ICC news

The Battle of the Rain Gods

David Hopps

May 30, 2012

Comments: 76 | Text size: A | A

Andy Flower and Paul Collingwood were left frustrated as the Duckworth/Lewis calculation left West Indies with an easy target, West Indies v England, World Twenty20, Guyana, May 3, 2010
Paul Collingwood and Andy Flower were frustrated by the Duckworth-Lewis method during the World Twenty20 in West Indies © Getty Images
Enlarge
Related Links
Numbers Game : The rain equations
Teams: England | India

The intricacies of the system used to recalculate run targets in rain-affected one-day matches are unfathomable to most people, but the ICC's cricket committee must grapple with the subject at Lord's over the next two days when it is presented with an alternative to the dreaded D/L method which has been adopted in international cricket in the past 15 years.

The Battle of the Rain Gods might not quite rival the best of Greek mythology - Hollywood is not yet thought to be interested - but it does not lack importance. A World Cup could one day depend on the outcome.

On one side are two reserved statisticians from Lytham St Annes, a peaceful seaside resort on the Lancashire coast, Frank Duckworth and Tony Lewis, whose system has benefited from an ever-increasing amount of data and has gradually won acceptance in cricket circles as making the best of a bad job.

On the other side is a persistent engineer from the southern Indian state of Kerala, V Jayadevan, who has had the audacity to challenge the established order by claiming that the D/L method "comprises several silly mistakes."

In its place, he proposes the alternative that he has worked on tirelessly for the past 15 years - the VJD method. VJD, to British minds at least, sounds disturbingly like the return of Mad Cow Disease, but Jayadevan insists that cricketing sanity is one of the advantages of his system, which has already been trialled in Indian domestic cricket and was also adopted in the now-defunct Indian Cricket League.

The D/L system was eventually introduced in response to a farcical finish to England's World Cup semi-final against South Africa in Johannesburg in 1992 when South Africa resumed after rain to find their target had been reduced to 21 runs from one ball.

There is a sense that Jayadevan's rival system is too much for the ICC to cope with. It took cricket officials years to respond to his letters. One set of rain rules was quite enough. But gradually Jayadevan, a deputy director in the Kerala Engineering and Research Institute*, won support, notably from the former India captain Sunil Gavaskar.

And so, the ICC annual meeting in Hong Kong last June received his proposals and, doubtless with a grimace, passed them down to the cricket committee at Lord's for their consideration. Dave Richardson, the ICC's chief-executive-in-waiting, has indicated that they will be taken seriously. Clive Lloyd, the former West Indies captain, chairs the committee and played in an era when the response to rain was to turn to the newspaper crossword and put your feet up.

"I wish they get time to go through my views that bring out the anomalies in the D/L method," Jayadevan has told India's Mail Today. "If the members read it, half the job is done.

"The inherent fear of people for mathematics seems to have helped D/L method being questioned beyond a limit. The D/L system comprises several silly mistakes. But somehow it has managed to create an impression in the entire cricket community that it's highly scientific."

The challenge for the ICC, and indeed the cricket public, is to go beyond nationalistic rivalries, judge the stringency of two complex systems and calculate the benefits they can bring to the game. There are differences and a few can be outlined in laymen's terms.

The battle between Duckworth and Lewis on one hand and Jayadevan on the other has been characterised as a battle between mathematics and engineering. The mathematician pins faith in the purity of the mathematical algorithm; the engineer is prepared to more emphasis on the evidence of what works. In other words, Jayadevan will adopt what one specialist called "intelligent use of trial and error" if it produces a better outcome.

To add to the debate that is raging in statistical circles, the view is growing that Twenty20 behaves very differently from ODIs and requires its own separate tables.

"Like in cricket, the ultimate result of a stroke is more important than how it is played," Jayadevan said. "The most important point regarding the acceptability of a method is its reasonability to adjust targets in a truncated match, and here my system is far ahead. A majority of cricketers and officials are looking for a change and hence it makes sense to give an opportunity to VJD system at least for the next two years."

Both systems recalculate a rain-reduced target based on the number of overs faced and the number of wickets remaining - described as the "resources" still available. There are, however, key differences.

Firstly, the D/L method relies upon a pure mathematical curve that assumes a team's scoring rate accelerates throughout a team's innings. Jayadevan argues that this is no longer the case because of fielding restrictions in the early overs which cause a rush of early scoring before mid-innings consolidation. His tables are adjusted empirically to take this into account. He even claims to take Powerplays into account.

Secondly, and this is where it gets difficult, the D/L method relies upon a single curve which is used to make adjustments to the target. Jayadevan uses two curves: a normal curve to adjust runs already scored, and a target curve to adjust runs still to be scored. The normal curve takes note of both runs scored and wickets lost, but the target curve takes note of only runs remaining.

The greater complexity of Jayadevan's system is less of a problem than it once was. Both methods are computer-based. Calculations are no longer made on the back of an envelope. In theory at least, you can just key in the match details and await the printout.

*08:30 GMT, May 31, 2012 The story has been edited to correct the name of Jayadevan's institute

David Hopps is the UK editor of ESPNcricinfo

RSS Feeds: David Hopps

© ESPN Sports Media Ltd.

Posted by   on (May 31, 2012, 23:55 GMT)

I know exactly how D/L works - whatawicket, perhaps you should take the time to learn?

Posted by   on (May 31, 2012, 9:05 GMT)

well, cricket has always been a game of uncertainties....and specially during natural interruptions, u really can't say anything exact about it. But still, we have seen how many disastrous results have the D/L method produced soo far. But really I don't think that any replacement would be beneficial either. Because even after imposing the new VJD system, you really can't say that its 100% exact. However, it was good to see more people thinking about such topics and are coming out with new ideas.

Posted by smudgeon on (May 31, 2012, 7:38 GMT)

some really crazy suggestions going on here in the comments section! i don't have an alternative method, so i guess it'll either be the status quoe (ie. D/L) or the new kid on the block. either way, the debate will continue until well after limited overs cricket is long gone, and no camp will ever be truly satisfied. me? i just enjoy when they play highlights on the tell from old ODIs when it rains :)

Posted by jmcilhinney on (May 31, 2012, 6:10 GMT)

I think that we're probably all in agreement that there simply is no perfect way to handle rain-affected matches. I agree with @CricketingStargazer's two points, i.e. there is no groundswell against D/L but the ICC have a responsibility to examine feasible alternatives. I see two issues though. 1. The ICC members themselves might have a hard time understanding the intricacies of either method. 2. All you've got to compare the results of the two methods with is intuition and that is rather subjective. It would be interesting to see the comparison though.

Posted by   on (May 31, 2012, 4:40 GMT)

I have this hope that ICC will use a simple modified version of what was used in the beginning: Make the target = the 1st innings run rate x the number of available of overs. With one important change; reduce the number of 2nd innings wickets by the % of overs lost. e.g. if you lose 20% of the overs the team batting second only has 8 wickets to lose (10 - 2).

Posted by Rocketman1 on (May 31, 2012, 1:44 GMT)

A WC did rely on it. 2007 Australia vs Sri Lanka. You can have all the arguments you want, but until the last ball is bowled, you don't know what the outcome will be. Bevan SCG 95/96, Matthews MCG 2010. That WC Final could have gone down to the wire, which would have been great to watch. At the end of the day, a close game is a good game.

Posted by BurningBright on (May 30, 2012, 21:01 GMT)

@Abhisheak Iyer :t hats a good idea..

Posted by CricketingStargazer on (May 30, 2012, 21:00 GMT)

First thing: if someone thinks they have a better system than D/L the ICC has an obligation to look at it. Second thing: despite the wild claim made by its creator, there is no great clamour for an alternative to D/L; in England, where the climate means that D/L is used more than anywhere else, people accept that it is usually pretty fair - absurdities are very rare. There are very little to go on in this article to make a sensible assessment of the new method, but there is one comment that worries me a lot and that is the claim that only runs needed, not wickets, are important in a rain-affected chase. This is patently absurd: if a side needs 120 runs with only 1 wicket left, it obviously has a far harder task than a side that needs 120 runs with 10 wickets left. However, I get the impression from the article that the new method is aimed mainly at T20 and, in this case, rain affected chases are so abbreviated that things change totally, but good batsmen will still svore more than bad.

Posted by makmakmak on (May 30, 2012, 19:23 GMT)

@sookha,

The standard is definitely NOT needed for grounds, the simple reason being both the sides have equal opportunity to hit boundaries on the same ground!.

Posted by whatawicket on (May 30, 2012, 19:04 GMT)

iv no idea how d/l works, who does, but try them both still using the d/l but run the other version running along side the indian version. i dont normally agree with randyoz but even the last remark makes sense. if india were to bring up a referal system that they think thats better go with that also.

Comments have now been closed for this article

TopTop
Email Feedback Print
Share
E-mail
Feedback
Print
David HoppsClose
David Hopps David Hopps joined ESPNcricinfo as UK editor early in 2012. For the previous 20 years he was a senior cricket writer for the Guardian and covered England extensively during that time in all Test-playing nations. He also covered four Olympic Games and has written several cricket books, including collections of cricket quotations. He has been an avid amateur cricketer since he was 12, and so knows the pain of repeated failure only too well. The pile of untouched novels he plans to read, but rarely gets around to, is now almost touching the ceiling. He divides his time between the ESPNcricinfo office in Hammersmith and his beloved Yorkshire.
News | Features Last 3 days
News | Features Last 3 days