Which has been the best era for batting in Test cricket's history?

The Wellington Test between New Zealand and Sri Lanka earlier this year - which the home team won by an innings and 58 runs - was the 2500th Test match in men's cricket. It wasn't a particularly memorable contest, but the landmark was a reminder of how long the format has been around.

The very nature of the five-day game means only a limited number of them can be fitted into an annual calendar. In 2022 only 43 Tests were played, compared to 161 ODIs and a whopping 531 T20Is (that last number is also a function of the huge number of teams that play the shortest format). While it has taken 146 years for Test cricket to reach 2500 matches, the other two formats have been rapidly adding to their numbers: in 52 years of the 50-over ODI, 4578 matches have taken place, while 2076 T20Is have been played in only 18 years.

That by itself means milestone matches come far less frequently in Tests. The 2000th Test was played almost 12 years ago, in July 2011, and the 1500th about 11 years before that. And that is super quick compared to the first 500, which took 83 years. (Even accounting for the 11 years lost to the two world wars, that's a long time.)

So how has Test cricket changed in this period, and which teams and players have been the most dominant in each of the five non-overlapping 500-Test blocks? Let the numbers tell the story.

Overall numbers in Test cricket, 1877 to 2023

For a format that has been played for more than 145 years, it is remarkable how similar the batting averages are in the five blocks. The early part of the first block featured lower averages due to uncovered pitches, but since then the number has hovered around the early 30s. The averages in the third and fifth blocks vary from each other by a minuscule 0.01, while in the second block (1960-1984), it was only about half a run higher.

In the first decade of the 2000s, the overall batting average jumped to almost 35. That was clearly the batting era, characterised by relatively flat pitches, several batters who averaged above 50, and bowlers who struggled to achieve sub-30 averages. That is the only block of 500 Tests in which more than 1000 centuries were scored - 1042, compared to 766, 854, 791 and 951 in the four other blocks (in chronological order). Since then, bowlers have fought back to re-establish equilibrium, and in the last five years, they have even turned the tables on the batters.

What has changed, though, are scoring rates, and the number of non-draw results. From a dour 2.54 runs per over in the first block, the run rate has gone well past three now, which seems to be the influence of the 50- and 20-over formats. That's the one data point where the number has consistently gone up in every period, apart from the last decade, when it dipped marginally in comparison to the previous one, largely due to the more bowler-friendly conditions on offer. The percentage of draws has drastically decreased too, from around one in every two games in the second block of 500 Tests, to one every five games in the last decade.

Team-wise trends in each 500-match block of Test cricket

To start with, England and Australia were the dominant teams. Of the first 500 Tests, in 178 those two teams were pitted against each other, while 310 involved at least one of them. South Africa were the only other team to play more than 100 Tests in this period. However, while Australia and England won many more Tests than they lost, South Africa largely struggled in this period, winning only 27 against 72 defeats.

West Indies emerged as an outstanding Test team in the next period, winning twice as many Tests as they lost (57-28). They had a wonderful period from 1962 to 1967 (15 wins, three losses), and then from 1980 were dominant through the rest of that decade. In the last 44 Tests they played in this block, West Indies won 22 and lost only two.

The third block of 500 Tests (between 1984 and 2000) was notable for Pakistan's rise and South Africa's return after their apartheid-era ban. Pakistan won 43 and lost 29, giving them a healthy ratio of 1.48, bettered only by Australia (1.84), West Indies (1.65) and South Africa (2.28), who came back international cricket with a formidable line-up in 1992. England struggled in this period (36 wins, 66 losses), but picked themselves up in the next, winning 65 and losing only 38. The last two blocks are also notable for India's rise (111-63 across the two periods), and West Indies' stunning decline (44-115).

In fact, India's win percentage of 52.99 since July 29, 2011 is the second best by any team in any of these five 500-Test blocks, bettered only by Australia's 65.32% wins in the 1501-2000 block.

In terms of win-loss ratio, though, India's 62-35 record in that post-2011 period (ratio 1.77) is in seventh place. Australia's record in the fourth block (2000-2011) was 81 wins against 24 losses, a stunning ratio of 3.38, which sits far ahead of anything any team has achieved. West Indies' slump is apparent from these stats: from a win-loss ratio of 2.04 between 1960 and 1984, the fourth best in any block, they have lost more than 50% of their Tests in the last two blocks (63 out of 113 and 52 out of 98), among the worst performances by any team.

Batting: which were the best and worst periods?

From the five non-overlapping blocks of 500 Tests (the last block is 503 games), it's clear that the period between Tests No. 1501 and 2000 - the first decade of the 21st century - was an extremely good one for batters: the average runs per wicket was 34.62 in that block of matches; excluding extras, the batting average was 32.48, compared to 29.7 and 30.41 in the 500-Test blocks either side of it. That's an increase of around 8% in that decade, compared to the periods before and after. Thirty-six batters made over 4000 runs in that period, of whom 16 averaged at least 50. In none of the other four blocks did so many batters enjoy so much success.

In fact, after July 29, 2011, only five out of the 25 batters with 4000-plus runs also averaged over 50: Steve Smith, Kane Williamson, AB de Villiers, Younis Khan and Joe Root. Virat Kohli, with an average of 49.94, just misses out. In the period between November 1984 and June 2000, it was five out of 23 - the chosen ones were again the cream of the lot: Sachin Tendulkar, Brian Lara, Allan Border, Martin Crowe and Steve Waugh (batters like Graham Gooch, Mohammad Azharuddin and Inzamam-ul-Haq missed out).

Within each period - which spans more than ten years - there could be specific periods when the averages went further higher or lower. That can be recorded by calculating the moving averages for a smaller number of matches. The next couple of graphs plot the moving averages calculated over 100-Test spans, which means the first plot point is at the 100th Test (the average over matches 1-100), the second is the at the 101st Test (average over matches 2-101) etc. There are thus 2404 plot points in the graph.

The highest peak among those points is 34.92, which is the batting average (excluding extras) for the 100 Tests played between August 8, 2008 and December 16, 2010. That period falls towards the end of the fourth block of 500 Tests. On the other extreme, the lowest average over a 100-Test period was in the very early days of Test cricket, between December 31, 1881 and July 26, 1909, when it slumped to 22.27. If you consider the last 60 years, the lowest phase came recently: between January 5, 2018 and August 5, 2020, the batting average dropped to 27.65, which is a 21% drop on the highs of 2008-10.

Breaking up these numbers by the top seven and bottom four batters in XIs, the dominant batting periods remain largely the same. The best 100-Test phase for the top seven was between October 17, 2008 and December 26, 2010, when they averaged 41.41. That average was matched in the 100 Tests between October 25, 2008 and January 3, 2011. On the other hand, in the period between April 30, 2017 and August 30, 2019, the average for Nos. 1-7 dropped to 32.87, which is also the lowest in the last 60 years. The percentage drop was again a significant 20.6%. Thus, the last 15 years have seen some of the best periods for batters and bowlers.

For tail-end batters too, the best phase was in the late 2000s and early 2010s: a batting average of 17.82, between December 2008 and June 2011. However, their worst lows were in the 1990s and early 2000s: they averaged 13.17 in 100 Tests between October 1990 and February 1994. And if you're wondering if lower-order batting has indeed improved overall through the years, here are the averages in the five blocks, in chronological order: 14.75, 15.34, 14.85, 15.77, 15.52.

When did bowlers thrive, and when did they need to toil?

If the batters had the time of their lives in the early 2000s, then it's obvious the bowlers didn't. Nineteen of them took 200 or more wickets between June 2000 and July 2011 (the fourth block of 500 Tests), but only eight of those had sub-30 averages. In each of the other blocks, almost all the bowlers with 200-plus wickets also had sub-30 averages. The first block of 500 Tests only had three bowlers with 200-plus wickets, simply because of the relatively small number of Tests played per year: till 1960, there were only two years with 20-plus Tests, and eight years with more than 15 Tests. That meant bowlers needed to have much longer careers to give themselves a chance of taking 200 wickets.

Alec Bedser, Ray Lindwall and Clarrie Grimmett were the only bowlers with 200-plus wickets in the period before 1961, and they all averaged under 25. In the next 500-Test block, both those numbers went up five-fold, and all 15 bowlers who took 20 or more wickets averaged under 30, from Joel Garner (20.73) to Graham McKenzie (29.78). Others on the list included all-time legends like Imran Khan, Michael Holding, Dennis Lillee, Richard Hadlee, Ian Botham and Kapil Dev. The 1984-2000 block had a 100% sub-30 record too, and included the likes of Malcolm Marshall, Curtly Ambrose, Allan Donald, Glenn McGrath, and the two Ws from Pakistan.

In the fourth block, though, only eight out of 19 bowlers passed the sub-30 Test. While McGrath, Shane Warne, Dale Steyn and Muthiah Muralidaran were all in that club, those who missed out included James Anderson, Anil Kumble and Zaheer Khan. In the most recent block, bowlers have again regained some of their pre-eminence, with 16 out of the 18 who have 200-plus wickets also averaging under 30. They include Pat Cummins, Vernon Philander, Steyn, Anderson, R Ashwin and Ravindra Jadeja. The only bowlers outside the 30 bracket are Nathan Lyon and Yasir Shah, and even they have averaged under 32.

Like for the batters, a 100-Test moving average will reveal a more nuanced movement of averages as opposed to the single number we have for each block of 500 Tests. Since the overall bowling average will obviously resemble the overall batting one, let's look at the numbers for pace and spin. Which were the best years for each of them, and were there periods when either type did well when the other didn't?

The moving averages graph for pace looks largely similar to the one for overall bowling averages, but there's a sharp dip in three places - 1956-60, 1980-84, and then again, more recently, in the last three years. The most recent dip is the most pronounced and reveals what a dominant period this has been for fast bowling. In the period from January 5, 2018 to August 5, 2020, fast bowlers averaged 26, which is the lowest they have averaged in any 100-Test period in around 100 years. The last time they did better was between December 29, 1894 and July 2, 1921, when they averaged 25.93.

For spinners, the returns haven't been as impressive, with the average hovering around the mid-30s for a while now. The last time they averaged under 30 over a 100-Test span was way back in the period between February 18, 1956 and January 19, 1962, when they conceded 29.62 runs per wicket. That was when Jim Laker, Gary Lock and Richie Benaud were all at their lethal best, taking 100-plus wickets at sub-24 averages. Since 2018 (100-Test sequence ending in 2018 or later), the best spinners have done is 31.60, in the 100 Tests between February 23, 2017 and February 28, 2019.