Your end or mine?

A look at bowling partnerships

Cricket is replete with stats on batting partnerships. These range from enormous ones: 624 by Kumar Sangakkara and Mahela Jayawardene for Sri Lanka v South Africa at the SSC in Colombo in 2006 remains the highest in all first-class cricket, to the career: Gordon Greenidge and Desmond Haynes still hold the Test 1st wicket partnership aggregate record with 6 482 runs at 47.31, to the niche (inevitably): Rory Burns and Mark Stoneman added 135 for the 1st wicket in both innings for Surrey v Middlesex recently to equal the first-class world record for identical opening partnerships for a team in a match with Kepler Wessels and John Barclay (Sussex v Somerset at Hove in 1979). But what about the bowlers, I hear you ask. Regular cricket-watchers will have heard commentators talk about a pair of bowlers putting together a good partnership. But you won’t have heard any stats on that. It is time to change this. To do this calculation we need ball-by-ball data. Fortunately, we now have this data for over 85% of Test matches. In the 21st century this is thanks to various internet sites, particularly CricInfo, that provide electronic scoring. Pre 21st century matches have been re-scored from scoresheets where they are available. Almost all this enormous task has been done by Charles Davis.

The first step to calculating a bowling partnership stat is to define what it is. Essentially two bowlers form a partnership in the same way two batsmen do. All the runs that the team scores while two batsmen are together are counted towards the partnership. The world record partnership of 624 consisted of 285 runs by Kumar Sangakkara, 309 by Mahela Jayawardene and 30 extras. A partnership ends when a wicket falls, there is a retired hurt or ill, or when the innings ends. For bowlers, the main decision is what runs and wickets to include in the partnership. It strikes me that a bowling partnership should include all the opposition’s runs and all their wickets while it is in operation, like a batting partnership. There is an argument that good bowling can create pressure that induces a run out. We love to measure everything but perhaps “Run Out Pressure Inducer” is beyond even the most dedicated statto or algorithm provider. (On the other hand “ROPI” does seem like a magnificent acronym). So, if a pair bowls 10 overs together and the batting team scores 50 for 3 including 6 extras with the bowlers taking 1-32 and 1-12 plus a run out, I am allocating 3 for 50 in 10 overs to the bowling partnership calculation. With a bit of intricate coding, we can calculate career bowling partnership records. Let us move on to the bit you have been waiting for (preferably with bated breath in these pandemic times). Who are the most successful bowling partnerships?

It should come as no surprise that Test cricket’s most prolific bowling partnership is Jimmy Anderson and Stuart Broad. After the Lord’s Test against New Zealand, they have a total of 1 134 Test wickets between them (Anderson 616, Broad 518). Of these 922 have come in the 122 Tests they have played together. That stat has been tracked for a while. Only Glenn McGrath and Shane Warne have more (1 001) in matches played together. But what are their numbers for when they were bowling in partnership, i.e. Anderson from one end and Broad from the other? England have taken 480 wickets at 22.25 while they have been bowling in partnership. This is well ahead of the second placed, er … this is where it gets a bit more complicated and the good old Rumsfeldian “known unknowns” come into play. The second pair is most likely Curtly Ambrose and Courtney Walsh. They have 287 at 21.04 in 76 matches they played together where ball-by-ball data is available. They also played 19 other Tests together for which we do not have ball by ball data. They took 762 in 95 Tests together and 655 in the matches with bbb data. So, pro rata we can estimate a total of about 334 (i.e. 762/655*287) when bowling in partnership. This is ahead of the 322 at 24.20 by McGrath and Warne and 313 at 28.26 by Anil Kumble and Harbhajan Singh. Trent Boult and Tim Southee are the only other pair who are definitely over 300. They have 303 at 24.55 with power to add. We have complete data for those three pairs. Two other pairs are within range of 300 if you use the pro rata estimate. Chaminda Vaas and Muttiah Muralitharan pro rata to 297 and Wasim Akram and Waqar Younis to 287. So, those pairs may or may not be over 300. It is not really puzzling that McGrath and Warne have significantly less when operating together than the great English pair despite being comfortably ahead in their total in matches together. McGrath and his new ball partner (Lee, Gillespie, et al) would inevitably have knocked over a few before Warne came on. So, McGrath and Warne would have bowled a lot less in partnership than Anderson and Broad.

In terms of averages the standout pair with a minimum of 100 wickets bowling together are England’s 1950s spin twins Jim Laker and Tony Lock. Laker’s 19-90 v Australia at Manchester in 1956 is one of cricket’s best-known stats. In the 1958 series against New Zealand, Lock produced the extraordinary return of 34 wickets at 7.47 in five Tests. While not quite as successful, Laker’s 17 wickets at 10.17 in four Tests in the series were not too shabby. All told in 24 Tests together their bowling partnership produced 158 wickets at an average of 14.40. We need to bring out the old ‘miserly’ to describe their economy rate: 1.77 runs per over. No other 100-wicket plus pair has an average below 20. Muttiah Muralitharan and Sanath Jayasuriya (well, mostly Murali to be honest) are closest at 20.24 (disclaimer alert: 9 of their matches together are missing ball-by-ball data).

Here are the pairs with over 200 wickets in bowling partnerships in Test cricket. “Est” is an estimate for those where full details are not available.

There have been six cases of two bowlers taking all 20 wickets between them in a Test. Have any of these seen all 20 taken while the pair were bowling in partnership? Laker’s match does not qualify as he took one of his 2nd innings wickets while Trevor Bailey was operating at the other end. But George Hirst and Colin Blythe did do it against Australia at Birmingham in 1909 when they bowled all but five of England’s overs in the match. There is no ball-by-ball data for two of the six cases but in both of these other bowlers bowled over 20 overs so it seems likely that some of the wickets would have been taken when our intrepid pairs were not operating together. #knownunknowns.

Will Broad and Anderson become the first pair to reach 500 when bowling in partnership? At their current average of about four per Test they will probably need five more Tests together to get there. Whether that happens will depend a lot on England’s rotation policy. That policy is whirling merrily these days, so it is difficult to tell. Maybe the good people who make the gizmo that measures rotation of the ball for spin bowlers can also produce one for selection policies?

An earlier version of this post made an erroneous reference to Malcolm Nash’s bowling figures in the innings that Garry Sobers hit him for six sixes in an over. It has been removed from here. But, I have left a link to the video for those who have not seen it.

You can watch Sobers’ six sixes here:

Please let anyone you know who may be interested in subscribing that they can do so here: 

If you feel so inclined you can show your appreciation of this column here:

And you can share this post from The Notcher's Natter:

You can leave a comment here: