THE CORDON HOME

BLOGS ARCHIVES
SELECT BLOG
November 1, 2008

Test cricket

A summary of Test cricket by period (Part 1)

Anantha Narayanan
Steve Waugh of Australia walks out to a guard of honour from his teammates during day one of the 4th Test between Australia and India at the SCG on January 2, 2004 in Sydney, Australia
 © Getty Images
Enlarge

This is a major attempt to generate a set of measures for Test Cricket by period. The purpose is two-fold. The first is to look at the way the figures change over the years, letting us get a handle on the evolution of the game. The second is to establish a criteria for adjusting any analysis we do which spans across the years. Many a time have I found myself in a situation needing to adjust a particular period's figures and I have re-invented the wheel every time. Now I hope to have a set of figures which can be used as a ready reckoner for such adjustments. Readers who do similar analysis are welcome to use these figures.

Readers should also realize that after I thought of this complex topic, I have put in nearly a month's work, on and off. into preparing this complicated analysis. I would appreciate avoiding of a superficial read and flippant off-the-cuff comments.

The analysis covers various aspects of Test Cricket. Since the article has become too long, it has been split into two parts. The first part covers Matches, Innings, Results, Partnerships and Extras. The second part covers Batting, Bowling, Keeping and Dismissals.

To start with let me divide the 130 years into 8 periods, taking into account the evolution of the game, years and the number of Tests played. The following are the periods.

1. 1877 - 1914  (Pre World war 1)
2. 1920 - 1939  (In between the two World Wars)
3. 1946 - 1959  (1940s & 1950s)
4. 1960 - 1969  (1960s)
5. 1970 - 1979  (1970s)
6. 1980 - 1989  (1980s)
7. 1990 - 1999  (1990s)
8. 2000 - 2008  (2000s)
These are logical and reasonably evenly spaced periods. Anything more will result in too many periods with consequent difficulty in following the tables and anything less will telescope multiple differing periods into one and we will lose out in analysis.

Even the formatting of the article required a lot of thinking. I tried having the periods horizontally. It was difficult to read. There was also the need to present the core data such as runs, wickets, balls, wickets et al to the readers. So I adopted a dual presentation approach. In the main body of the article I show the calculated measures in a grouped form and the base core data in the supporting pages. That way all the information is shown and the main report is not cluttered. I have also avoided showing the variance of each period figure to the all-Test averages to avoid showing too many numbers. That will indeed be the key figure to make adjustments.

Let us get into the analysis results.

First the base Match analysis.

1. Match analysis 1 (Balls/Runs/Wkts per match)

Period     Mats   B/M  R/M  W/M

Pre-WW1 134 1799 812 33.6 WW1-WW2 140 2171 976 29.9 40s-50s 209 2303 912 30.4 1960s 186 2409 1003 31.1 1970s 197 2259 1014 31.0 1980s 267 1985 949 29.2 1990s 347 2018 963 30.5 2000s 409 1967 1046 31.1

All Tests 1889 2093 973 30.7

During the first period, timeless Tests and 3-day Tests alternated. Later 3-day, 4-day, 5-day, 6-day and timeless Tests were played through the years until 1979, from which year almost all the 1000+ tests have been played over 5 days. As recently as 1973, 4-days tests were played between New Zealand and Pakistan. Please remember these pertinent facts while perusing this table.

Surprisingly the Balls per match figure during the first period has been quite high despite the number of 3-day tests. This, despite 4-ball overs during most of these years requiring more change over time. During 1960s the balls per match is the highest. More than the match days, I feel this is certainly a result of lot more drawn matches during this period and to a lesser extent the 1970s.

The runs per match is the highest during the current decade and the lowest during the first period when batting was indeed difficult. The relatively high 1960s and 1970s figure must no doubt be due to the number of drawn matches.

More Wickets per match fell during the first period. Barring this period the figure has remained fairly static.

To view the complete table please click here.

2. Match analysis 2 (Runs/Wkt, Runs/Over)

Period      RpO    RpW

Pre-WW1 2.71 24.2 WW1-WW2 2.70 32.7 40s-50s 2.37 29.9 1960s 2.50 32.2 1970s 2.69 32.7 1980s 2.87 32.5 1990s 2.87 31.6 2000s 3.19 33.7

All Tests 2.79 31.7

The RpO figure is the most important measure we have seen until now. It has varied quite significantly over the years. Surprisingly the Rpo figure was quite high during the first two periods despite the pitches. It fell drastically during the post-112 period, certainly due to a combination of accurate bowling and defensive batting and attitudes. The figure picked up later and has crossed the key value of 3.0 for the current decade, where it is 14% higher than the all-test average. This has been the result of most teams, led by Australia, scoring quickly in a bid to go for a result.

There have been 4/5/8 ball overs at different times in Test cricket, however all RpO figures have been standardized to 6 bpo for this table.

Barring the first period the Runs per wicket figure has remained fairly stable. The figure is highest during current decades. For most of the periods the RpW figure has exceeded 30.

To view the complete table please click here.

3. Inns Analysis (Runs per completed inns, Low and high scores)

Period      R/CI  I<100   I>500

Pre-WW1 231 12.71% 2.63% WW1-WW2 289 3.14% 6.68% 40s-50s 256 6.46% 6.79% 1960s 284 1.93% 5.80% 1970s 276 2.62% 5.24% 1980s 278 2.23% 5.64% 1990s 269 1.71% 5.47% 2000s 282 3.42% 9.87%

All Tests 272 3.84% 6.49%

The average completed innings size has followed the pattern. Quite low (15% below all-Test average) during the first period and then around the all-Test average mark subsequently, barring the low-scoring 40s-50s period..

During the first period, there was an extraordinarily high instances of sub-100 innings. Over 12.5% of the innings completed (53 out of 494) were below 100. The second period was a major drop in the sub-100 innings. However the figure almost doubled during the 40s-50s. Then it has settled down. The 1990s had the lowest figure. Surprisingly the current decade's is double that of the previous decade. There have been 36 such instances out of 1052 completed innings.

I was so intrigued by this sudden escalation that I decided to make a detailed study. As expected the culprits were Bangladesh with 8 sub-100 scores and Zimbabwe with 7. However the situation has been worsened by the West Indian decline. They have had 5 sub-100 scores. At the other end, Australia and South Africa have had one instance each.

Predictably there were very few 500+ innings during the first period. Then the % stabilized to the all-Test average during the next 6 periods. There has been a noticeable increase during the current decade with 147 of the 1489 innings crossing 500. Remember that these are not just completed innings but all innings.

Australia leads with 28 500+ scores while India is close with 24. At the other end Zimbabwe has only 2 scores in excess of 500 while Bangladesh has not crossed 500.

The paradoxical current decade situation of high number of 500+ scores and high number of sub-100 scores is a pointer to the wide gap between teams as well as the drive to achieve results.

To view the complete table please click here.

4. Partnerships analysis (Opening & Last 3 wkts)

Period     Open OP100+ OPSub10 Last3W

Pre-WW1 29.8 5.9% 37.2% 47.3 WW1-WW2 40.5 11.3% 28.1% 47.5 4os-50s 36.3 8.1% 27.4% 40.7 1960s 38.2 7.5% 24.9% 49.0 1970s 38.3 8.0% 27.0% 47.5 1980s 34.2 6.4% 27.5% 50.0 1990s 35.7 8.2% 30.3% 48.2 2000s 39.0 8.9% 28.7% 49.8

All Tests 36.7 8.1% 28.7% 47.8

This is an analysis of two types of partnerships. The first wicket partnership is the most important one since it lays the foundation for the innings. The average first-wicket score has been reasonably scattered around the all-Test average of 36.7 barring the first period when it fell below 30. In between the wars the partnership average went past 40, possibly owing to the strong opening partnerships of England and Australia.

Even though I am not a fan of measuring quality through individual 100s (I always treat the 100th run as nothing more than the run(s) scored around the 99 mark), a 100 partnership is more significant since it delivers a psychological blow for the team. A fairly low number of partnerships during the first period crossed 100. Surprisingly this was followed by a doubling during the next period with over 10% of the partnerships crossing 100 (56 out of 494). There has been a recent increase during the current decade, also at a good scoring rate.

The next is a measure of opening failures. These are the sub-10 (single digit) partnerships. This includes only instances where the first wicket has fallen. During the early days, especially during the first period, well over a third of the partnerships have been failures. This figure improved over the years but has picked up now and we are back to a fairly high (either side of) 30% figure. It may have to do with the attacking attitude of the opening batsmen nowadays. I could have done a "opener dismissed at 0" analysis. However I feel that a single digit partnership is a failure and a 0 is no worse than a 5 or 9.

The next measure is the number of runs added for the last 3 wickets. This has not varied much barring the 40s-50s when it fell to around 40. For the current decade the value is around 50, indicating a more committed late order batting set-up with better techniques and application.

To view the complete table please click here.

5. Extras Analysis - per 1000 balls (Extras/Byes/LegByes/NoBalls/Wides)

Period     E/Tb B/Tb L/Tb N/Tb W/Tb

Pre-WW1 22.6 12.8 6.5 2.3 1.0 WW1-WW2 21.2 9.3 8.6 2.8 0.5 40s-50s 16.8 7.6 6.3 2.3 0.5 1960s 18.9 6.6 7.3 4.4 0.5 1970s 27.9 6.5 9.2 11.3 1.0 1980s 32.0 6.0 11.8 12.5 1.7 1990s 33.1 5.9 12.4 13.4 1.4 2000s 34.0 6.9 11.9 12.2 2.9

All Tests 27.4 7.1 9.9 8.9 1.4

All the extras calculations have been done per 1000 balls. This is just a convenient measure and is to be used only as a relative measure for comparison. All the extras components have been analyzed.

The number of Extras per Tb has increased over the years and the current decade figure is the highest, about 20% higher than the all-Tests average.

The Byes per Tb started at a high figure and now stands around the all-Test average. Have the keepers become that much better?

Leg Byes follows the reverse pattern. Starting at a low level it is now at a fairly high level.

No Balls per Tb have increased significantly. They were extremely low during the first 70 years and suddenly zoomed up during the 1970s and have remained there. Possibly the changing of the No ball rule during the 1960s must have contributed to this increase.

Wides per Tb have also increased during the current decade, almost double of the all-Test average and the previous period of value. Possibly the bowlers are striving for too much. May also be that the unmpires, no doubt influenced by the ODI experience, are calling wides more often now.

The increase in LB/Nb/Wides per Tb has more than odffset the drop in Byes per Tb and this has resulted in the overall increase in Extras per Tb.

To view the complete table please click here.

6. Results Analysis (Results/HomeWins/AwayWins)

Period     Res% HW % AW % Dr %

Pre-WW1 82.1 44.0 38.1 17.9 WW1-WW2 62.9 35.7 27.1 37.1 40s-50s 65.1 36.8 28.2 34.9 1960s 52.2 30.6 21.5 47.8 1970s 57.4 35.0 22.3 42.6 1980s 53.9 32.6 21.3 46.1 1990s 64.3 40.9 23.3 35.7 2000s 77.0 46.2 30.8 23.0

All Tests 64.9 38.6 26.3 35.1

This is a very interesting table. The overall Results % started at an incredible 82+ value during the first period, dropped to a low 50+% during the miserable 1960s and has risen again now to a near-80% value. Australia might be stuttering now. However they are the team which started the equivalent of "total cricket", hard, attacking and always striving for results. Due credit should be given to them for changing the face of Test cricket, especially after the miserable 1960s-1980s periods.

A similar pattern emerges in the Home wins measure. The first and last periods have high Home wins % values.

The best period for Away wins was the first one when the 3 month sea travel seemed to have done something good since 38% of the matches finished with Away wins. This value has since dropped and stood at its lowest during the 1960s when "Not to lose" was the motto. The value has picked up significantly during the current decade with over 30% Away wins.

The Draws % shows low values during the first and last periods. The most boring period in Test history was during 1960s when nearly half of the matches ended in draws, not all of them the exciting ones.

To view the complete table please click here.

The second part of the article will follow in a week's time covering the Batting, Bowling, Keeping and Dismissals aspects. I will also try and do some changes based on any significant comments. I invite readers' comments, both on these areas and the ones being analyzed next week. At the end of the second part, if readers so desire, I will also make available an XL sheet containing all the measures analysed, including % variances to the all-Test averages.

Anantha Narayanan has written for ESPNcricinfo and CastrolCricket and worked with a number of companies on their cricket performance ratings-related systems

RSS Feeds: Anantha Narayanan

© ESPN Sports Media Ltd.

Posted by Martin on (November 7, 2008, 12:42 GMT)

I like to do a bit of statistical analysis myself but with individual players performances, and find many of today's players more often do better than most of even the best of the 70s to 90s players. Trying to find some reason I have left statistics aside and studied financial motivation. In all sports especially football, but also very much in cricket the money incentive has changed the reason for playing well or achieving results. Your fascinating analysis seems to show that the last 20 years' cricket has been energised and players have become more professional and motivated. How about comparing players test match incomes and performances over the years?

Posted by CM on (November 7, 2008, 6:45 GMT)

The increase in Leg Byes seems to correspond to the period when wearing helmets became common. Perhaps this figure indicates that now they are better protected batsmen are increasingly prepared to get behind the ball and stay there, resulting in more hits on the body and hence more LB's.

Posted by SRK on (November 5, 2008, 14:13 GMT)

I am sorry for troubling you again. However I am still confused. Can you explain how the entries in the table in the paragraph 6 are calculated? [[ Ananth: I think there is a small blind spot which is preventing you from understanding. Happens to all of us. Also please read my earlier reply. Let me explain. If India and Australia play two series of 5 tests each, one at India and the other at Australia. The first series in India ends in a 2-1 win for India and the one at Australia ends in a 3-1 win for Australia. There have been 10 matches, 7 results and 3 draws. That is simple. Out of the results, there have been 5 home wins (2 for India and 3 for Australia) and 2 away wins (1 each for India and Australia). So this analysis will yield 5 home wins, 2 away wins and 3 draws. As I have explained already, do not think about losses, only wins. ]]

Posted by redneck on (November 5, 2008, 5:58 GMT)

very interesting read! with all this talk these days of batsman friendly wickets through out the test world, im suprised to see the current periods has a higher result percentage over every other period bar the golden age. excelent summary Ananth

Posted by Mish on (November 4, 2008, 10:47 GMT)

In my defence, I wasn't suggesting that you removed Zimbabwe/ Bangladesh from the figures. I was more interested in how much the increased results of the 2000s were due to those weak sides or whether it was part of an overall trend due to increased run rates and generaly more agressive play.

Posted by Ananth on (November 3, 2008, 5:38 GMT)

This is in response to Daniel's comments.

I have treated a tied match as a draw. That is the correct interpretation. To treat a tie as a result means there has a winner, consequently a loser which is clearly not the case. Neither team deserves that tag.

I accept that a tie may be a positive result. No more than many a great drawn match. There are draws and draws and let us consider a tie as a wonderful draw.

I have not considered Penalty runs separately since I felt that it would have very little impact on the overall conclusions. Only purpose would be to be 100% accurate.

Posted by Gareth on (November 3, 2008, 4:37 GMT)

There was a law change for wides (amongst other things) in 2000. Previously to be a wide, the ball had to be "out of reach of the striker"; now it is a wide unless the batsman can "hit it with his bat by means of a normal cricket stroke."

Posted by Cas on (November 2, 2008, 17:13 GMT)

I'll probably get a lot of stick for this but I find it hard to asknowledge the first period as real TEST cricket. Just a couple of teams playing against each other? From what i've read so far, the wickets were very contrasting/ Lohman getting his wickets at that average. I mean was he really that good or was it the batsman who were so crap!

Posted by Chris on (November 2, 2008, 11:54 GMT)

Byes were more prevalent in the early days as most keepers stood up to nearly all bowling, irrespective of the pace. The marginal increase in byes over the last decade is probably a result of the drive to pick the best batsman among keepers rather than the best keeper that has become prevalent in many countries during recent years.

Posted by James on (November 2, 2008, 11:43 GMT)

A look at the variation in the numbers you give doesn' appear to really support the statement about a 'more committed' tail, e.g. 49.8 doesn't seem very different from the 48.2 and 50.0 of the previous two decades. Not being a statistician, would the use of measures of statistical significance (p values and all that) to objectively assess qualitative statements such as 'tail order batting has improved in recent years' be applicable? It would be useful to differentiate between likely random fluctuations and significant trends. Many thanks for an interesting article [[ Ananth: Chris/James, a common reply. Both of you are correct. The current decade figures are only at par with the previous two decades. It is however true that the current decade and 1980s have seen a 5% increase over the all-time average. 5% is a significant variation. I agree that the only DEFINITE statement I can make is on the 1940s-50s when the figure was nearly 15% below all-test average. ]]

Comments have now been closed for this article

ABOUT THE AUTHOR

Anantha Narayanan
Anantha spent the first half of his four-decade working career with corporates like IBM, Shaw Wallace, NCR, Sime Darby and the Spinneys group in IT-related positions. In the second half, he has worked on cricket simulation, ratings, data mining, analysis and writing, amongst other things. He was the creator of the Wisden 100 lists, released in 2001. He has written for ESPNcricinfo and CastrolCricket, and worked extensively with Maruti Motors, Idea Cellular and Castrol on their performance ratings-related systems. He is an armchair connoisseur of most sports. His other passion is tennis, and he thinks Roger Federer is the greatest sportsman to have walked on earth.

All articles by this writer