Here we continue our journey of determining and comparing cricket’s greatest all-rounders.
Following on from part one, the average runs scored per Test of the top 20 batsmen I assessed was 76.82 and the average number of wickets taken per Test of the top 20 bowlers I assessed was 4.12.
The ratio difference of runs to wickets is 18.65, meaning the true value of each wicket might actually be closer to 20 runs.
Interestingly, the range of runs scored per Test of the 20 batsmen sampled ranged from Brian Lara (91.24) to Alec Stewart (63.63).
The top five were: Lara (91.24), Kumar Sangakkara (89), Matthew Hayden (83.74), Virender Sehwag (82.56) and Rahul Dravid (81.02).
Sachin Tendulkar averaged 79.61 runs per Test (exclusive of total innings).
The rate of wickets per Test of the 20 bowlers sampled ranged from Muttiah Muralitharan (6.02) to Chaminda Vaas (3.20). The top five were: Muralitharan (6.02), Dale Steyn (5.072), Dennis Lillee (5.071), Richard Hadlee (5.01) and Shane Warne (4.88).
I also noted that back in the days when five Test series were far more prolific than they are today, the magic figure for a batsman was to reach 400 runs in the series, while the magic figure for a bowler was to take 20 wickets.
So if you either scored over 400 runs or took 20 wickets, you could be pleased with yourself at having a very good five Test series.
That’s a ratio of 20 runs for every wicket (20 x 20 = 400), so now our methodology is complete.
To find an all-rounder’s productivity quotient (PQ), you add the total number of runs scored plus number of wickets taken (x 20) divided by the total number of matches.
Using this methodology, the top 15 all-rounders in Test cricket based on PQ (x 20) are:
1. Gary Sobers (136.90)
2. Aubrey Faulkner (135.76)
3. Richard Hadlee (135.56)
4. Imran Khan (126.44)
5. Ian Botham (126.08)
6. Chris Cairns (123.87)
7. Vinoo Mankad (121.57)
8. Trevor Goddard (121.37)
9. Jack Gregory (118.53)
10. Keith Miller (115.60)
11. Jacques Kallis (115.27)
12. Alan Davidson (114.72)
13. Richie Benaud (113.67)
14. Shaun Pollock (113.00)
15. Eddie Barlow (110.53)
Wasim Akram rated a PQ of 107.48 and Kapil Dev rated a PQ of 106.32.
Wasim probably loses ground with his batting, while Kapil loses ground with his bowling (both compared to other all-rounders).
It seems to me sanity has been restored to the world! It is only fitting that Hadlee should occupy the all-time number three Test all-rounder’s position. Most cricket fans would agree he is the greatest bowling all-rounder in the history of the game, and this methodology recognises this.
Cairns’ high ranking is surprising until you look closely at his record, it is very good. I guess whichever system you use, it will throw up anomalies.
Gregory pips Miller as Australia’s best all-rounder. Their batting averages are virtually identical, but Gregory averaged 3.5 wickets a Test to Miller’s three, thus edging ahead of him.
On the other hand, Gregory only played 24 Tests to Miller’s 55. However, Miller is still Australia’s best all-rounder.
As is often been mentioned, Miller didn’t care about stats and mostly operated on idle. But he could do the unbelievable and incredible when the occasion demanded it.
Australia’s top 12 Test all-rounders are: Gregory (118.53), Miller (115.60), Davidson (114.72), Benaud (113.67), George Giffin (106.39), Monty Noble (105.17), Bobby Simpson (101.44), Ray Lindwall (99.38), Charlie Kelleway (94.69), Shane Watson (93.02), Armstrong (92.06), Macartney (86.60).
(Note Watson at no.10.)
Fans of Kallis will be perplexed at how he has lost ground to Sobers.
The clue might lie in the fact that Sobers averaged 86.37 runs per Test and took an average 2.52 wickets per Test, compared to Kallis’ figures of 80.05 runs per test and 1.76 wickets per Test.
Clearly, Sobers was busier, and therefore more productive.
So that’s the greatest Test all-rounders according to my methodology of finding the productivity quotient for each player.
But what all about first-class cricket? Who is the best here? Let’s take a look.
I warn readers some of the results here will shock. But I also remind you that playing standards varied enormously from country to country and from decade to decade.
Nevertheless, it is a fascinating education all the same. Here’s the top 20.
1. Aubrey Faulkner (130.05)
2. Vijay Hazare (128.73)
3. Gary Sobers (128.39)
4. WG Grace (126.89)
5. Mike Procter (125.38)
6. Trevor Goddard (122.73)
7. Richard Hadlee (122.37)
8. Warwick Armstrong (121.93)
9. Jack Gregory (121.25)
10. Richie Benaud (118.22)
11. Vinoo Mankad (116.88)
12. Imran Khan (113.90)
13. Wilfred Rhodes (111.76)
14. Jacques Kallis (109.86)
15. Shaun Pollock (109.47)
16. Wasim Akram (108.95)
17. Chris Cairns (108.94)
18. Keith Miller (106.74)
19. Monty Noble (106.67)
20. Ian Botham (106.56)
Wow, where did Hazare come from? There’s no doubt he was a top-shelf cricketer, but perhaps standards in India in the 1940s and 50s were not of the highest order. Nevertheless, the stats support his quality as a cricketer.
The productivity quotient also tells us that the fuss over WG Grace is justified.
Sobers remains the greatest Test all-rounder, but Grace (from the dawn of international cricket 1877-95) and Faulkner (from the golden age of cricket 1896-1914) have certainly stood the test of time.
Also, those of us who make such a fuss of Mike Procter are justified in our enthusiasm of his ability. It’s a crying shame he only played seven official Tests. He had so much to offer the game.
The point of showing the leading player’s PQ for first-class cricket was designed to bring the likes of Grace and Procter into the overall picture, as they played little Test cricket.
Finally, a word on Donald Bradman. If ever there was an example of a cricketer who could bat each of five days in accumulating 1000 runs, Bradman is it.
This has been an article on all-rounders, yet such has been the massive weight of runs made by Bradman (40-60% better than the next best) that despite little bowling he comes in with a Test PQ of 135.31 and a first class PQ of 123.02.
Which would rank him fourth and sixth respectively in those categories.
There’s no doubt about it, Bradman is a nonpareil (without equal).
The point of assessing all-rounders via PQ is to demonstrate their productivity or busyness in a match, bringing both batting all-rounders and bowling all-rounders to a common, or level playing field.
Of course, the example of Gregory and Miller demonstrates that PQ only shows part of the story, but not the full story.
Gregory was better on a day-in, day-out basis. But Miller was the guy to do the unbelievable with bat or ball when the occasion demanded it. He had the ‘X-factor’, and we will need an entirely different methodology to assess that!
Irrespective of whether you agree with my PQ methodology or not, it is important that you have enjoyed reading this journey through cricket history.
Rusty
Roar Guru
I think you are being hard on the batsmen Sheek - they only get two chances per match, sometines only one to make that impression. A bowler can have a shocking over and then reel in 3 wickets... tough balance
sheek
Roar Guru
BB, We all have our heroes. I can see Kallis is a hero to Rusty & yourself. Sobers is the best all-round cricketer I've seen. Like I've said plenty of times through these two articles - stats are a guiding hand, but they're not the whole story by themselves.
sheek
Roar Guru
ak, I still think 20:1 is fairer to bowlers. We're forgetting for the moment that a bowler's potential haul is finite - 10 wickets per innings, 20 wickets per match. On the other hand, a batsman's potential haul is infinite by comparison. He can score a hundred, or 200, or 300 in an innings. Even nearly 400. He can score a 200 & 100 in the one match. He can two hundreds in the one match. Any ratio of runs to wickets ought to reflect this conundrum.
ak
Roar Guru
Sorry. It should be per match.
Rusty
Roar Guru
if all great batsmen scored 80 and innings then they would average over 80 -So I think you are setting the bar a little high there for the batsmen unless you are Don Bradman. So assume you meant 80 per match?
ak
Roar Guru
If stats are analysed, we can say that great bowlers on an average pick say 4.75 wickets per match whereas great batsmen score 80 per innings. 80/4.8=16.67. Maybe we can take 16 or 17 for calculating PQ. By the way an all rounder called Kapil Dev who does not find a place in this list has turned 55 today. Happy Birthday Kapil !!! Really cannot believe such a long time has passed. Everything seems to have happened just a while back.
Rusty
Roar Guru
Hey Sheek - love your work as always! I compared Sobers to Kallis as thats where a lot of focus drifts but its more a general point. It doesnt really matter how many tricks/variations you have in the bag if you arent taking wickets. So for a first change paceman or strike bowler 90+ is a pretty poor return and on pure bowling terms would generally not be good enough to make a team. Call me tough but I dont think Sobers was actually good enough for a first change paceman. If anything he was more in Kallis's mould of a fourth with the added value of spin late in the match. You mention the holding role - well consider this fact. One of the worst holding/attacking spiiners I have ever seen, Imran Tahir has a strike rate of 72.8! and lets be honest he is pretty rubbish. So I think that strike rate is an important factor in your value calculation somewhere - although it will probably again devalue the batsmen all rounders.
SandBox
Roar Guru
Yeah, it's going to be very hard. Haddins' effort in this series shows how some runs are more valuable than others
JGK
Roar Guru
Agree. There is no perfect way based purely on the stats. However, it is good to see the likes of hugely under-rated and mostly forgotten cricketers like Gregory and Faulkner get a mention in an article like this. A couple of years ago when they were trying to come up with a name for the Trophy that Aust and Saf played for, I suggested the Trumper-Faulkner Trophy after the two great early contemporaries who in back to back Tests became the first two players to score double tons and lose a Test match.
Deccas
Guest
I think its got something to do with the era of professionalism which began around the mid to late 70s though I wouldn't discount the influence of pitches. As specialisation and training have become bigger and bigger parts of the game pure talent is less likely to set someone aside as better than their peers. There is a wonderful book on sporting statistics written by a scientist called stephen jay gould called Life's Grandeur. He uses baseball statistics because he is an American but there are a few references to cricket in there. What he discusses is that professionalism see's a shortening in the range of statistical appearances. Its a wonderful read and anyone interested in sports and statistics would be well served by picking it up. That goes for you to Sheek, I think you'd find it fascinating and would probably give you encouragement to write a part 3! Thhanks for this series, they have been some of my all time favourite Roar articles.
trev
Guest
Also runs and wickets per test (as opposed to averages makes it difficult for allrounders who had to share the wickets with other bowlers (Jacob Oram) or didn't get to bat enough - ala Shaun Pollock
trev
Guest
What about Ashwin? He has certainly made a stunning start to his test career (albeit mostly in India). Looking into the crystal ball, as a kiwi I'm also excited by the progress of Corey Anderson (and Ben Stokes - a Christchurch lad who got away).
sheek
Roar Guru
Hi Andy Roo, Agree entirely that stats "can be interpreted in so many ways to prove or disprove almost any hypothesis." Of course, everything would run so smoothly if everyone just agreed with me! ;-)
sheek
Roar Guru
Hi Sandbox, You might be waiting a while... ;-)
sheek
Roar Guru
Rusty, Thanks for your kind words. However, I must take umbrage that Sobers was an average bowler. Even you must understand it is a dangerous practice to take strike rates & economy rates at face value. Also, by the time Sobers began playing tests in the near mid-50s, wickets were regularly being covered. Sobers & Kallis were used differently by their teams. Sobers was often used as one of four bowlers, but played two roles - first change paceman & second spinner. Kallis was most often used as a fourth/fifth paceman in short, sharp spells. In these circumstances, even though you would expect Sobers to take more wickets, you would expect Kallis to have a better strike rate. Remember also that spinners, by the nature of their craft, require more balls to take a wicket than a pacemen. By all means, let's use your scenario but make it all wickets in all conditions in all countries. Sobers will be able to mix up his bowling according to the conditions & the opposition - leftarm pace, leftarm swing, leftarm orthodox spin, leftarm chinamen spin. Kallis has his rightarm medium-fast swing & seam. How will Kallis go if he can't effect many breakthroughs within 10 overs? Can he bowl 30 overs, as Sobers often did? Can he contain as well as strike?
SandBox
Roar Guru
I'm looking forward to you determining an X factor ranking. Some variables would include: quality of opposition, condition of pitch, weather conditions, quality of the team you are on, the overall pressure on the innings/bowling. May I ask did your PQ multiplier at the 18.65 come up with different results again from 15&20?
On the contrary mate. i checked the batting strike rates a few days ago. Firstly not all the balls faced by Sobers was recorded, however when you look at those recorded it clearly shows that Sobers had many innings where he would bat as slow as two runs an over all the way up to 4.5 runs an over. On close inspection Kallis had a good number of innings where he would bat above 4 runs an over. As for being a stroke player, Kallis hit as many boundaries (fours and sixes) as Ponting did in his career, only difference was where Ponting were flamboyant, Kallis was technical and did it without fuss. Smilarly Kallis up to 2007 had a strike rate of roughly 2.5 runs per six balls whereas his rate was over 3 runs per six overs in the latter part of his career. Indicative of the role he played prior to Smith, Amla and de Villiers. Besides, strike rate in bowling suggests the effectiveness of a bowler to take wickets whereas strike rate in batting is merely indicitave of the pace of a batsman and not his ability to score runs, which is the primary focus of a batsman.
Andy_Roo
Roar Guru
Hi Sheek, A good article and very detailed methodology. I am reminded however that statistics can be interpreted in so many ways as to prove or disprove almost any hypothesis. What about bringing fielding ability into the equation. Catches taken and run-outs achieved. (I know there are no detailed statistics on run-outs)
Rusty
Roar Guru
Fantastic article series and an interesting view point. However I'm not buying into the Sobers bowling myth - in an age of uncovered wickets etc with a strike rate at over 90 he was an average bowler. To put it another way - if Sobers and Kallis were to do nothing but bowl to an entire team until they were out, using their relative averages and strike rates. Sobers would take 153 overs and concede 340 runs Kallis would take 115 overs and concede 326 runs So, how is it that Sobers is a more effective/valuable bowler when he would take nearly 40 overs more to take the same number of wickets on what were more bowler friendly pitches? Its a pity we dont have balls faced and strike rates for the batting side of things to make it interesting
sheek
Roar Guru
BB, You Saffies can't have it both ways with Kallis. His bowling strike-rate is good because he was often used in short, sharp busts. If he had bowled more, he would have probably taken more wickets. But like Sobers, his strike-rate would have blown out as well. ;-)