What history tells us about the length of Test matches

Would Headingley 1981 have been as remarkable if it was a four-day Test? Getty Images

The recent announcement by the ICC that they are looking at reducing the duration of Tests to four days has set the cat among the pigeons. It is clear that the proposal has not pleased many, and there are interesting reactions from various people involved in the game. The reactions could be summed up as being approximately 75% against and 25% for the proposal.

I throw my hat in the ring with this article. In the first part, I will analyse all the data available to us and present the facts as insights to digest. In the second, I will list all the points that do not have to do with the numbers, and suggest possible solutions.

I will look at each of the 2382 Tests that have been played so far from the results and duration aspects and present insights from these two data sets. All the data analyses will be at the macro level, and will not factor in how the individual teams played. The fact that in the past 20 years England have won 25 Tests in three days or fewer should have no bearing on the final decision. Other teams may have quite different numbers. The analysis is current up to the fourth Test between South Africa and England played at the Wanderers last month.

I have split the 147 years of Test cricket into periods, as below. The base analysis is done by period so that we can get a clear handle on the development of Test cricket as the years went by. There is a case for splitting the first period into two halves, either side of World War I, but the number of Tests in the resulting two segments will be quite low, and I want all periods to have comparable numbers of Tests.

Period 1: 1877 - 1948 (307 Tests)
Period 2: 1949 - 1968 (336 Tests)
Period 3: 1969 - 1985 (390 Tests)
Period 4: 1986 - 1999 (447 Tests)
Period 5: 2000 - 2009 (464 Tests)
Period 6: 2010 - 2020 (438 Tests)

Let us move on to the period views. In the six graphs below, the results of every Test ever played have been plotted, along with the number of overs bowled in each of these matches. In order to have clear horizontal separation, I have split the data across two tiers within each graph. The duration is capped at 600 overs - only nine Tests exceed this value. Blue lines indicate Tests that ended decisively and black lines drawn Tests.

The first period was a long and complex one: three-day Tests, uncovered pitches, frequent rule changes, a giant of a batsman setting the cricketing world alight, Bodyline, timeless Tests outside England, and so on. The high percentage of decisive Tests confirms this. Most drawn Tests were in England; 97 of the 101 Tests held in Australia were decisive. There were three stretches of decisive Tests early on in this period: 22, 18 and 17 respectively.

The duration of Tests outside England was also very long at times in this period - confirmed by the seven Tests capped at 600 overs, and one Test reaching 907 overs. However, many results were achieved in fewer than 200 overs. It is also clear that the two halves either side of WWI were chalk and cheese.

In total contrast, the second period, which started after Bradman's retirement, was dull and dreary. The bleakness of the post-war years seems to have left an impact on the cricket. With the abandonment of timeless Tests, new teams were adopting safety-first tactics (there were 0-0 scorelines in three five-Test series - all involving India), defensive bowling was the order of the day and the win percentage dropped to 59.

There were three streaks of seven, nine and eight draws during this period. Despite the overall prevalent dullness, there were flashes of lightning, such as Durban 1950, Brisbane 1960 and Manchester 1961.

The 17-year span after 1968 followed the trend of the previous period. It was a defensive stretch, with only 57% of Tests being decisive. There were many sequences of five consecutive draws. Things began to change with the emergence of attacking fast bowlers in West Indies and Australia. However, it is also true that the first objective when teams visited West Indies was to avoid defeat. West Indies themselves did not dominate the way Australia were to do in later years. They were content to carve out series wins sprinkled with draws. A few magnificent contests, like the MCG Ashes Test in 1982, lit up this third period.

The fourth period consists of two distinct halves. In the first half Australia were trying to find their feet and West Indies were on the wane. It can clearly be seen that the number of decisive Tests was only around 50% during these years. There was a stretch of seven successive drawn Tests.

West Indies fell off in the 1990s, but Australia, first under Mark Taylor and then under Steve Waugh, started dominating. The period finished with a healthy 61% decisive Tests. South Africa came back with a bang and the Indian approach also turned towards the positive side.

These changes, first seen during the late 1990s, have continued till date. Despite their decline, West Indies won a few stirring contests during this period - the Adelaide heart-stopper in 1993 and the Bridgetown classic in 1999.

The millennium started on a positive note. The first fireworks were in Kolkata, with India's dramatic turnaround orchestrated by VVS Laxman, Rahul Dravid and Harbhajan Singh. That stopped the Australian juggernaut. From then on, Australia-India Tests would generally always be competitive and great contests.

One of the greatest Tests series ever - the 2005 Ashes contest - falls in this period.

South Africa competed with India and Australia vigorously this decade. England were formidable at home. The period saw over 75% of decisive Tests. It can be clearly seen that the average number of overs bowled in decisive Tests was also getting lower and lower. There was a stretch of 18 consecutive wins in the first half of the decade.

If the first decade of the millennium was very good for Test cricket, the one that just went by was even better. Four out of five matches were decisive. There were two streaks of 20 and 22 decisive Tests during the second half. The number of exciting Tests went past single figures. There were many stellar individual performances, and at least two - Kusal Perera's and Ben Stokes' all-time great innings - were played in the last year of the decade.

A perusal of the chart will indicate how infrequently the dark lines indicating draws appear. Also, look at the number of Tests finishing below 300 overs during the two decades of this century. The attacking captains of the leading teams threw the gauntlet at the feet of their oppositions, saying, "Let me see you play for a draw. Win if you can." So much so that magnificent match-saving innings like those of Faf Du Plessis in 2012 and Gautam Gambhir in 2009, and Michael Atherton in 1995, are like gold dust.

Now let us move on to the summaries.

I have looked here at the result percentages and the average number of overs bowled in all Tests by period. The result percentage is the highest in the last decade - in excess of 80%. The draw percentage was highest between 1969 and 1985 - over 43%. That's almost every other Test. The average overs bowled per Test was the lowest in the last decade - 326. The previous decade was also quite close. It is interesting to see that the average overs figures in the three columns are all virtually identical for the first period.

The highest average overs-per-draw figure was in the 1949-68 period, 414. (It must be remembered that there were quite a few six-day Tests during this period.) The average number of overs per decisive Test was very low in the last two decades - 317 and 319 respectively - which means the match does not even reach lunch on the fourth day. All the five Tests won by Australia in their 2019-20 home season did not see the fifth day.

The 1932 Test between Australia and South Africa needed only 106 overs to produce a result. In 1935 the England v West Indies Test needed only 112 overs. In 1888, England beat Australia in only 131 overs. At the other end, Test Nos. 178, 179 and 180 between Australia and England in 1929 took 701, 695 and 707 overs respectively. These three Tests were played over 22 days in total.

Finally, a look at the average number of overs bowled per day across the periods. This calculation is not easy since a day can consist of ten overs, 50 overs or 100 overs. The last days of Tests are difficult to define. Data on day-end scores is not complete, especially the number of overs bowled. However, I have made a determined effort to arrive at an average figure as accurately as possible. Any minor inconsistencies will apply across all periods.

Now let us look at the results. The first period is very good - the best ever one. Across 72 years, teams bowled an average of 96 overs per day, and there were days when over 120 overs were bowled. Despite the otherwise bleak nature of the second period, teams bowled 94 overs per day. Then the 1970s dawned and that average dropped like a stone - the next four periods have averages below 82 overs per day. It even dipped below 80 during the '80s and '90s. Whether it is the cause of teams with all-pace attacks or due to go-slow tactics, this was very poor indeed.

The above summary looks into decisive Tests more closely. Fewer than 200 overs were bowled in over 11% of Tests during the first period: that is a finish in two days or just over. But then three-day Tests and uncovered pitches meant that these were bowlers' years. In the last two decades, the percentage of Tests that finished in the 200-250 overs range was quite high - about 14%. In the fourth period, 21% of Tests finished between 250 and 300 overs.

In the decade just passed, over 30% of Tests finished in the 301-350 overs bracket. The third period leads in the 351-400 overs group - that is a fifth-day finish. Finally, the second stage had the highest 400-plus-overs finish - well into the fifth day. Compare the 40% for this grouping in the 1950s to around 12% now. It is clear that Tests finish much earlier nowadays.

The above chart is about days played. I have only taken the last two decades. Let me clarify that a two-day win in a Test means there were only two days of play. It does not matter whether the two days were the first and fifth day (like in the Centurion Test between South Africa and England in 2000). The days need not be consecutive. I have decided to include the Centurion game, and Pakistan's forfeiture of The Oval Test against England in 2006. They were results and not including them would be wrong, however much I do not agree with what happened at Centurion and The Oval.

There were 704 decisive Tests in the years between 2000 and 2020, including England's win over South Africa in Port Elizabeth last month. Out of these, seven finished in two days' play. That is 1%. For those interested, these are the Tests that finished in two days: the Centurion game mentioned above (147 overs), England v West Indies at Headingley in 2000, Pakistan v Australia in Sharjah in 2002, (148 overs), South Africa v Zimbabwe in Cape Town in 2005, Zimbabwe v New Zealand in Harare the same year, South Africa v Zimbabwe in Port Elizabeth in 2017, and India v Afghanistan in Bengaluru in 2018. One hundred and thirty Tests, representing 18.5%, finished in three days of play. A further 274 Tests (38.9%) finished in four days.

That leaves us with 293 Tests that finished on the fifth day; 187 of these, representing 26.6% of the total, took 392 overs or less. Only 106 Tests took more than 392 overs. That means that a total of 85% of decisive Tests required less than 392 overs.

Conclusion
Going only by these numbers, it would seem that a switch to four days, and 392 overs, will mean that only 15-20% of the decisive Tests will be affected. This number is quite small, making the four-day Tests a positive step, freeing up a number of days. Purely based on numbers, the switch seems to be a logical and sound move, provided the change is in favour of something positive. The tactics might change, with captains going for results in different ways.

However, the numbers only tell one side of the story. It is certain that the average number of overs needed to produce a result, which is 327 during the past 20 years, is valid only because five days were available for play. Similarly, the percentage value for decisive matches during the same period is valid because of the basic presumption of five-day Tests. If only four days are available, the dynamics change and the percentage value of decisive matches will come down. There are many non-numerical factors that should be kept in mind, as I have outlined below.

1. Very few teams bowl 90 overs a day, either because a team is stuffed with pace bowlers, or because of tactical reasons. As such, it is the height of optimism to expect that 98 overs will be bowled in a day's play.

2. Perhaps the penalty for not maintaining over rates should not be fines. For many players, those fines represent an hour's earnings, and they might not waste a minute thinking about them. Loss of WTC points would mean something only to teams fighting for the top two places. To hit the teams/captains hard, it is recommended that the team is penalised 12 runs for each over not bowled within the prescribed time, after making due allowance for all stoppages. For South Africa at the Wanderers, 36 runs would mean a lot more than six WTC points.

3. A fifth-day pitch is a great leveller and allows the spinners to have their day in the sun. This factor will not be in play in four-day Tests. Spinners' roles in winning Tests will be undermined severely.

4. A four-day Test could very well become the equivalent of four one-day matches. The strategies used extensively in a five-day Test will disappear.

5. How many Tests have we seen recently that finished with a fifth day fightback or defensive tour-de-force or great chase? These will disappear. (For an example of such a Test, you only need to go back one month, to England's superb fifth-day win in Cape Town.)

6. Will we see a Cheteshwar Pujara or Dean Elgar or Tom Latham get dropped from a Test side because they do not fit in with the four-day Test's strategic requirements, that batsmen have to score at strike rates of 50 or more? Is that a welcome change?

7. A five-day Test can allow for the loss of, say, a day's play and still produce a result. Many Tests that have finished in around 350 overs or so have been spread across five days. That level of flexibility is absent from four-day Tests.

8. A severe rain disruption of nearly two days in a five-day Test could still produce results (like the New Zealand v Bangladesh Test in Wellington last year). However, such a disruption would kill a four-day Test instantly.

9. Fascinating draws have always been part of the wonderful genre of Test cricket. Many a draw has had more excitement than one-sided wins. To play through the last day and save a game often requires more skill than a swashbuckling hundred that secures a win.

10. On mosaic floors masquerading as pitches, the four-day Test would produce the same dull draw as the five-day format.

11. In the five Tests of the Australian summer of 2019, we saw wonderful spells of spin bowling by Nathan Lyon. These were not the square-turning tracks of the subcontinent but good pitches where a bowler had to bowl many overs at a stretch to earn wickets. Lyon needed over 200 overs to take 27 wickets. The fact that all five Tests finished in four days is beside the point. If required, Lyon would have bowled 100 more overs to take these wickets. Will we ever find this facet of the game in four-day Tests?

12. Let us take a scoreline at around tea on third day - Team 1: 450, Team 2: 200. In a five-day Test, most captains will bat on and set a target in the fourth innings, giving the chasing team maybe a 5% chance. In a four-day Test, there is no alternative for the team batting first except to enforce the follow-on. This makes their task simple: take ten wickets in 4-5 sessions. However, the team batting second can only play for a draw. Terrific Tests like Manchester 1961, or Headingley 1981 or Kolkata 2001 will just disappear from the scene.

13. The players' views are very important. Many of them often work hard to finish a Test ahead of schedule, savouring the extra day of rest they thus earn. Taking these days away and stuffing in a few more T20 matches is likely to meet with a lot of resistance. It will also be counterproductive.

If the objective is to free up enough days to conduct other more commercially lucrative events, that is the wrong reason to make this fundamental change. On the other hand, if one of the objectives is to correct the drawbacks present in the World Test Championship, it has some sense. Maybe if the ICC uses the switch to four-day Tests to standardise three-Test series between all teams, then the proposal is worth considering, possibly in the 2021-23 cycle. While the numbers tell a certain story, the non-quantifiable and traditional factors have to be given a lot of weight.

I personally would not like the five-day format to be tampered with, other than for very significant reasons. However, being one of the smallest cogs in the wheel, let me leave it at that.

Email me with your comments and I will respond. This email id is to be used only for sending in comments.