, because a partial year with its label on is more useful than silence.
The bottom line
Test cricket in our data is faster than it has ever been, wickets fall sooner, batters leave fewer balls alone, and the draw has fallen from more than a third of matches to something close to a rarity.
The format is not dying of boredom. It has quietly become the most decisive version of itself we have ever measured.
All figures computed by CricNerd from our own ball-by-ball Test records, January 2005 to 25 July 2026.