The case of the shrinking Test match

Four years ago I presented a comprehensive analysis of the duration of Test matches. Since then, over 150 Tests have been played and it is obvious to everyone that the number of decisive Test matches has gone up and Tests are becoming shorter all the time. No one can forget the Newlands Test early this year, which lasted a little more than a day.

In this article, I will attempt to look at the concept of four-day Tests again, with special reference to the last four years. These analyses are current up to and including the Wellington Test this year between New Zealand and Australia.

In the first table, I will look at the result percentage and average duration of Tests, by period. I have looked at this only from 1980 onwards - the modern Test era, for the purposes of this analysis. For details of earlier times, you can refer to the previous article. In any case, two-thirds of the total number of Tests played are covered.

The average duration of a completed Test in the 1980s was 315 overs. It went up in the next decade and then dropped back to around the 1980s mark. It stands at 295 overs now. That is about three days and a session. It is a clear indication that the average duration of decisive Tests is on the decline. The current period sees a 23-over drop from the average for the five decades considered.

The overs required for the shortest match in each period saw a steady drop from 174 in the 1980s to 107 in the current period. The longest decisive Test tells its own story. At the SCG in 1987, in the Ashes series, Australia and England battled 470 overs in five days to achieve a result. There is no information available as to how many overs remained to be bowled at the end. Even if we add five overs, we are looking at 475 in five days - an average of 95 overs per day. These were not subcontinental spin attacks delivering an over in 150 seconds. The bowlers included Graham Dilley, Gladstone Small, Ian Botham, Merv Hughes and Bruce Reid. Today, with three spinners bowling, teams deliver about 85 overs on extended days.

Let us now look at the result percent values. From a little above 50% in the 1980s, the result percentage has moved up steadily. Each period has seen a significant increase from the previous period. The current period has a staggering 85% decisive results. That is more than five in six Tests. If this trend continues, not far from now we will reach a point when we see nine Tests out of 10 ending decisively. The World Test Championship set-up also contributes to this increase. There are no dead rubbers; the incentive to win and secure WTC points is ever present. The ICC must be complimented for this significant change. Five Tests each in England and the West Indies have been drawn during these four years. No Test in South Africa has been drawn.

The average duration of a completed Test in the current four-year period started at a reasonably healthy 311 overs but has come down to 281 in 2024 (though we are only a little over two months into 2024). That means that a Test now finishes, on average, in just over three days. Anyone booking tickets for the fourth and fifth days of matches is being optimistic. The shortest Test during these four years lasted 107 overs. If that tells a story, the longest, which lasted 442 overs, tells another. However, in 2023-24, the longest decisive Test lasted only 377 overs - that is a little over four days' play.

Forty-seven Tests were played in 2020-2024, of which 41 were conclusive, most not even going into the fifth day. That clearly illustrates the way Test cricket is going: quicker scoring, lack of defensive batting, and an almost never-to-appear fifth day. During these four-plus years, there was a long sequence of 23 decisive Tests in 2022. Currently we are in a 20-match unbroken sequence of decisive Tests.

The result percent during the first Covid-19 year was the highest of the lot. In the current year there have been decisive results in every game. Is there a draw in the near future? We might have to wait for one produced by weather.

The next visual is about the duration of Tests during the periods 1980-2019 and 2020-2024. It uses overs bowled rather than days played as a measure, since a Test could be played on five days but might have only 100 overs because of rain. And I will take the expected value of 90 overs per day, although very few teams achieve this norm. So two days mean 180 overs, three mean 270 overs, and so on.

In the 40-year period between 1980 and 2019, only around 1.8% of Tests finished within two days (less than 180 overs) - that is, about one match every two years. The majority of Tests (45%) finished on the fourth day. Over 30% of Tests went into the fifth day.

Since the start of 2020, the number of Tests finishing in two days (180 overs) has moved up to over 6% - one every six months or so. Most Tests (45%) still finished on the fourth day. However, in a significant change, less than 20% of Tests went to the fifth day, as against over 30% before. This drop seems to have been made up by the rise in the three-day-finish category.

Now for a look at the batting and bowling averages during these periods with a view to determining the reasons for this phenomenon of significantly early finishes to Tests.

Look at how the batting average has dropped this century. It started at around 36 in the 1980s and 1990s, peaked at just over 38 during the first decade of the millennium, dropped in the second decade and then again significantly over the last four years. We are now at 34.8, which is a massive four-run drop from the highs of 2000-2009.

The reason for this steep drop could be a combination of T20-style batting, inability to play long innings, judicious use of the DRS, result-oriented pitches, a decline in "correct" batting technique, and so on.

The pattern for bowling averages is almost identical to that of batting, though where the batting stats only looked at batters Nos 1 through 7, the bowling averages include performances against all batters. Starting at 32 in the 1980s (thanks to the army of quicks that West Indies unleashed), going on to a slight drop in the 90s, a steep increase in the first decade of the millennium, then two successive significant drops. The bowling average for the current period stands at a low 30.4. That corresponds with the drop in batting average. The reasons I outlined to explain the drop in batting average can also be applied here.

A low runs-per-wicket value has led to a situation in which drawn Tests are very rare. Groundsmen are probably instructed to produce result-oriented pitches. Five-day draws are passé , which is good for the game. But think of how many playing days are lost because games finish ahead of their scheduled duration, and the subsequent loss in gate revenues. Or am I barking up the wrong tree? Television revenues are fixed at a certain figure per match, with no reference to the actual playing hours. So maybe the boards, especially the ones that control the game, do not care.

I am not comfortable with the term Bazball. I even coined my own alternative term. However, since Bazball is so widely used, that's what it will be.

England have had a lot of success with the philosophy and there is no denying that it has revolutionised Test cricket. While there is more to Bazball, it is clear that one important aspect of the approach is the increase in scoring rate. It is possible that the overs bowled might be on the lower side but the runs scored in them match the figures in earlier periods since the scoring rate has gone up. Bazball could also be one reason why teams are not looking at draws as a valid alternative.

In my analysis from four years ago, I talked about the concept of four-day Tests extensively and presented many arguments for retaining the five-day set-up. However, given the increasing number of Tests that do not reach the fifth day, maybe it is time to bite the bullet and have four-day Tests. Maybe compensate somewhat by increasing the hours of play to six and a half per day. With the additional 30 minutes, there is a fair chance teams will achieve at least 90 overs per day. Another good idea is to cut the number of free days between Tests. Now teams are finishing Tests in two to three days and sometimes seem to get a week between matches.

Shortest ODIs
In each of my articles, I present a numerical/anecdotal outlier relating to Test and ODI cricket. This time the outlier query is "What are the shortest matches in ODI history?" This is partly influenced by the recent Canberra match - which misses the cut to make the top four, since it took all of 31 overs. The top four values, based on a cut-off of 25 overs, are below.

1. Nepal, with their strong spin attack, dismissed USA for 35 in 12 overs in Kirtipur in 2020 and overhauled that score in 5.2 overs. This match is the shortest on record, with only 17.2 overs needed for a finish. Sandeep Lamichhane took 6 for 16.

2. At the SSC in Colombo in 2001, Sri Lanka dismissed Zimbabwe for 38 in 15.4 overs and rattled off the paltry target in 4.2 overs. The match ended in exactly 20 overs.

3. In the recently concluded Asia Cup, India dismissed Sri Lanka for 50, scored in 15.2 overs. They needed 6.1 overs to reach this target and the match lasted only 21.3 overs.

4. In a 2003 World Cup game in Paarl, Sri Lanka repeated their quick 2001 win, this time against Canada, who scored 36 runs in 18.4 overs. Sri Lanka overhauled that score in 4.4 overs, making it 23.2 overs in all for the game.

Talking Cricket Group
Any reader who wishes to join my general-purpose cricket-ideas-exchange group of this name can email me a request for inclusion, providing their name, place of residence, and what they do.