Stats Analysis

An explainer on the performance ratings used to measure the best ODI bowling performances of all time

Pitch quality, batting quality, importance of the match and more are taken into consideration

Pat Cummins and Jhye Richardson celebrate a wicket, India v Australia, 5th ODI, Jaipur, March 13, 2019

Pat Cummins and Jhye Richardson took identical figures of 3 for 37 against a very strong Indian batting line-up in Ranchi in March this year  •  Getty Images

Wickets Captured
This is the base measure. The computations are such that a nine-wicket haul will get maximum points. Currently, the most wickets taken in an innings by one bowler is eight and we don't know how long the wait will be until a nine-wicket haul. This is unlike the "Runs scored" parameter, in which the maximum runs have already been reached and is capped at 200.
Even though Chaminda Vaas took 8 for 19, Rashid Khan's 7 for 18 trumps Vaas' spell on this measure since Vaas' spell was against the lower-ranked Zimbabwe and Rashid's spell was against the (surprisingly weaker but) traditionally more powerful West Indies. This takes into account the fact that traditionally strong teams remain tough even though they might have gone off the boil.
SWQ (Spell Wkt Quality)
This important measure takes into account the quality of the batsmen dismissed and the score at which they was dismissed. Thus, it can be seen that dismissing a top batsman like Virat Kohli for a low score carries a lot of weight. However, it is also important to not allow the better batsmen to bat for a long time. Dismissing Kohli at, say, 76, still carries decent weight.
On this measure, Vaas has the highest value. The sheer number of wickets he took carried the day for him. He just gets ahead of Waqar Younis, whose 7 for 36 against England in 2001 is the only other spell with SWQ greater than 50.0.
Bowling accuracy
Unlike the batting strike-rate measure, this is a complex measure, since this combines the base economy rate and the ratio of the bowler's economy rate to the team economy rate. The base economy rate has a higher weight. The idea is that a spell of 10-0-45-1 has a lot going for it in this measure if the other team scored 350 but is very poor if the other team scored 160.
Phil Simmons' spell of 10-8-3-4 against Pakistan in 1992 has the highest value on this measure. Courtney Walsh's 4.3-3-1-5 against Sri Lanka in 1986 and Wasim Akram's 7.2-2-4-2 against India during 1986 follow closely.
Bowling SR
Like the bowling-accuracy measure, this also takes into account both the absolute strike rate of the bowler and the strike rate of the bowler relative to the rest of the team. This ensures that whether the batting team lost five wickets in 50 overs or ten wickets in 25 overs, the analysis is fair. In both the accuracy and strike-rate measures, what the bowler achieved in absolute terms and what he achieved in relative terms mean a lot.
The bizarre spell of Ajay Jadeja, 1-0-3-3, against England in Sharjah in 1999 takes top position on this measure. Robin Singh's spell of 5-0-22-5 in the match against Sri Lanka in Guwahati in 1997 is also highly rated.
Batting quality
This is based on the career-to-date bowling averages and economy rates. This ensures that the rating values calculated do not vary for future computations. It also allows for career swings, which are common for most bowlers.
Recent Indian and England batting line-ups have been among the greatest of all times. Pat Cummins (3 for 37) and Jhye Richardson (3 for 37), in the recent India-Australia match in Ranchi, faced one of the toughest batting line-ups of all time. Kuldeep Yadav's magnificent spell of 10-0-25-6 against England last year takes second position.
Pitch quality
Similar to the batting analysis, this is based on the RpW of the particular match. There is no use going into any history since the RpW values move drastically in this regard across time.
Think of the plight of the bowlers in the match in Jaipur between India and Australia in 2013 in which six wickets fell for 721 runs - a match RpW of over 120. R Vinay Kumar, who took 2 for 73 in this match, gets the highest points value for this measure.
Innings status
This parameter is based on the enormity of the task faced by the bowler at the beginning of the innings. For the first innings, the bowlers face reasonable challenges. Across the years, the bowling par scores have ranged from 191 to 234, and this value is used as the target. This is the score that gives the bowling team a 66.7% chance of success. For more details on this, please refer to my recent article on redefining par scores in ODIs. However, for the second innings, the targets have ranged from 36 to 482 and the measure takes into account the target ahead. Unlike batting, this cannot be a dynamic measure since there is no record of when the bowlers bowled for over half the matches.
In an ODI against England in Bristol in 2003, Zimbabwe were defending 92 and Heath Streak took 4 for 21. This heroic performance, albeit in a losing cause, got the maximum value on this measure. Of course, Douglas Hondo, while defending a total of 35 against Sri Lanka, might have faced the toughest of challenges, but cases like these are excluded. A minimum of three wickets have to be taken.
TS/L/R (Team Strength/Location/Result)
This measure is a combination of three factors: the relative team strengths, the location and the result. In addition, very close results, such as win margins of under ten runs, are considered and the concerned bowlers get additional credit. Close results like one- or two-wicket losses are also rewarded a little bit. Ties carry more credit than no-result matches since a tie is considered to be a result.
In 2002 in Delhi, England scored 271 and won by two runs thanks to Ashley Giles, who took 5 for 57. Giles tops the list on this measure.
Match Importance Index
Multiplying factor in the range of 95% to 125%. The 11 World Cup finals carry the maximum index value.
Parameter weights
The overall weights for bowling spells with two wickets and more (14,796 in total) are given below. In view of the importance of wickets in the final ratings work, it is not possible to relax this criterion any further. The weight percentage will be distorted a lot, especially for the first two parameters.
Wkts: 13.3%
SWQ: 8.5%
Acc: 12.5%
StRt: 4.9%
BatQ: 20.3%
PQI: 11.8%
Tgt: 12.9%
T/L/R: 15.9%
It should be remembered that the weight of SWQ is somewhat low but the real influence is much higher for spells in which four or more wickets are taken.

Anantha Narayanan has written for ESPNcricinfo and CastrolCricket and worked with a number of companies on their cricket performance ratings-related systems