The Web


World Cup 2003
Match Reports
Graphical Analysis
WC Format
Fantasy Cricket
Discussion Groups

Home > Cricket > World Cup 2003 > Columns > Srinivas Bhogle

The Duckworth/Lewis factor

March 06, 2003

Now that we are at the Super Six stage in the World Cup, with an additional day to play in case of rain interruptions, the Duckworth/Lewis (D/L) factor is likely to recede in the background.

This might therefore be a good time to reflect on the D/L method, and wonder why everyone seems to really enjoy rubbishing the method.

I am frankly puzzled. The D/L method is still about the best that we have to reset targets in rain-affected ODI matches. I have argued that Jayadevan's method might just score over the D/L method -- but not by much.

Let's look at different applications of the D/L method in the 2003 World Cup.

We must start with South Africa's confusion about whether 229 or 230 was the winning target at the end of 45 overs.

First, the D/L method itself can hardly be blamed for this confusion. Blame the South African team management; blame the organisers for not displaying the D/L par score on the scoreboards (the par score is always displayed on scoreboards in England), but not D/L! In fact, if I understand D/L well enough, the method sets impeccable targets after more than 30 overs are played, more than 3 wickets are lost and the scores hover in the 225-275 interval.

It's likely that South Africa made at least two mistakes. First (if we recall Shaun Pollock's statement), they probably confused the par score for the winning score (the par score is the score at which the rival teams tie; if the par score is 229 or 229.3, the winning score is 230).

Second (and Eric Simons has said this too), South Africa blundered if they thought that the target of 229 (or 230) was fixed. D/L par scores, that depend on how many balls and how many wickets remain, are changing practically on a ball-to-ball basis (that's why captains are provided printouts). Boucher should always have looked for a 'safe' run -- off the last ball or indeed off every ball bowled to him.

Let's next consider the other South Africa match against New Zealand. Replying to South Africa's 306/6, New Zealand faced a couple of interruptions till they happily discovered after another longish interruption that only 44 were needed off 51 balls to win! The popular view was that this target was easier to reach at that stage of the match. The average cricket fan felt that if there hadn't been that last big interruption New Zealand would have found it more difficult to win.

The D/L counter-argument is that, having lost only one wicket, New Zealand were anyway well placed to win that match -- that's why the apparently "easier-to-get" revised target was actually fair.

Here we run into a well-acknowledged weakness of the D/L method. If teams are chasing big targets, and there is the prospect of rain, it is more sensible not to lose wickets and score at just an above average rate (if the asking rate, for example, is 6.1, it's enough to score at 4.75 or 5 an over for the first 20-25 overs). Everyone knows this weakness. Certainly, Stephen Fleming knew it, and his team played brilliantly to exploit it! The only way to counter this is to strike hard, set truly attacking fields and capture those first three wickets quickly. South Africa and Shaun Pollock failed to do this (let's not forget though that simple catch that Boucher dropped. If Boucher had held that one, and SA had got another wicket from somewhere, NZ might in the end have required something like 70 runs off 51 balls).

Consider a third match: England vs Namibia with A J Burger blazing away. England were in real trouble there and there were at least five uncomfortable overs (overs 25-30) when rain would have meant certain defeat for England. Fortunately (or more probably because Nasser Hussain knew his D/L well, unlike Pollock), England quickly got a few more wickets to seal the game in their favour.

How would Jayadevan(J)'s method have fared in these three examples? In the South Africa vs Sri Lanka match, South Africa would have required 232 to win in 45 overs using the J method (instead of D/L's 230). For the South Africa vs New Zealand match, I am almost certain that the J method would have set NZ a much stiffer target (I don't have the exact details of all the interruptions and the J calculation in case of multiple interruptions is much more complicated than D/L).

I remember Sunil Gavaskar, who is the chairman of the ICC Technical Committee, remarking on ESPNStar's "Follow Through" programme, that he thought the J method set slightly more realistic targets. So it seems almost certain that the D/L method is heading for a rather severe review.

I don't know if D/L will get the axe. D/L could modify their method somewhat, but they can't really do too much more with their exponential decay functions. We must also remember that the Jayadevan method is still not truly tested. There could still be weaknesses that might show up once the method is introduced on the cricket field. If I might make a guess, the likely weaknesses in the J method will probably show up in innings-end scenarios or in scenarios where a heap of wickets fall in the space of an over or two.

Still, I cannot understand the unwarranted criticism of the D/L method. It's almost fashionable for cricket commentators to say that D/L is rubbish. It isn't. The average run rate method was rubbish, the 1991-92 'most productive' overs' method -- that haunted that year's World Cup -- was bigger rubbish. Duckworth and Lewis were the first to come up with something that was much fairer.

Perhaps the J method could become the eventual winner. Even if Jayadevan wins, he has much to be grateful to D/L. There was a time when I found myself acting as the moderator in e-mail exchanges between Jayadevan and Duckworth/Lewis. I recall that early versions of the J method were rather poor. If the method is now a serious contender it is in no small measure because Duckworth/Lewis pointed out many major deficiencies. Jayadevan, to his credit, was skilful enough to quickly react to this feedback and correct most of these deficiencies.

The best way, however, to defeat rain interruptions might be to do what we did in 1999 in England: convert "one-day internationals" (ODI's) to "limited-over internationals" (LOI's) and allow matches to spill over to the next day. It would be even better if there were no unnecessary political boycotts. Then we might have had England and South Africa in the Super Six instead of Zimbabwe and Kenya!

Related articles:

The dummy's guide to Duckworth-Lewis

Is Jayadevan's proposed method better than the Duckworth/Lewis method?

More Columns

Share your comments

 What do you think about the story?

Read what others have to say:

Number of User Comments: 28

Sub: RE:duckworth-lewis method

You have not realised the pressure situations. The bottom order of any team is not as strong as the top. For example if a team ...

Posted by Deepak

Sub: Flaw / No flaw

I accept that D/L method is the best but the flaw is that teams must be prepared to be ahead every ball from the 25th ...

Posted by Deepak

Sub: Why all this complaining?

I very much agree with the D/L method being used in cricket. If a match is interrupted it's not the same anymore. It's more like ...

Posted by Anand

Sub: The trouble with weather prediction

There is a fundamental flaw in the D/L system which is, perhaps, somewhat inevitable. When it is used to set a target before an innings ...

Posted by Prasanna Ganesan

Sub: J vs D/L

I understand your expressions of admiration for the D/L method when it is under attack unfairly. But, while doing that, you are slightly diminishing Jayadevan's ...

Posted by Chandra Mouli


Article Tools

Email this Article

Printer-Friendly Format

Letter to the Editor

Copyright © 2003 rediff.com India Limited. All Rights Reserved.