October 15, 2010


Test series analysis - part 2

Anantha Narayanan
Gordon Greenidge: Best batsman in the 5-0 win against England in 1984  © Getty Images

Last week I did an analysis of the Test match results by series. In view of the length of the article and the need to get and incorporate user responses, I decided to do this as a 2-part article. I am glad that I did this way. I have had excellent responses from the users and this has added a lot of value to the follow-up article. I feel that this is one follow-up article which was turned 180 degrees through the user responses. I have made significant improvements and have also changed the very basis for measuring the teams.

First, the summary of changes.

1. The basic method of determining the performance index has been changed based on excellent suggestions by and the extensive dialog I had with Raghav Behani. He himself added value to suggestions put up by Topa Singh. More details later.

2. Coming-from-behind series wins are recognized. Thanks to Dhaval/Anand for suggesting this. More details later.

3. There will be an adjustment in points secured in the series with 1/2 matches.

4. A 3-1 win will carry lesser points than 3-0, a 2-1 win will carry lesser points than 2-0 and so on.

Shankar had made a useful suggestion that all the away series which did not have neutral umpires should get higher weights. The suggestion is eminently valid as all teams, without exceptions, had some dubious members of the umpiring fraternity. However the whole thing is muddled with one neutral umpire scheme starting at various times in different countries. I am not able to work out a clear time-line.

Performance Index

For the Test match analysis, I had done a 0-1-2 points for wins. It came out quite well. For the Series analysis I did not like this over-simplified point allocation because of the widely varying number of Test matches. Then Raghav came back with a suggestion that I do this based on the maximum points available for each series. This would reduce the imbalance problem and provide proper weight. This idea of keeping the denominator at the maximum points did not strike me until Raghav pointed that out. This was a fantastic suggestion and my thanks to Raghav and Topa Singh for this. As Raghav mentions, this will let us do some important analytical studies of some of the famous rivalries like Ashes, India-Pakistan, India-Australia et al.

Of course it required some fine tuning. It is quite easy to win the one Test in a 1-0 series and get full 100% credit. To do that in a 6-Test series is very difficult. Why, it has never been done in history of Test match cricket. Hence some downward and upward adjustment of points is called for, as summarized below.

1-Test series: Multiply secured points by 0.75.
2-Test series: Multiply secured points by 0.875.
3-Test series: No change.
4-Test series: No change.
5-Test series: Multiply secured points by 1.125.
6-Test series: Multiply secured points by 1.25.

Sum for all series of {Series points x Above weight} Index % = ---------------------------------------------------- Sum for all series of {Maximum points}

Series win points

The following table is self-explanatory.

Max   <----------------------Series result---------------------->
1-0 2-1  3-2 : 2-0 3-1  4-2 : 3-0 4-1 : 4-0 5-1 : 5-0 : 6-0

1 2.0 - - - - - - - - - - - 2 3.0 - - 4.0 - - - - - - - - 3 4.0 4.5 - 5.0 - - 6.0 - - - - - 4 5.0 5.5 - 6.0 6.5 - 7.0 - 8.0 - - - 5 6.0 6.25 6.5 7.0 7.5 - 8.0 8.5 9.0 - 10.0 - 6 6.0 6.25 6.5 7.0 7.25 7.5 8.0 8.5 9.0 9.5 10.0 12.0(yet to occur)

Coming from behind

My definition of coming from behind is different to Cricinfo/Statsguru. They treat any series in which a team has gone behind as a coming-from-behind series. Even a series in which a team loses the first Test and then wins the next four. Mine is a stricter definition which is that I will only reward teams which come from behind by a margin of 2 Tests, to either win or draw series. The summary below will explain this. No team has ever gone 0-3 behind in a 6-Test series and gone on to draw the series.

SeriesId               From   Final

0-2 to win the series 3-2 (+20%) Home win: 67 1936 Aus vs Eng 0-2 to 3-2

0-1 to win the series 2-1 (+10%) Home win: 12 1888 Eng vs Aus 0-1 to 2-1 Home win: 302 1990 Win vs Eng 0-1 to 2-1 Home win: 350 1994 Saf vs Nzl 0-1 to 2-1 Home win: 398 1997 Slk vs Nzl 0-1 to 2-1 Home win: 399 1998 Eng vs Saf 0-1 to 2-1 Home win: 440 2001 Ind vs Aus 0-1 to 2-1 Home win: 550 2006 Saf vs Ind 0-1 to 2-1 Home win: 562 2007 Saf vs Win 0-1 to 2-1

0-1 to win the series 2-1 (+15%) Away win: 323 1992 Aus vs Win 1-0 to 1-2 Away win: 352 1995 Zim vs Pak 1-0 to 1-2 Away win: 358 1995 Pak vs Slk 1-0 to 1-2 Away win: 439 2001 Slk vs Eng 1-0 to 1-2 Away win: 565 2008 Nzl vs Eng 1-0 to 1-2

0-2 to draw the series 2-2 (+10%) Home draw: 45 1928 Saf vs Eng 0-2 to 2-2 Home draw: 111 1956 Saf vs Eng 0-2 to 2-2

The team strength adjustment continues to be done as explained in the last article. The points secured by the stronger team will be reduced and points secured by the weaker team increased proportionately. I will show the detailed graphs for "All Tests" and the eight periods based on the revised method of determining the Performance Index. The analysis is current upto and including the recently concluded India-Australia Test series.

I have also made significant changes to the graphs. To the extent possible the team colours are used. Tough task because of 3 shades of blue, 3 shades of green and 2 shades of red. The descriptive table at the right now contains both series played/won/drawn/lost information as also the number of 1/2/3/4/5/6-Test series played by the teams. A treasure-trove of information.

graph of Test series results across the years
© Anantha Narayanan

Australia leads the all-Tests table by a comfortable with an index value of 55.2%. When one sees that nearly half the series Australia have played are the longer 4/5/6-match series, their index figure assumes even greater value. West Indies have moved into the second position at the expense of England. Then come South Africa and Pakistan. Now there is a switch. Sri Lanka moves into sixth place, displacing India. One possible reason could be that Sri Lanka has never played even one 4/5/6-Test series. However it must be noted that they lose out on the 1/2-Test series. And India has been quite average, until 1990. They are carrying a lot of baggage.

graph of Test series results in the 2000s
© Anantha Narayanan

The 2000s decade has Australia on top, way above South Africa. South Africa, England and India follow next. There can be very few questions on this positioning. Here also Sri Lanka edges out Pakistan. If their away form had been better they could have challenged India.

graph of Test series results in the 1990s
© Anantha Narayanan

During the 1990s, Australia were on top, followed by South Africa and Pakistan. West Indies had not started their slide. Sri Lanka edges India out of the fifth place.

graph of Test series results in the 1980s
© Anantha Narayanan

The 1980s was the time when Calypso was king, and how. They have the best ever decade of any team, winning 14 series and drawing 5, often away. Not one of the other teams even crossed 50%. Even series wins were rare for teams like India.

graph of Test series results in the 1970s
© Anantha Narayanan

England, less troubled by Packer and WSC, led the 1970s decade, comfortably ahead of Australia and West Indies. India were also quite passable during this decade, no doubt due to Gavaskar and the spin quartet.

graph of Test series results in the 1960s
© Anantha Narayanan

The swinging 60s were the time of the swinging giants from the equatorial islands. Led by Sobers, West Indies were on top. Australia and England were in the next two places. Pakistan were quite poor as they were in the decade in between Fazal and Imran Khan.

graph of Test series results in the 1950s
© Anantha Narayanan

As expected, Australia, first led by Bradman and then by Hassett, were way out on top with 73.4%. England and West Indies also did well.

graph of Test series results between WW1 and WW2
© Anantha Narayanan

Bodyline notwithstanding, this was the era of Bradman and the Australians. Another 70+% index performance. England were the only other team which competed. the others just made the numbers. However I can assure the readers that wins against these weaklings have been seriously under-values.

graph of Test series results before WW1
© Anantha Narayanan

This was an even period, with England taking the edge. They travelled well. South Africa were quite a good team towards the end of this period.

The following are the three top-ranked series in terms of points secured.

256 (1984)    West Indies defeated England 5-0 away. This was a clean sweep.
213 (1978-79) England defeated Australia 5-1 away. Australia were depleted 
but this being a 6-Test series and a huge away win gained lots of points. 296 (1989) Australia defeated England 4-0 away.

Users can view/download the following tables. Some interested users could even think of importing these tables and deriving some statistical insights.

All-Tests and Period summary: Please click/right-click here.

Summary of all series - Chronological : Please click/right-click here.

Summary of all series - by Team : Please click/right-click here.

Caveat Lector: Since I use the published scorecards of Cricinfo, I have done the Series analysis by comparing the Series descriptions provided. There are some discrepancies over the years. So it is possible that a series might be split into two series. It is impossible to be 100% accurate. An example is the one Richard Mackey pointed out. For the first two Tests the description was "India in Australia, 2003-04" and for the last two Tests, it was "Border-Gavaskar Trophy, 2003-04". This has since been corrected.

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 Meety on (November 1, 2010, 6:56 GMT)

Good work, I find these articles very interesting, I wonder if this analysis can be reduced from essentially a decade analysis to Calender Year? I think this analysis + the comment earlier regarding Sliding scale from Topa Singh would be good reading. On a different tangent - you could analyse a nation against its lifetime score. For Example, Oz's all test score is 55% it could be said that in the 1980s 32% would be considered below par, so the diff of 23% could be presented by a block below the "par line" of 55%. This all makes perfect sense to me but it might of come out as jibberish - hop you know what I was trying to say - LOL! [[ Andrew The peer comparisons of players was a major step in that direction. Ananth: ]]

Posted by Taimur Khan on (October 25, 2010, 9:07 GMT)

Dear Ananth Well researched and thought out as usual! Thanks for all the pleasure and understanding you bring to us of the game we all love. To all the readers....we follow cricket because we love the game not because we are indians Pakistanis etc. Otherwise why are we not following hockey, squash etc. with the same fervour. So lets not bring jingoism into it. Second, there is room on this site for all levels of analysis. If this analysis is too obscure for some readers, they have the option of moving on. Some like myself really appreciate its depth and wisdom so let us enjoy.

Posted by Pawan Mathur on (October 19, 2010, 21:58 GMT)

Ananth, firt of all i must congratulate you that you are actually publishing some genuine cricketing stuff and have not caught the "pick my all time XI" fever which has infected cricinfo. Now coming to my query, i just want to know whether any wider quantitative analysis can be done of captaincy skills (more than just matches played/won/lost/drawn). Because i think that such an analysis could. be an avid compliment to your current test analysis as well as as earlier analysis of decade wise supremacy.

Posted by Anand on (October 19, 2010, 5:26 GMT)

I dont mean to be offensive to anyone but if someone says numbers are non-sense they shouldnt be watching or enjoying cricket at all. A team wins based on the NUMBER of runs scored. I am yet to see a bowling analysis that reads "beautiful overs-lot of maidens-few runs-some wickets". While the exact values of the weights in any statistical analysis may be debatable what I see is the philosophy in assigning the weights (relative values) and if that makes sense then the whole thing makes sense. By the way if someone feels winning in India in 1984-85 was tough, check out this stat. Between 1982 and 86 India played 44 tests and won ONLY ONE. At home, between 1982-87, they played around 32 and won 3, 2 against Sri Lanka (who were not much better that today's Zimbabwe). Again one can look at them as facts or numbers nonsense. By the way I am not sure if Bradman would be the same if we associated a NUMBER like 29.94 for his average instead of 99.94 or is that also just a number non-sense?

Posted by Chi Square on (October 19, 2010, 2:24 GMT)

Much ado about nothing! I and Anantha were just having a simple conversation and everyone decided to jump in. welcome everyone :) Just Joking. Yes, I would like to be of help, but at the risk of blowing my own trumpet, I am a singularly busy person hence have little time to follow my passion except read selective stats blogs and offer advice: which is sometimes misunderstood. However, if you could send me the relevant data sets in a excel spreadsheet, I can take a stab at it. I love stats and analysis and do have access to the relevant analytical resources. [[ CS Will send within before end of day. Thanks. Ananth: ]]

Posted by Richard on (October 18, 2010, 22:43 GMT)

Sad to hear some of the flak you are getting. These comments pages at cricinfo are full of it unfortunately. Chin up mate, keep going please as many of us value an open mind and a love of cricket itself. The squeaky wheels always get the grease!

Posted by Rohit Singh Loomba on (October 18, 2010, 19:30 GMT)

Dear Mr. Garg,

I understand what you are saying and respect that but you on your can not decide on the primary objective of any blog. The primary objective simply appears to be (I don't know for sure) to do inetersting things with numbers to inform the cricket lovers. Sometimes it will be simple, sometimes it shall be complex depending upon the nature of the analysis. Please don't create this false division between the 'elite' and the 'masses'. We are all cricket lovers and as Hema says some of us just want to enjoy the drive. But that means someone has to figure out the transmission and braking etc: we can not a drive a toy car. The numbers do tell us a lot but one needs to put ear to the ground as well to listen. I reckon Anantha is doing a great job in combining both. [[ Rohit I feel too much is being made of Vineeth's comments made in a nice acceptable manner. ALl of us have our rights to make our suggestions. As he himself has requested I suggect let us close it. Ananth: ]]

Posted by Vinish on (October 18, 2010, 17:53 GMT)

Hema A

I think you could not understand my comment in right sense. If you see carefully, I have in fact appreciated the value addition brought in by 'technical comments'.

See respponse by Ananth to my comment, he makes it clear enough. At no point did I sound discouraging or pulling down something. I just made a general comment from the perspective of 'common readers' who are not always too interested in 'technical aspects'. Nonetheless, we can drop it for now, and move on.

I shall try to be more clear.


Posted by Hema Adhikari on (October 18, 2010, 15:06 GMT)

Vinish, Well it is disappointing that you are not into in-depth analysis. The truth is deeper the analysis more meaningful the result. You don't have to know how the car engine works to appreciate a good car. Just see the end result. This blog is called 'it figures' and if people are collaborating on something to make it a little better, I think you should be supportive and not discouraging. See what one can do to make it better and not try to put each other down. Thanks. [[ HA The problem was that some of the mails were extremely condescending and the type of "what sort of nonsense articles are these". The readers are also scared that this blog would become very intensive, difficult-to-understand. As long as all of us understand this it would be fine. Additional insights are welcome but should be explained and understandable to most readers. The putting-down was from the detractors' side and hurt me quite a lot. This, without the readers seeing some of the extremely nasty unpublished mails. One guy saying I should get Pakistani citizenship and another one saying that I was churning out non-intensive articles just to make more money. When you know the quantum of work I put in and the honorarium I get, you guys would all LOL and think I should be certified. Ananth: ]]

Posted by Vinish Garg on (October 18, 2010, 8:28 GMT)

For last 2-3-4 blog posts by Ananth, I see many technical comments such as using z-factor, t-value and how rankings for ATPI/FIFA are done!

I understand the objective and value addition by such comments, however, the 'general' readers might feel put-off if there are too many comments of this nature. As a cricket enthusiast, I look forward to read 'what I understand' more often (not less than 19 out of 20 times).

Letus not forget the *primary* objective of Ananth's blog. If there are 10 readers posting comments (whether technical/general, there are 90 readers who read the blog but dont comment).

We CANNOT have two different discussion threads - one for *technical* and another for *generic* comments. Can we have a convention to distinguish the two cases? For example, a user posting a *technical* comment should begin it by writting Technical at the beginning of comment? I am not sure of how well and how acccurately they do so... I leave that to Ananth and to such users!

Regards [[ Vinish I will never lose sight of the main objective which is to bring understandable cricket analysis to number of average cricket enthusiasts. Just to take the game beyond the normal batting/bowling average level. I have no prooblems with people commenting on the absence of statistical depth. I only want the comments to be positive in nature. If my work has analytical depth, I am satisfied. It is difficult for the commenters to preface their comments. However, where warranted, I will put in a caption, summarizing the comment. Ananth: ]]

Comments have now been closed for this article


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