Following on from our initial batting analysis of the all-rounders fighting for the spot in our ‘Best ODI XI’, here we examine their abilities with the ball in hand – before combining batting and bowling to finalise our selection.

As it stands, Ben Stokes was the highest-rated batting all-rounder according to our Batting Index – however, he has already been selected as an out-and-out batsman in our number five slot, meaning even if he is the all-rounder with the highest overall All-Rounder Index he will not be eligible for selection. However, for means of comparison, he remains part of our shortlist.

Just to recap, the criterion for making our shortlist was to have scored at least 800 runs and taken 30 wickets since January 1 2016. The men who did so were: Stokes (2138 runs and 38 wickets), Shakib Al Hasan (1925 and 54), Jason Holder (1449 and 88), Sikandar Raza (1409 and 39), Mohammad Nabi (1350 and 78), Kevin O’Brien (1206 and 36), Thisara Perera (1153 and 50), Sean Williams (1086 and 34), Jimmy Neesham (1048 and 41), Marcus Stoinis (1046 and 33), Moeen Ali (986 and 52), Hardik Pandya (957 and 54), Gulbadin Naib (889 and 50), Imad Wasim (876 and 35) and Rashid Khan (869 and 120). For clarification, all bowling statistics are based on games against Test-playing nations.

Economy, Average and Strike-Rate

Our initial assessment will be looking at their respective economies (how many runs they concede on average per over), averages (how many runs they concede per wicket taken) and strike-rates (how many balls they take to get a wicket).

The first notable takeaway from this is the positive correlation demonstrated by the data – i.e. as a player’s average increases so too does his strike-rate, which makes sense when considering that if he is taking more deliveries to get a wicket (higher strike-rate) then he’ll be conceding more runs in the meantime (higher average).

The absolute king of this section is Rashid Khan, the leg-spinner possessing a ridiculously low average of 18 alongside a strike-rate of 26 – both of which are the lowest of anybody in the group by a considerable margin. The 21-year-old has developed into one of the best limited-overs bowlers on the planet since his emergence in 2015, and that statement is backed up by the figures here (not to mention the 120 wickets he has taken in this time). 

Behind the Afghani superstar when it comes to average is fellow countryman Nabi at 29.52, whilst second for strike-rate is New Zealander Neesham with a wicket every 32.2 deliveries. At the other end of the spectrum, Moeen is neither a wicket-taking threat nor tight in his bowling, being the complete antithesis to fellow spinner Rashid by holding the highest average (56.32) and strike rate (61.8). Imad is the only other player with a strike-rate of around 60, but his average of 48.77 is lower than his English counterpart.

Moving onto their economy, it is once again the two Afghan spin twins who are the best represented.

Rashid is able to limit his overs to go for, on average, just 4.14 runs, whilst Nabi is not far behind with 4.21. There are three other bowlers who possess an economy of fewer than five runs per over (RPO): Williams (4.64), Raza (4.73) and Imad (4.87) – and it is telling that all five are spin bowlers. 

The role of spinners is continually becoming more important, especially in limited-overs cricket, with a high-quality spinner being vital to a side if they wish to be successful. These five have demonstrated the constraints that slower bowlers can place on batting sides, and their presence in the highest echelon here is an indicator of how well spinners can affect a game. In fact, the lowest position of any of the seven spinners is ninth place – Moeen, with an economy of  5.46. This shows how well spinners have adapted to the modern-day white-ball game, and how effective they are at limiting the opposition’s run rate. The fact that the highest place of any pace bowler when it comes to economy rates is seventh paints its own picture.

Elsewhere, the most expensive bowler is Australian Stoinis, who concedes 6.16 RPO on average, slightly ahead of Perera at 6.09 and Neesham with 6 exactly – a somewhat surprising figure given his position as the player who takes wickets with the second-most frequency of the group. 

Are they better against right- or left-handers?

Just like when analysing the batters in our side against pace and spin, here we have compared their effectiveness when bowling against both right- and left-handers to try and highlight any possible glaring strengths or deficiencies they may have when bowling to either kind of batter. Initially, we will plot their respective average against both forms of batsman – as is below.

When we do so, there is no distinctly clear pattern or trend emerging from the data as was the case when comparing average and strike-rate, although it would probably be fair to conclude that when a player’s average against right-handers goes up it does so against left-handers too. 

There are no cases of bowlers having a striking favourability against either sort, with Williams the man to hold the greatest difference as he averages 19.31 runs lower against right-handers than he does left-handers. Stokes holds the second greatest average difference, again in preference to right-handers, averaging 30.29 compared to 48.42, whilst Raza and Neesham both possess averages differences of around 12 runs higher against left-handers – 12.04 and 12.76 respectively.

In order to give ourselves a clearer picture of a bowler’s all-round ability, adding both averages together allows us to form a cumulative total.

When doing so, it, perhaps surprisingly, highlights Stokes as the weakest bowler within the whole dataset at a cumulative average of 78.71, accompanied only by Perera in possessing a cumulative average greater than 70. Rashid is the man to hold the lowest average at 40.4, with Nabi not far behind at 43.34. There should also be mentions for Holder, the West Indian captain only marginally behind Nabi with his cumulative total of 44.01, and Shakib, accumulating an average of 45.13.

Taking this one step further, whereas with the batsmen we subtracted their average difference from their cumulative total  – as the aim is to have as high an average as possible – here we can add their average bowling difference to their total in order to form an Overall Average, due to the aim being to have as low an average as possible. This then rewards those who are as good against both types of batsmen and hampers those who show clear favouritism to one kind – and it changes the table quite significantly.

Rashid loses his place at the top as he slips to fourth with Nabi also dropping down, falling one place to third, whilst Shakib moves to fifth by virtue of his difference of 9.09. On the rise are Holder and Pandya, both boosted by a minimal difference between their right- and left-handed averages (Holder of 1.11 and Pandya of 0.52) to move up to first and second place respectively. 

Do they get the best batsmen out?

Now, here we get to the final constituent metric when it comes to highlighting their ability in a particular aspect of bowling: namely, their success against the best batsmen.

Although career averages are of course very worthy and telling in their own right, they can perhaps be skewed by a bowler’s ability to dismiss tail-enders as opposed to genuine batsmen. Therefore, we now dive into their skills against the top level of batsmen the game has to offer, as we compare respective averages against opposition top five batters, done so by working out their mean average against numbers one to five.

When doing so, as was the case in the previous section, it is once against Stokes who is not reflected in the best light. His average of 62.22 is by far the worst of the whole set, almost 15 runs higher than Stoinis’ 48.63 in the penultimate position. Another surprising position is Rashid in twelfth, his average of 39.9 eye-opening given his perceived world-class ability to bamboozle batsmen. 

Looking at those who fare best, it is the experienced Bangladeshi Shakib who appears to have the best ability at containing higher-order players thanks to a meagre average of 21.31, although both Pandya (27.06) and Holder (28.51) can be proud of their sub-30 averages too – the only three players in the group capable of boasting such a statistic. Imad (30.85) in fourth and Moeen (30.88) in fifth may surprise a few given their high overall averages and strike-rates as previously discussed.

On the graph below, the red area indicates a player’s average against top-five batsmen, with the green indicating their ‘normal’ average in this period. Therefore, those that have green showing perform better against the perceived better players than they do on the whole overall – the more green showing means the more they outperform their expected levels against the top batsmen.

Bowling Selection

When taking all of these into account, as with the batsmen previously, we have devised a unique Index to provide a measure of just how good each player is. Our index takes into account all of the areas discussed previously and generates a number that is a reflection of all of these statistics. Whereas when batting the aim is to produce as high an index as possible, when it comes to the Bowling Index the idea is to have as low a score as possible. 

Therefore, according to our Index, the best bowling all-rounder is Rashid Khan. His Index of 136.19 is almost 20 runs lower than Nabi in second place – a testament to the level of cricketer that Afghanistan produce that they can have the top two bowling all-rounders. 

Holder (160.02) and Pandya (163.96) are third and fourth, whilst Shakib occupies fifth place with his Index of 166.07. Looking at the other end of the scale, it is English duo Moeen and Stokes who sit 14th and 15th respectively, Stokes out on his own with an Index of 261.75. So, how do these findings impact the overall selection?

All-round selection

Usually, the traditional way of judging an all-rounder’s competence is to subtract his bowling average from his batting average, and the higher the remaining figure the higher all-round standard the player is of. If this was the way we were going to select, the all-important chart can be seen below.

With this method, it is Stokes who is displayed the most favourably thanks to his figure of 6.82 squeezing out Shakib on 6.74. However, we don’t do things traditionally here at TCA.

Our method is the same, but instead of using generic averages we are using our Batting and Bowling Indexes – and when doing so, it shakes things up significantly.

When subtracting the Bowling Index away from the Batting Index to form an All-Rounder Index (ARI), Stokes plummets down the table to tenth by virtue of his very high Bowling Index. Instead, it is Shakib who reigns supreme, his ARI of 39.77 being over ten points higher than Pandya (29.33) in second – himself a significant riser from 12th place in the Average Difference table. Rashid retains third spot, whilst Holder moves from 14th in the Average Difference table to fourth in the Index Different table.

On the whole, Shakib is clearly the strongest ODI all-rounder in the game over the past four years. His run-making ability and wicket-taking nous against the best players in the world have highlighted him as the stand-out performer – meaning he is perfect to occupy the number six slot in our side.

Stay tuned in the coming days as next up we will be analysing the best wicketkeeper-batsmen the game has to offer. 

Current Team

  1. Rohit Sharma (India)
  2. David Warner (Australia)
  3. Virat Kohli (India)
  4. Ross Taylor (New Zealand)
  5. Ben Stokes (England)
  6. Shakib Al Hasan (Bangladesh)