Duckworth-Lewis-Stern (DLS) Method Explained: Rain Interruptions in Limited-Overs Cricket
The Duckworth-Lewis-Stern (DLS) method is the mathematical formula used by the ICC to adjust targets in rain-interrupted limited-overs cricket matches. Originally developed by statisticians Frank Duckworth and Tony Lewis in 1997 and adopted by the ICC in 1999, the method was revised by Steven Stern in 2014. DLS replaced the simpler Average Run Rate method that had produced unfair results, most famously in the 1992 Cricket World Cup semi-final between England and South Africa. The method models the runs a batting team is expected to score based on wickets in hand and overs remaining.
When rain or bad light interrupts a limited-overs cricket match, the two teams may face different numbers of overs. Simply reducing the target proportionally — using the Average Run Rate method — produces unfair outcomes, especially when the interruption affects the death overs (the most productive for batting) differently than the early overs. The Duckworth-Lewis-Stern method was developed to resolve this by modelling the relationship between wickets in hand, overs remaining, and expected runs — the two resources a batting team has available to score.
How DLS Works: Resources and the Table
DLS is built on a table of 'resource percentages' — the proportion of total expected runs a team can score from a given combination of wickets remaining and overs remaining. A team starting an innings (10 wickets, 50 overs remaining) has 100% of resources. A team that has used 20 overs and lost 3 wickets (30 overs, 7 wickets) has a different (lower) resource percentage. When overs are lost due to rain, the DLS method calculates how much resource was lost and adjusts the target accordingly. If the team batting second has fewer overs than the team batting first, the target is reduced by the percentage of resources lost. If rain reduces the first team's innings after they have batted, the target for the second team is set based on what the first team would have been expected to score using their full 50 overs. The revised target is sometimes higher than the first team's score — which causes confusion but is correct: it reflects that the first team's actual score was made under sub-optimal conditions (e.g., they scored conservatively early knowing the rain would not interrupt, and rain cut short an expected acceleration).
History: The 1992 World Cup and Why DLS Replaced ARR
The most famous failure of the pre-DLS Average Run Rate method was the 1992 Cricket World Cup semi-final between England and South Africa at the SCG. South Africa needed 22 runs from 13 balls when rain stopped play. Under the ARR rules, when play resumed with 1 ball remaining, the target was recalculated using the Most Productive Overs formula — South Africa needed 22 from 1 ball (later corrected to 21 from 1 ball), an impossible target. England progressed. The result was widely considered the most unfair application of a rain rule in major cricket history and directly motivated the development of the Duckworth-Lewis method, which was adopted by the ICC in 1999. The 2014 revision by Steven Stern updated the resource percentages table with modern data from thousands of ODI matches, improving the model's accuracy for high-scoring matches (early versions of DLS slightly underestimated the pace of scoring in the 300+ era).
Frequently asked questions
Why does the DLS method sometimes set a higher target than the first team's actual score, and is this fair to the team batting second?
The DLS method sets a target for the team batting second that is sometimes higher than the first team's actual score, and this is one of the most frequently misunderstood aspects of the method. Understanding why requires understanding what the first team's actual score represents under conditions of a rain interruption. Scenario: Team A bats first in a 50-over ODI. After 25 overs, rain arrives, and the umpires reduce the match to 40 overs per side. Team A had scored 120 from their first 25 overs, a run rate of 4.80. When the match resumes at 40 overs, Team A now knows they only have 15 more overs (overs 26–40) instead of 25. They adjust their strategy: they might sacrifice wickets to accelerate, or play very aggressively from the 26th over onward. Suppose they score an additional 120 from overs 26–40 (run rate 8.00), finishing at 240 from 40 overs. Why the DLS target might be above 240: Team A scored their last 120 at 8.00 per over knowing exactly how many overs they had. They deployed their most aggressive batters optimally, knowing the death-over situation. Team B, if given a straightforward target of 241 to chase in 40 overs, would actually face a slightly different match context — they knew from ball one they had 40 overs, potentially allowing more conservative early-over play. The DLS calculation adjusts for what Team A would have expected to score with their full 50 overs from the outset — and if the rain interruption caused them to bat more efficiently than they would have over 50 overs (because they compressed their scoring into fewer overs), the DLS target reflects the full 50-over expectation, which might be higher than the compressed actual score. In practice, the DLS target for Team B exceeding Team A's score typically arises when rain cuts the first team's innings after they had a conservative start (a low run rate in the first 20–25 overs) followed by a heavy reduction in available overs. Their conservative start is not punished in DLS terms (the target is based on expected scoring, not actual), so Team B faces what would have been an optimal 50-over score from a team that started well. Critics argue this creates unfairness for Team B — particularly in T20 cricket where the DLS calculations are compressed and small errors in the resource table produce larger percentage errors in the target. Defenders argue it is fairer than ARR and is the best available approximation given real-world data on how teams score.