APPLIED STATISTICS IN DECISION MAKING – STUDENT BOARD ROOM 1
Group 10 – Mohit Choudhary, Simran Bhatia, Atithi Subramaniam, Ravi Parekh
ANALYSIS OF DLS METHOD - Answers to the given questions -
1. In your group’s opinion, do you think the DLS Method is fair and unbiased? Justify your
answer.
Duckworth Lewis (later modified to Duckworth Lewis Stern (DLS) ahead of the 2015 Cricket World
cup) method was adopted by the International Cricket Council (ICC) from 1997. DLS method
considers both wickets and overs as resources, and revises the targets based on the availability of
those resources assuming resources available to both sides are equal. A published table of resource
remaining percentages, for all combinations of wickets lost and whole overs left is given in the case
study.
Let us analyse the application of DLS method in various typical scenarios as may arise during a
one-day cricket match.
Case 1: Completion of Team 1 innings but curtailment of Team 2 innings: [50-over match]
Team Actual Score Overs left Wickets lost Resource percentage
Team 1 250 in 50 overs Play Does not 100%
uninterrupted matter
Team 2 199/5 in 40 overs 10 5 (100 – 26.1)% = 72.5%
So, par score of Team 2 after 40 overs as per DLS method should be 250 x (73.9/100) = 184.75
(Team 2’s par score = Team 1’s score x (Team 2’s resources/Team 1’s resources). In this case, since
Team 2 were on 199, they would be declared winners by 14 runs.
Case 2: Rain interruption in 2nd innings in an already shortened match (50-over match)
Particulars Actual Score Overs Wickets lost Resource percentage
left
Team 1 200 in 40 overs 40 0 89.3 %
Team 2 Start of innings 40 0 89.3 %
Team 2 140/5 in 30 overs 10 5 26.1% (resource %age at the
stage of 2nd interruption)
(Play stopped again)
Team 2 140/5 in 35 overs 5 5 15.4% (resource %age left at
2nd resumption)
(Play (5 overs further reduced
resumes) due to rain)
Team 2 - = 26.1 - 15.4 = 10.7%
Resource
percentage
further lost
Net Available = 89.3 - 10.7 = 78.6 %
Therefore, revised target for Team 2: 200 x (78.6/89.3) = 176.03 (~176 to win). Hence, now 36 runs
required in 5 overs. Note that here DLS method turns out to be quite fair, as it takes into consideration
the number of wickets lost and the number of overs left to calculate the revised target.
Case 3: Interruption by rain to Team 1’s innings: [50-over match]
Score Overs Wickets lost Resource percentage left
left
Team 1 100/2 in 25 overs 25 2 60.5% deprived, therefore
resources left = (100- 60.5)%
(When it first
= 39.5%
rained)
Team 2 0/0 25 0 66.5%
(At the start
of 2nd innings)
Here, Team 2 has 27% more resources than Team 1. So, for the target there would be an upward
revision on Team 1 score by 27% of 245 (G50 i.e., average score in a 50-over innings) = 66 runs
more than Team 1 score. Therefore, target for Team 2 will be 166 runs in 25 overs (under the
Standard Edition). Under the current scenario in International Cricket, the same target would be
computed taking resource values and benchmark target score from a computer program.
Case 1 Analysis: The above result was quite fair as Team 2 was very strong when the play was
interrupted and most likely would have won the match. Average Run Rate (ARR) method in this case
would have given the par score as 201 – So probably, Team 2 would have lost the match if any other
methods were to be followed.
Case 2 Analysis: The DLS method proves to be fair once again, as it takes into consideration the
number of wickets lost and the number of overs left to calculate the revised target. Of course, how
high the target should be for Team 2 is still a debatable issue given differences in qualitative
characteristics of the two teams.
Case 3 Analysis: It should be noted that here the advantage to Team 2 from knowing in advance of
the reduction in their overs is neutralized. Other target resetting methods make no allowance for this
advantage. They would set the target to a lower score compared to that set by DLS and this would be
clear injustice to Team 1 who were pacing their innings to last 50 overs when it was interrupted. DLS
counters this by setting for Team 2, a higher target than the number of runs Team 1 actually scored.
Therefore, it would be fair to say that the DLS method is relatively fair and unbiased compared to
other available methods as it takes into consideration the majority of factors and attempts to
neutralize any prior advantages and team would have in application of the method. Although, it does
suffer from a few shortcomings as explained below:
The DLS method has been criticized on the ground that wickets are more heavily weighted than
overs when determining the revised target / par score.
The D/L method also does not take into account variation in field restrictions which would apply in
case of a full 50-over match.
Even though Steve Stern felt that he had improved on the D-L method by adjusting the formula to
reflect changing realities in high-scoring ODIs and T20 matches, it is still a matter of concern
whether the DLS model has kept up will be able to keep up with the dynamic changes brought
about in the game of Cricket by the shorter formats i.e., T20 and T10 matches.
2. Compare the DLS Method with the previously adopted Average Run Rate (ARR) Method and
Most Productive Overs (MPO) Method. Which of the three do you think is the fairest to both
teams? Justify your answer.
The Average run rate method was adopted in the 1960’s until it was replaced by the Most productive
over method in 1991, with DLS method being adopted 1997 onwards.
Average Run Rate (ARR) method – Example - Team 1 made 250 in 50 overs, and Team 2’s innings
is reduced to 25 overs.
Target for team 2 – [(Team 1’s total /Over used by team 1) X overs available to team 2 ] + 1
So, target = (5 x 25) + 1 = 126.
As can be clearly seen above, the ARR method is usually in favour of the team batting second as:
- It is easier to maintain a given run rate for a reduced number of overs as compared to doing the
same for higher number of overs.
- It does not take into account the number of wickets lost, but simply how quickly the second team
was scoring when the match was interrupted.
Unfair match example - Third final of 1988-89 World Series cup - Australia scored 226 from 38
overs, WI were 47 for 2 in 6.4 overs when rain curtailed play. As per ARR they were given a revised
target of only 108 runs in a reduced game of 18 overs.
WI won with 4.4 over remaining.
It can clearly be seen in the above example ARR was heavily biased towards Team 2. Australian fans
loudly booed this unsatisfactory conclusion, which was highly criticised by the media and Australia's
captain Allan Border as well.
Most Productive Overs (MPO) Method - New target for Team 2 from their total of X overs = Runs
scored by Team 1 in their highest scoring X overs + 1.
While the ARR method is heavily biased towards team batting second, the MPO method Is biased
towards the team batting first.
- It penalizes Team 2 for good bowling by ignoring their best overs when setting the target.
- Takes no consideration of wickets lost by the 1st team
- If the least productive overs were maiden overs, Team 2 would be left with lower number of overs to
play without reduction in runs required to win.
Unfair match example - Eng vs SA World Cup 1992 – Semi-finals [50-over match]
England scored 252/6 in 45 overs; Rain interrupted play when South Africa were 231/6 in 42.5 overs.
Two overs [as per MPO method, 2 least productive overs yielding 1 run in total] were lost due to rain
and so South Africa required 21 off the last ball to win the match [further, the scoreboard incorrectly
flashed 22 runs to win] (Example given in the case study in detail).
It can clearly be seen that in the above example MPO was heavily biased towards Team 1.
To summarize, despite a few shortcomings DLS appears to be the best method available among all the
available methods for the purpose. It attempts to arrive at a formula that takes into account as many
variables as possible and also tries to properly reflect the efforts of both teams in its projections. That may
be the reason why although it has been updated a few times and probably will continue to be updated in
future, it is generally considered the most accurate system used in international cricket so far.