Wow......thanks for the analysis. I never even knew that a person called GA Faulkner ever walked the earth let alone think that he is one of the best all round players ever to have played the game.
[[
Ali
Pl look him up and if you can, please read Christoper Martin-Jenkins' description of GAF in his excellent book on Test players.
Ananth:
]]

A few things that strike me after this analysis Ananth.

Given that Kallis is such an extraordinary player, why he isn't talked of in the same breath as Gary Sobers? His ARIdx1 is higher than Sobers and Imran. Yet Sobers and Imran command higher praise from most followers of the game.
[[
I have always felt and mentioned quite a few times that Kallis has to be classed amongst the greats.
Ananth:
]]

Great analysis. I am also surprised to see that in terms of per test analysis (i.e. an even playing field) Keith Miller edges out Imran as a better allrounder.

Also in your analysis both Imran and Miller come out as better all rounders. They were both more of bowling all rounders. While Sobers and Kallis are batting allrounders. Any reasons for that?
[[
If you study the graphs closely, Miller and Imran, especially in the first and more important graph, move into the top right quadrant. Hence they are truly great all-rounders. For both Sobers and Kallis their bowling figures pull them down slightly. There is no denying that their batting makes up for it.
Ananth:
]]

But a wonderful wonderful analysis.

I have my own memories of watching the great test all-rounders from 1956 to about 1980. A move to the USA has meant only occasional glimpses of the greats of the last thirty years. No surprise then that Gary Sobers is my number one.
I do appreciate your analysis - especially since it gives me some idea of how players such as Imran and Kapil Dev stack up against the players of my memories. I have always thought that the 1970 series between England and Rest of the World was worthy of test status [as first designated and in light of the recent World XI inclusion]. Memory tells me that Sobers statistics would then move closer to the "top-allrounders" quadrant.
Then if you were to quantify the third dimension of a great all-rounder - fielding - I suspect that there would almost be no contest, and the hero of my youth would reign supreme!
Keep up all the good work.
[[
Edward
You are correct. The factors not covered in this analysis, for the very reason you have mentioned, are the fielding numbers. Also Sobers' bowling versatility is another key, but subjectiove, factor.
Sobers is Sobers. And we have to admit that Kallis comes close to Sobers.
I also feel that ICC should designate some of these great series as Test matches. If that mismatch between Australia and ICC can be termed as Test matches, these could certainly be.
Ananth:
]]

gr8 analysis. i always thought of miller and cairns as top class and kallis as hyped. can one recall any game kallis and sobers winning any game for their bowling. they were at best stock bowlers. kapil was mostly good as batsman on indian pitches. surprised not to see jack gregory, a fantastic fielder additionally, 15 catches in a series.
[[
Ashish
Jack Gregory scored only 1100+ runs.
Kallis and Sobers are world class. Not every all-rounder needs to be the no.1 bowler for his side. They are the no.4/5 bowlers.
Ananth:
]]

better make it 2 or at least 1.5 (so that shastri qualifies)wickets per match and at least a batting average of 25 with bowling average no more than 40 (so that prabhakar and phadkar?qualify.
[[
Ashish
We can always move the qualifying criteria around. These will only meet individual requirements.
Ananth:
]]

Great Analysis. I never though Imran Khan n Richard Hadlee r such greats. Could u pls do it for ODI's as well.

ananth,, good work... you try hard to quantify the relative value of runs vs wickets in the second analysis.. i believe that you missed a third harder to quantify category... catches...sobers, botham, kallis are all great slip catchers too and their contributions there are just as important... outfield catching is not as critical a skill as infield catching.. I would call someone with >1 catch a match as a critical catcher and >2 as bradmamesque except keepers where the ratios are worth a whole another exercise

Ananth - u had equated a wicket to 20 runs to come up with the ARIdx2 stat .. but then in the per test value graph, u have taken 3 wickets and 48 runs ... shudn't u have taken 3 wckts and 60 runs to apply the same logic ... the top allrounders will change after that (miller, botham and cairns move out of the group) ...

also ... how did u decide on the batting avg of 33 and bowling avg of 29 for the averages graph ... taking batting as 38 and bowling as 30 will change figures as imran and miller wont qualify as top allrounder ...

also .. sobers is no.1 and 2 in the 2 indexes u created ... and kallis tops miller in both of them ... then what was the basis of excluding them from the top 5 all rounders ... the 2 indexes are most analytical than the graphs which are subject to the decision criteria of averages and runs/wckts per match ... so i think the decision shud have been based on the indexes and not the graphs ...

but great work indeed ... keep them coming ...

Great work. An interesting fact about Imran, he avereged almost 50 with the bat and 19 with the ball in his last 9 years or so. Not that it makes any impact on your analaysis, just an interesting fact.

Why weren't wicketkeepers considered in the analysis?
You could have related the catches and stumpings to wickets, with a different weight...probably 0.5 for a catch and 1 for a stumping.

As always, great analysis. All the importat numbers have been presented in a clear manner. Thanks Ananth.

Kallis is better than Miller in both the indexes. Yet you have not included him in the list of top allrounders while includig Miller. Can you clarify why is that. I understad Miller is in the top allrouder block in the charts, but Kallis is better on both the indexes

add to this excellent analysis, imran's magnificent contributions as a captain, and you probably have the most complete all rounder ever! what do you think ananth?

i doubt if any other all rounder in cricketing history managed to galvanize an oft rift driven side into world beaters! as benaud once said that "nothing in Pakistan cricket has been of the same quality since imran left" .... i would dare say nothing in world cricket has been of quite the same quality in terms of complete allrounders ....

What an analysis. Congrats-
is there away of loking at per test value - counting opposition strength for batsmen and bowlers also

well.....An excellent analysis again, annath. But, i would like to make 1 point here. I, for one, feel that a true highest class, quality all rounder should be able to win matches with both bat n ball.Now, shaun pollock would hardly win any match with his bat(2 100s and 16 50s in 108 test matches).For the same reason kallis wud rarely win a test match with the ball(5 5fors in 140test matches)n similar arguments for imran n hadlee as well........ so, I, for 1, think that no. of 100s or 50s per test match and no. of 5 for or 4 for per per innings should also be taken into account and with this Ian botham comes out as one of d top contender(14 100s, 27 5for).... and we have all seen Botham winning test matches equally with bat n ball........ i agree dat his batting n bowling avgs don't support this claim but in his pomp, it was hard to find better bowler or batsman than him..........
Miller n faulkner will also be among top all rounders as they already are.....

Anantha Ji, a very good work but I have a few points to question the statistical approach. (1) Isn't it comparing apples and oranges when bowling quality, protective gears and state of the wickets make it totally a different game while playing it 80's or now a days. I don't think we have better batsmen today than 80's only because there are 4 times more batsmen with 50+ averages today than in Imran and Kapils times. (2) Does it matter for some statistics that Imran played 48 matches as captain (app. 10 years) where his batting average is above 52+ and in all away matches it is 50+ (with variables of 80's like wickets etc.) and his bowling average as captain is 20.He wasn't a great player before becoming captain as in these 40 matches his batting av. is mere 25 and bowling av. 28. So in my opinion, Imran the captain (last 10 years) is not only the greatest allrounder but the best ever cricket captain (surely not by win/lose ratio alone as that depends upon player material) too.

Good to see Aub Faulkener getting the recognition he deserves. He was the Jacques Kallis of his day and if he were here today he would be even more useful to Sth Africa than JK is. His batting would probably improve given modern conditions and his bowling average would probably stay the same given there are so many poor techniques to spin out there. He would be exactly what SA needs, an old-time spinner who bats, to be number one.

Thank you very much for enlightening the world by your knowledge. A very American analysis; completely useless. You seem to know a lot about cricket, you analysis clearly indicate the breed self-proclaimed game analysts. Please go back and learn what cricket is all about, it is not always about numbers, it depends on the opposition you are playing against, quality of the team you are part of.

In my opinion an allrounder should only be considered world class if he is among the top 4 bowlers and among the top 6/7 batsmen in his team. Sobers is a top class batsman but a very mediocre bowler. Kallis goes the same way but could have been better had he bowled in or before 80's. I will rate imran, a much better allrounder than the above mentioned two gentlemen. Some one mentioned fielding to be a factor worth considering, but in that you can't say a player is a better fielder if he has taken more catches. It's the catches which he dropped should matter.

Agreesive intent and Great Bowling Action...

To Me Imran was better than Sobers.... He was better then Hadlee, Botham and Kapil .... He was an ultimate specimen ... The Great Legend....

Super analysis. Now, my request/ question would be - should all-rounders only be judged as people who can bat and bowl with nearly comparable felicity? Sobers, due to his excellent fielding and Imran for his brilliant man-management/ captaincy skills probably are the alpha males among the lot. Kieth Miller was a top allrounder - the fact that there were times when he'd give away his wicket for he felt there was a mismatch magnifies his achievements ( though definitely not statistically). What about Adam Gilchrist? - in my view the only certainty in an all time World XI aside from Bradman and Sobers. Given that they contribute in two departments equally, should Mark Boucher (probably a more avg. allrounder for his batting skills), Denis Lindsay or M.S.Dhoni be considered?

I am glad it dispels a couple of myths: Kallis is a great allrounder, Flintoff is not.

A very interesting piece that confirms a few instinctive suspicions. The analysis shows conclusively that of the genuine all rounders, Sobers was the greatest batsman and Hadlee the greatest bowler. Obvious, yes, but if the analysis doesn't demonstrate that it is flawed.
I'm particularly struck that Botham and Miller do so well, notably when considered on a per test basis, as they have always struck me as the 2 most likely to win a game with either bat or ball. Of the 2 measures, I find the 2nd measure more useful as it reduces the impact of those who have accumulated runs or wickets (e.g., Imran & Kallis) through longevity, but were rarely match winners in both disciplines (which is not to say that Imran was not a fine batsman or Kallis a fine bowler).
A few additional suggestions: catching surely has to come into the equation somehow. More importantly, there should be a way of weighing match-winning contributions i.e., factor in 50s/100s & 4/5-wicket hauls in victories.

Very interesting. Hadlee being so far out on his own in terms of wickets per test confirms that he was one of the greatest fast bowlers of all time, let alone being an all-rounder (and NZ didn't really need him as a batsman so much with Crowe, etc in the team).

On Steve Waugh, his wickets per test in the first part of his career, before his recurring back injuries stopped him from bowling and forced him to become a pure batsman, would surely have seen him qualify - although his batting in the first part of his career wasn't great, so would perhaps have disqualified him!

Waugh was a Vettori in reverse - Vettori started as a specialist and has become an all-rounder (whose bowling has suffered), while Waugh started an all-rounder and became a specialist (one of the best batsmen of his day).

In view of the number of comments to be answered, I have given a common response. My apologies for that.
1. Catches will become a third dimension and cannot be accommodated in this analysis. Similarly the Captaincy related measures.
2. However Andrew (Gray) has suggested in a direct mai a few days back that I should do a comprehensive player contribution analysis incorporating batting, bowling, catching and captaincy. It is a mammoth task and should answer most of these queries.
3. The dividing positions were determined, not arbitrarily, but to have approximately equal players, above/below and left/right. These could be changed but on a justified basis.
4. Someone has mentioned why not 3 wickets and 60 runs. The reason is that the runs per test has always been a relatively lower figure, comparatively speaking, than wickets per test. Barring the freak value of Bradman, no one has exceeded 100 rpt. On the other hand 11 bowlers have exceeded 5 wpt. Hence getting a cut-off of around 50 rpt was essential. Also do not forget the lopsidedness caused by Kallis' and Sobers' figures.
5.I have said this a few times. I have tried to ensure that the top all-rounder quadrant does not contain any player who is below par in one of the roles. Hence Kallis/Sobers on the one hand and Hadlee/Akram/Benaud could not be considered.
6. I do not agree that Pollock could not win a test for South Africa in either capacity. The fact is that he was not the leading SA bowler only because of the presence of bowlers like Donald. He could as well have batted at no.6 in his entire career.
7. Despite my repeated affirmations on the type of analysis this is, some readers bring in opposition strength, match conditions etc. I do not know how many times should I tell that this is solely based on unadjusted career figures. There is a place for the other type of analysis. Why mix these up.
9. Do not forget that average batsmen, amongst all-rounders, are here because their bowling compensates and average bowlers, again amongst the all-rounders, are here because their batting is great. The second table is a clear indication of the value to the team.
8. Re considering Gilchrist etc pl see response no 2. They will certainly come through there.

Hi Ananth

Great analysis. One question for you- Do you think wicketkeepers should also get a look in in the context of allrounders . A cut off of minimum 200 dismissals and 2000 runs would bring lot of champion players into this , guys like Gilchrist, Rod Marsh,Jeff Dujon, Alan Knott, Mark Boucher,Ian Healy, Tim Evans etc into the picture

Regards

Deepak

I was suprised to see,based on a scientific analysis, that Malcolm Marshall did not make the cut,wheras Carl Hooper did. We in the West Indies consider the late Marshall as a better all-rounder than Hooper.
[[
Kadim
You are probably correct in your assessment. However Marshall scored 1800+ runs at 18.95. And if we let the great Marshall in, he will also open the door for two other bowling greats, Warne and Kumble !!!
Ananth:
]]

You need to deduct 10 runs from the batting averages of players who played after 1992. This is because they played in smaller grounds. So it was easier to score boundaries and batsmen had more courage in attempting to clear the boundary.

This is a great analysis, and one that needs to be made more often. Too often, people talk about kapil dev, and hadlee in the same light. Kapils bowling wasnt so great and hadlees batting wasnt that good, but they could win matches sporatically. Imran is my hero and im glad that he is up there as one the best in history. Just would like to know how his last 9 years stacks up against the others, im betting its better than anyone. Maybe an analysis of the best 40 matches of the players would be interesting. If we took an entire career to weigh people, then flintoff wouldnt be as talked of, and Marvan attapatu started horribly which lead to his average being only 40, but he scored had the record number of double centuries at one point later in his carreer.

It is very unfair on your part to have omitted mohinder amarnath as a good allrounder.We won the 1983 worldcup because of his performance.
[[
I do not want to be sarcastic or harsh. However should not you have known that this is a Test match analysis. Amarnath has captured 32 wickets in 69 Tests.
Ananth:
]]

Hey Kallis is great so is Shane Watson and kapil is not an average all-rounder.

Ananth - i said 3 wickets and 60 runs because u had equated them in finding the index 2 ... hence for consistency sake i said that and it does change tables ... also, the fact that kallis and sobers were so good in one is working to their disadvantage which in unfair ... if u are consireding 56-57 as unusually high batting avg, then 21-22 is also unusually low for bowlers ... imran can make it to my world xi purely as a bowler also ... but def not as a batsman ... similar logic appplies to sobers ... to say that imran cud win a match with his bat but sobers cud not with the ball in very subjective as the decision is based on 34(sobers) and 37(imran) . i wud say both the numbers signify secondary trade ...
hence ... also the top graphs are, like i said before, very subjective .. but not the indexes ... so i wud still say that sobers, imran and kallis are top 3 ... miller can only be 4th ... or 5th faulkner is also there ....
[[
Rachit
All of us can have our variations and interpretations. Nothing is set in stone. I have never claimed that and would never do so. I see no reason why a collection of 4 all-rounders such as Sobers/Kallis/Akram/Hadlee are not better than, say, Imran/Miller/Faulkner/Botham.
Ananth:
]]

Anantha, may be you do realize this but without the optimum slope (as you have admitted you had tried to find it) the rest of the analysis is dependent on quite subjective assumptions. Meaning size of the boxes will not be appropriately determined rendering who falls into them quite meaningless.Second, BCG has been abandoned in the world of business for purely that reason. Could not separate cash cows from the dogs because determination of cut-off point was difficult. I don't think 40 and 20 is necessarily better than 50 and 30 as you potentially have a Gavaskar and Kapil rolled in one player with the second set. Which set of all-round numbers depends on what kind of team you are playing in and playing against. So your assumption should be changed on that.
[[
Topa
The business requirements are quite complex. Here we are dealing with a simpler matrix.
I exampled 50/30 against 40/.20 only to show that the divisor is a better method. Surely the bowling difference of 10 is more significant that the batting difference of 10. Even on a % basis the 30-20 is far more of a drop than 50-40.
If 50-30 is Gavaskar/Kapil then 40-20 is Mark Waugh/Marshall.
Anyhow a sloping line will lead to many justified protests since it will be quite subjactive. Here at least I have a clearly laid down objective.
Ananth:
]]

I always thought that the term all rounder has been greatly over used. An all rounder is someone who would make his team for either his batting or his bowling. Vettori would do that for New Zealand right now and Botham, Cairns at times, Sobers, Akram and Kapil Dev (maybe) and Imran Khan, Miller and Faulkner (all 3 definetely) would have too.Shane Watson? Maybe not. Kallis maybe not for the reason that he does not win games with his bowling.Ditto Hadlee for batting. Adam Gilchrist and Jeff Dujon were genuine all rounders as is Brendan MacCallum.
You have rightly included fielding as a criteria and obviously captaincy skills are also a vital part of cricket. Pollock,Benaud, MS Dhoni and S Waugh come to mind immediately as players who deserve extra weighting on that score
Drifting over to the one day debate Sanath Jarasuriyah has been incredible with bat and ball for Sri Lanka over a very long number of games as has Kallis
The "ultimate" cricket team would include say 3 all rounders
[[
Michael
I presume you are referring to the MVP type analysis I have referred to.
Ananth:
]]

Hai All
I am from Bangladesh,you have not mention our top alrounder Sakib Al Hassan.I do'nt understand why do you always underestimate us.Last one year Sakib is the ICC's top alrounder.
[[
Tarik
My automatic program-based selection did not select Shakib since he has scored 1179 runs. Maybe he deserves a special consideration.
Ananth:
]]

Very refreshing, and i will relook up many past greats. Thank you

VIPIN GARG : U really made me Smile "IMRAN & HADLEE rarely made a Match Winning Performance with their BALL or either BAT"... ha ha ha... My Dear Pls Check Records, dont get blind in praising IAN BOTHAM.
[[
Next time if you do not give ytour name, the comment will not get published.
Henceforth no comments will be published if name is not provided. I understand that in most cases it is an oversight rather than deliberate action. Also the blog control program should not allow comments without name to go through. However ultimately it is essential that the reader gives a neme.
Ananth:
]]

Can u remind me of few matches where pollock's batting had won matches or even bailed Proteas out of trouble??Similarly Kallis running through a decent batting order......Pollock was at best only a useful lower order batsman and batted even lower down to Boucher for the same reason....Don't u think, a high class all rounder should be able to win matches with both bat and ball, at least at their best, in their pomp?????
[[
Vipin
I have too many things on my plate to look for such matches. I would suggest that you should go through scorecards to decide one way or other.
Ananth:
]]

Great work Ananth... a quick one, If Kapil is an avarage allrounder then most of the guys in this list are average all rounders. One should keep in mind that Kapil played 80% of his matches on dead subcontinent pitched.

As Ananth has pointed out these are statistical analysis. There are many things that will remained unanswered if you dont see beyond statistics. One, Imran Khan averaged 52 in batting and 20 as bowling during his tenure as captain. I dont argue much against his bowling being top class but with that batting average is highly inflated due to not outs. His runs per tests in those tests were less than even botham's career RPT and we all know what happened to later part of Botham's career. Botham was most gifted all rounder of the 80's quartet but injuries and that 81 series (which got into his head) ruined his chances of becoming the greatest of them all. Just after 5 years into his career he had a belly of 3 month old pregnant woman and after 8 years (at the age of 30!) he was bowling gentle medium.... what a waste of talent!! On the other end Imran and Hadlee kept themselves fit till almost 40.

great analysis. i would disagree with people who are vouching for sobers and kallis. all round performance means both ball and bat. if a player is even slightly off balance in either of one departments he should not be considered for the top quadrant.we need to see this analysis from a very specific point of view. that is the all round capabilities. an average of 40+- is very decent but a bowling average of over 30 is not. i would agree with ananth's list.good job

("I do not agree that Pollock could not win a test for South Africa in either capacity. The fact is that he was not the leading SA bowler only because of the presence of bowlers like Donald. He could as well have batted at no.6 in his entire career.")

I SO-SO-SO MUCH AGREE. The fact is: ***He could have batted almost anywhere.*** But he JUST missed by an hair's breadth being in that brown square due his below-par batting (by his own top-class standards) in the few test at the end of his career that brought his batting-avg down from 35-something to whatever he finished at. And I myself was just madly hoping for him in those last tests of his, that somehow he will could/will rise again to touch THAT batting average of 33 or above. Unfortunately he couldn't.

@sheharyar - i disagree ... kallis and sobers are off balanced cos they average too high with the bat ... that shud not b a negative thing ... also imran and miller average 37 with the bat .. that is definitely not better than averaging 32-33 with the ball ... and according to you. 40+ is a decent average with the bat ... then none of the top allrounders qualify

@Ariz - not out is not an inflation factor in tests ... if anything that goes to show how tough it is to dismiss a batsman ... surely a man cannot average 52 with the help of not outs ... also bothams messing up his career does not prove anything ... flawed genius will always be that ... flawed ... a person is only as gifted as his performances ...

Great job Ananth. I am sure all of us have our own opinions as is always the case. Personally I am a great fan of Jacques Kallis and therefore I always HOPE that he will somehow emerge to be the best. He doesn't always.

I do wonder however if it is not also essential to see who player played with. Someone like Kallis has always been in good bowling team and hence he has not taken as many wickets as some other players due to that fact(like Heath Streak who would not have made a SA or Aus side)(how did this affect Richard Hadlee?). Also some could have made more of their batting had they not been forced down the order.

Captaincy is not essential to this analysis of yours since not all players can be captains, although it counts in favour of an already topnotch allrounder like Imran Khan.

I expect that in future we will see less genuine allrounders due to professional cricket forcing all to focus on their best traits and neglect the other to some extent.

Great post none the less. Thx

My pick is Botham.

Reason : In 1980s he was the one of top leading batsman, bowler and fielder for England in every game they played. He had even scored a 100 and took ten wickets in India which is near impossible for a foreign player. He rolled over Aussies and had match winning perfomances against Pak, NZ and India.

His only blemish is not having a hundred against West Indies.

But with his crowd pleasing game and antics he is the best ever allrounder.

Sakib, does not require any special consideration.You are bias in making such remarks.Sorry to point out
[[
Tarik
I think you are losing it. I only mentioned that Shakib did not meet the criteria of 1500 test runs and hence if he had to be included it had to be a special consideration. Pl do not ascribe bias when it is totally missing. I have a lot of respect for Shakib. But he does not need such type of support. As and when he crosses 1500 test runs he will be automatically eligible for consideration.
Ananth:
]]

Bedi is greatest Indian bowler ever!
How you said it. average is not only thing.
but your analysis is great.

wonderful outstanding analysis... technically sound..
great so see people who have this much indepth sense of cricket...

Anyway some spots in graphs were overlying and not visible properly.

great effort though, waiting for more from you
[[
Nothing could be done anout the overlays. You van deduce the player yourself.
Ananth:
]]

Good work, keep it up. If Kapil cared bout his batting ( should have been a little less agressive perhaps, he was at par with Botham with Imran and Hadlee far far behind) and used any antics while bowling on dead indian pitches, he is and would be recognized as greatest all rounder bar none.
Finally stats is stats it will never measure impact of 4 sixes in 4 balls to save the follow-on. They are just 24 runs.

@vipin garg-check out Kallis bowling against England in England a few tours back-that will put your argument to bed! Its a pity Clive Rice and Mike Procter did not have the opportunity to play any/many tests respectively.Imran generously described Ricey as the best allrounder in the world! Batting order duties and bowling role play a huge part-Kallis batting at 3/4 against the new ball very often and bowling no.4 or 5 seamer has a vastly different role from Imran or Botham.

Nice analysis. Is it possible to do one for ODI all rounders and eventually arrive at a methodology to combine test and ODI standings? Thanks.
[[
Karthik
Yes can be done. One day has got additional dimensions.
Ananth:
]]

Comparing Kallis/Miller/Hadlee is subjective, except to say one's bat-dominant, one more balanced, and the other ball-dominant. Given the choice of one of them for you team, it'd depend whether your team was batting-light (Kallis would be a top player even he'd never bowled a ball), short of a bowler (even if Hadlee had averaged below 10, he'd walk into the modern Indian team) or you're seeking better balance overall (e.g. Miller frees up a batting or bowling slot).

In theory an all-rounder can do "two men's work". In practice they might not be good enough at both for that, but even if they're worth 1.5 or 1.2 players, they earn their place on the teamsheet. I wonder if it is possible to judge how many players an all-rounder is worth? If a decent top 7 bat scores 13% of team runs (in 1st inns, to stop distortion if unfinished 2nd inns), and a good bowler 20-25% of wkts, how many of both is an allrounder making 10% of your runs and 15% of wkts worth?
[[
Mike
An excellent idea. Let me see how I can work this into the MVP analysis I have talked about earlier.
Ananth:
]]

An excellent analysis, as usual, Ananth. Well thought-out.

Just a thought, though:
Why are, for example, Imran is a very good all-rounder, and, say, Sobers not as good because his record is a bit batting-heavy? I know the restrictions you have imposed, but just thinking of it, who should one prefer in the side, just based on batting and bowling averages?
Sobers 57.78 and 34.04
Imran 37.69 and 22.81
I know one shouldn't take the differences (batting averages are larger, hence the difference between two batsmen is always greater than that between the difference between two bowlers in terms of bowling average). If we take the ratio,
Sobers 1.697
Imran 1.652
You have based your analysis on cut-offs. What if we remove cut-offs and add the ratio to the calculation as well?
I'm moving the rest of this to another post.

See, if BAT represents batting achievements, BOWL represents bowling achievements,
You have classified the all-rounders based on min (BAT, BOWL)
I would suggest it to be a function of min (BAT, BOWL) and also the individual values of BAT and BOWL.

Isn't the combined value of batting and bowling what we seek for in an all-rounder? If you fix 33 and 29 as your cut-offs, a person with averages 33 and 28 gets into the Top Allrounder quadrant, while people missing out can have 55 and 30 (Kallis/Sobers) or 32/20 (Pollock is the nearest, but there can be future entries).

Doesn't a 55-30 person make a better allrounder than a 33-28 person? Adding the ratio to the pile might eliminate this. Please let me know your thoughts.
[[
Abhishek
For the purposes of graph I have to use only two facets of the all-rounder's characteristics, in this case, Batting and Bowling. I cannot complicate it by pusing in complicated ratios.
However please note that the two tables given provide an alternate insight. You coulkd make your selection on either or both of the tables. For instance add the ranks in the two tables and then select the top-5. Purely in terms of these two Idx values, there is no way to keep Sobers (total rank 3) out of the top-5 and probably Kallis (total rank 10.
You could do that exercise and send me a new table. I will post it.
Ananth:
]]

Hi Ananth,
Great analysis..i have only one doubt as to why keep 48 as the benchmark for runs/test and not a round figure like 50..is it to prevent Flintoff from falling into the average category..i certainly feel Flintoff was an average allrounder and the batting benchmark should be 50 runs/test.
[[
Dikshit
Normally I would have been quite upset at reading your remarks. However I realize that this somment is out of ignorance of the methodology or not reading the article fully.
The center line is selected not to keep any specific player in or out but to have approximately same number of players on either side. That is all.
Ananth:
]]

excellent analysis!
Normally stats speak for themselves, but the numbers really dont give the entire scenario. Genuine all rounders are a rarity today, especially with the advent of the shorter versions of the game. The impact that the 4 contemporay cricketers had on world cricket ie Imran, Kapil, hadlee and Botham, its hard to find any of the present day cricketers do the same. I would say, Imran probably was the best all rounder, because besides being a wonderful bowler and a batsman who could both defend and attack, he was a great captain who transformed talented individuals into a team of match winners. The other three had their brilliance, but could not compare to Imrans status as a captain and astute statesman

Looking through the interesting but entirely statistical analysis reminded me of the importance of Cairns, and the fact that he is sooooo under rated as an allrounder by everyone. His stats are similar (if not slightly better) than Flintoff, but Flintoff is regarded as a great allrounder by alot of people. As far as I can remember most of Flintoffs greatness came from one series (2005 ashes) in home conditions.

Slowly the world is may realise what an excellent talent Cairns was, and what a loss to NZ cricket.

And Grieg is also somewhat under rated

Ananth,
Can such analysis be done of career performances taking the quality of opposition into account. For example, a Kallis hundred against Bangladesh is vastly different from a Kapil Hundred against South Africa in 1992.
Perhaps you can use an earlier table of yours in which you outlined the test teams of every decade
[[
Pawan
For this types of analysis adjusting career figures is not worth it. Kallis' 100 against BD would certainly be compensated by a Kallis' 100 against Australia, over a long career.
Ananth:
]]

I wish ppl who have not seen Imran Khan bowling at his best could watch his videos. He was the best bowler and the second best batsman in Pakistani team. he was a tiger who used to stay at the wicket inspite of wickets tumbling on the other side. He was not a natural superstar. But worked harder than anybody else on his new bowling action and batting technique. He was unlucky to be out of team becsause of serious injury when he was on top of his cricket. There was nobody in the world who could face him. Ask Gavaskar. His indipping yorkers and incutters used to cut biggest batsmen in the world into halves. To me there is nobody better all rounder as Imran Khan.

The "two men's work" concept also applies to Bradman - a clear example of one man who could make the contributions of two, despite the quality of his legbreaks!

In terms of teamsheet considerations, a player like Bradman actually works like an all-rounder. It allows lower-order players to be selected more purely on bowling or wicket-keeping. When England had collapsitis in the 90s, they often compensated by stiffening up their order, adding late batting capacity at the expense of specialist wicketkeeping or bowling - resulting in a surfeit of "bits and pieces" cricketers. A "double-worth" batsman would likely have changed that policy.

When some posters here talk about captaincy/fielding/keeping as broader dimensions of all-roundership, I think they're asking "what added value does this player bring?" Can that be measured in "players-worth"? Hard for captaincy (maybe stats can show a good captain is worth eg an extra "half a bowler" in the field?) but sounds plausible for the others.
[[
Mike
The term MVP or Player Value aptly defines what we are taling about. And Captains have to be assigned clear values. As do wicket-keepers.
Ananth:
]]

Mike's idea is a good one on the issue of "how many players someone is worth". I think this should be a separate analysis not restricted to just allrounders. Bradman with a batting average of two batsmen was likely worth 2 players. So was Murali, who did the job of 2 bowlers. I think all players should form part of the analysis (with some suitable cutoff ofcourse). It will give us a good perspective on whether an allrounder is better than a specialist.

Hai Ananth
Thanks,for making it clear.However in future please do, accommodate some Bangladeshi young cricketer like Tamim,Sakib,shariar nafees,Ashraful etc in your write up.An exceellent analysis on Great Alrounder's,becox I follow world cricket since 1971.Keep it up
[[
tarik
Will keep this in mind. For what it is worth let me say that Shakib deserves his place in this elite group.
Ananth:
]]

Stats are like bikinis or so goes the story. It is easy to see why. In this analysis Botham seems just another all rounder but I just happened to read articles on him by Mike Slevey and S. Rajesh (Legends of Cricket series) and their numbers show a different story. He was one of the greatest if not the greatest all-rounder before his back injury after 51 tests or so. His stats are so good for that era that Kapil, Imran and Hadlee are not even close. You should have included a matrix based on MOM award-a true indicator of any players worth especially all-rounders. Hope you do that for MVP analysis. I think, as others have pointed out, in this analysis there is too much weight on bowling average (comparatively) that has made some players looking better or worse.
[[
Rahul
If ever Rajesh did an article on Imran Khan, he could certainly do the same thing to Imran what he did to Botham. After all, all these analyses, as you say, are numbers based.
Incidentally both the averages get equal weight.
MOM is non-existent for over 100 years. What does one do.
Pl see my response to Abhishek.
Ananth:
]]

@ Akbar Khan, yes he was a great player but so were other all-rounders. I think you are hyping him a bit too much. He was definitely a better bowler than the rest( except may be Hadlee)but Botham and Kapil were probably better attacking batsman than him. His batting average improved significantly later on when he started coming way down the order (not outs) and played almost like a conservative lower order batsman. Imran unlike Botham has not won many test matches due to his batting. He had a great in-swinger, no doubt, but what happened in that series is best left untouched.

Interesting analysis - if we peel the onion further and analyze Imran Khan's career in 2 nearly equal segments - pre-1980 (handyman and occasional bowler) and post 1980 (batting avg. 50+ and bowling avg. c. 20), the latter would show him as better than Sobers or Kallis - also, Sobers' career where you trim out his formative years as a teenage cricketer ( 3 years) and last 3 years back up to the 1970-71 India tour of WI - he was batting at No 6 or further down and easing younger players into the WI side.

Hi Ananth,
The analysis is very fulfilling..but i may be wrong..but is runs/test or wickets/test a right way to judge a player..i know it compensates for longevity..but players in certain dominant teams like Australia do not get to bat 2 innings often and would not get many 2nd chances whereas players in teams like Bangladesh get 2 innings in almost every test..same case for the bowlers..bowlers in weak teams often do not get the chance to bowl the opposition twice..dont u feel that these parameters of runs/test and wickets/test is debatable..as many players would score more runs in a test if they get a 2nd innings more often and same for the bowlers

good stuff. it all looks pretty much like i had in mind. i'll pat myself on the back for being an objective analyst! i'd like to see an ODI version to see how Sanath J features. Shame he underbowled in tests

Dikshit has fair points but there are complicating factors. True, good bowlers in weak teams get fewer chances to bowl than in strong teams (fewer fourth inns to bowl at, and even those are against lower targets) but folk like J. Gillespie suffer a counterbalancing problem: competing for wkts with other top bowlers. One factor in Murali's extraordinary wkts/match ratio is how he carried the Lankan bowling. That's not a criticism - the fact he bowled so many overs and took so many wickets is an impressive feat of endurance - but he'd have taken less wkts/match if supported by a top pace quartet.

Averages are "purer" measure but do get distorted. Second spinners eg Herath play on more turning tracks. Openers are more likely to bat twice on a high-scoring pitch than lower order bats. 5th/6th bowlers (Kallis, North?) get called to bowl when strike bowlers have had no impact, denting their averages. Different numbers reveal different aspects of the story, none show absolute or final truth.

Great points Mike. This is the problem of dealing with summary stats. They Just don't add up once you start look in. Central point of any analysis is to control for confoundings as much as possible. These are some very important ones you have mentioned and there are some more, I hope a truer picture of allround excellence emerges soon.

I have always observed that we tend to attach more greatness to former players than current players. This is just like past was always better than current. This is not correct in every case. This is the reason why Kallis is not given the credit that he thoroughly deserves. I am sure 10-20 years down the line he would be regarded as the finest all-rounder.

Hi Ananth,
Good Analysis!! Gives a lot to think about... and also takes allrounder at a specific perspective.
This would help people drawing a bigger picture in terms of allrounder as whenever the allrounders are talked about Kapil, Imran, botham and Hadlee appear as the greatest of them all.
Coming out from your analysis are people like Miller, Faulkner and Pollock along with the greats mentioned before.
I do not want to question the 4 "Real" allrounders but am glad that people like Miller, Pollock also make the mark.

Ananth, here's a suggestion: define an all-rounder (a-r) to have certain basic criteria including number of Tests / innings. To this add one more criteria of averages, for ex - bat-avg > 30 AND bowl-avg 45 AND bowl-avg 25 AND bowl-avg > 15.

Now, the best way to calculate the effectiveness of an a-r is to do the a-r-ration or a-r.r = Bat Avg / Bowl Avg. This number tells us that an a-r.r of 1 is a decent Test a-r. Any a-r.r of 1 is a better Test a-r. The guy with the best a-r.r is the best a-r ever (we have already included the subjectivity thru criteria filtering). The ratio gives us a relative score of all the a-r's and also tells us how far is one from the other.
[[
The ARIdx1 is exactly what you are talking about. Pl see my response to Abhishek. You could do your own selection using the two tables instead of graphs.
Ananth:
]]

Anant,
Thanks for this analysis. I have always maintained that Miller was an awesome (bowling) allrounder and had him in my World XI as one (with Sobers as my batting allrounder).
Is there anyway to do a slight more complex analysis (tried on statsguru but have not succeeded).
Is there anyway to give me the averages of top batsmen when they played against bowlers with the top 15 or 20 lowest averages (in their era ofcourse). e.g. runs made against McGrath are more valuable than runs against say a Kasprowicz(sp?)
Conversely, can you tell me the best bowlers against batsmen with highest career averages.For wickets against a Tendulkar should count for more against a VIkram Rathore...
[[
Aloke
When I do the BCG analysis of Test batsmen I am going to set up the weighted average bowling quality as one of the variables. That is a clear indication of the bowling attacks faced when the batsmen scored the runs.
Ananth:
]]

One problem with this analysis is that it only looks at the mean and not the variance. Using your data, I calculated the standard diviation of batting and bowling, and they are 7.9 and 5.9 respectively. What does that mean? A higher variance in batting means that within this sample of great cricketers, it's *easier* to be a better batsman than it is to be a better bowler. You should redo this same quadrant analysis using z-scores rather than simply the averages, and see how the players spread out. I did this, and it shows Sobers and Kallis being outstanding (and outliers) and Imran leading the field of people who are good at both departments, but he is not that much better than Miller or Faulkner.
[[
Arnie
Thanks for the insight. I am not very familiar with the z-scores methodollogy. I have looked it up in Wikipedia, that is all. Why don't you send your Z-scores table to the mailid mentioned in the article so that I can do a graph and post the same. It will be a nice participatory exercise.
Ananth:
]]

Very good analysis. Just wanted to add one thing here. Imran left the team which consists of such name like Wasim, Waqar, Saeed, Mustaq, Aqib, Injamam, Amir Sohail who keep winning games for next 10 years.

Great article. Thanks for the insight on Miller and Faulkner. However guys, this is JUST one way of calculating things - not the end of the World. Everyone have their favorite player and that's a value statistics cannot match. This is just for fun. So relax and read it for fun and don't question on how the author decided to come up with it.

Excellent stuff.

Two comments :

1. Why should all rounders be match winners? There role could be to bulk up batting and bowling alone - so as long as you could consistently be in top 6 and top 4 bowlers, you are an all rounder.

2. Would be silly to include catches. Why not fielding skills then? What if a player was a fantastic cover fielder..how many would he have caught? And catches come to you, while you make the runs and wickets
[[
Prakash
I will take catches only in the Valuable Player analysis.
Ananth:
]]

interesting analysis - as you say statistically based so many variables suggested are not pratical. one i would like to see considered, if you have a spare million hours or so, is the % of team runs and/or wickets (particularly at test level) taken by the variuos players. this may give some insight into their relative importance to the teams they played for.
[[
Bruce
All these will be considered in the Valuable Player analysis suggested by Andrew and supported by Mike.
Ananth:
]]

Hi Ananth
Again I am here,by now you must have heard the result of 1st ODI between BAN & NZ.It's Shakib's briliant performance which got through BAN a historic win.But to be world best AlRounder he need's to be more consistance, still long way to go.
[[
Tarik
The quality of Shakib is never under question. And I hope Bangladesh goes on to win the ODi series.
Ananth:
]]

Fascinating analysis Ananth. Actually, looking at the charts, I was reminded of basic microeconomics and how individuals choose between bundles of two goods. Typically, the indifference curve, which consist of all the bundles that make you equally happy, is bowed inwards. This reflects the idea that you prefer to have a mixture of two goods rather than an extreme of one- for instance, you might prefer 4 apples and 4 oranges to 10 apples only or 10 oranges only.

Maybe you could use the same idea for all-rounder analysis? That is, draw indifference curves from the team's point of view that bow inwards towards the centre (assuming that teams prefer someone who is very good in both batting and bowling to someone who excels in one department but is only moderate in the other).

This would give us a way to compare, say, Kallis with Pollock, or Sobers with Imran, while quantifying the idea that 'all-round' all-rounders are more valuable.
[[
Nahim
Intriguing idea. However I need more information. You could mail me directly.
Ananth:
]]

Anant - just a quick something I realized about the second, per test analysis. I feel it unfairly penalizes those bowlers who had were either change (as many allrounders are) or were fighting with many other good-great bowlers for wickets. Alternatively, it favours someone who did not have much competition in his team to take wickets.
I tested my hypothesis with McMillan and Hadlee who I felt were perfect examples of this (no this was not reverse engineered :)). From table 1 and 2 numbers, McMillan goes from 5 to 11 and Hadlee goes from 24 to 23. Maybe not appreciable numbers but hopefully you get my thought process. I think runs per test is the way to go..and maybe strike rate (as opposed to wickets per test) should be the second axis.
Tell me what you think - btw- love the time you take to respond to people, even if some are just not worth your time.
[[
Aloke
The Runs per test and Wickets per test are two compensating measures. It allows a player to make up for one with the other. Since these are the two per-test measures, I would like to retain the same values. Bowling strike rate does not belong here. Also if some one has a Wpt value of, say, 2.0, his Strike rate is likely to be quite high and vice versa. We may very well end up with a similar graph.
Ananth:
]]

imran khan is better than all once his analysis another he played the decade of 80s beautiful times of cricket he is greatest also has captaincy.kallis is not a perfect allrounder you should analysis match winnig performance and which player scored or tacking wicket against which type of team or player

Anant,
The more I think of it, the per test match analysis may be flawed. The reason I say so is that there is no specific place an allrounder must play (say 6 or 7). McMillan played often times at 8 if I remember correctly and Greig at 5.
What this does is gives Greig a higher chance to score more runs per test. Similarly, someone who was a first change (e.g. Miller) and who had awesome opening bowlers (Lindwall and Davidson I presume) would be unfairly dinked as 5 wickets may have often times been taken by the time he was brought on...

I feel # of 50s per innings batted (somehow accounting for notouts? like a 35*) and bowling SR might be better as it shows how they did when they had the chance themselves..and were not impeded or benefited by the state of the game or the composition of their team.

Just a thought..

If this has been raised please forgive me as I have not read all of the comments, but no mention of allrounder in the stats account for catches, surely a 'great' allrounder should be great in all of the disciplines.
Secondly can a player be judged by stats alone, few players have almost single handedly won test matches with bat ball and in the field but Botham and Sobers certainly fit this bill!

How good would Flintoff figures have been if the England team took just a fraction of the chances that went to hand, one of the unluckiest bowlers ever!

Stats don't lie: they all tell truths, and different statistics reveal different truths. For instance, "bowling ave" and "wkts per match" bring different perspectives and insights. The value of a great anaylsis like this one is how it helps us see the game in a fresh light.

Z-scores are unsuited for cricket because few stats have a symmetric, let alone normal, distribution: e.g. batting aves are strongly skewed to the right. "Fred's ave is 2 standard deviations above the mean" (Z-score=2) is unhelpful.

A better way to put averages in context is by quantile e.g. "Fred's ave is in the top 4%". This reveals just how much better a batting ave of 55 is than of 45 - the former surely being in the top 1 or 2% of all batsmen. Similarly a bowling average of 23 would be revealed to be far more unusual than one of 28. Best of all, quantiles can be equalised by era: "Fred's bat ave is in the top 4% and bowling ave in the top 8% for the period he played in" (ie for tests between his 1st & last).
[[
Mike
From what you say, Z-factor may be more meaningful if we take all batsmen (Bradman ???) in a batsmen analysis.. I may be wrong of course.
What you have suggested has given me an idea. Last year I did an article "How far ahead is the top one" in which I assigned 100 to the top ranked measure and relative values to the lower placed players. Bradman-100, Pollock-60.8, Headley-60.5 and so on. Maybe that sort of measuring can be revived in the all-rounder analysis.
Example. Figures not accurate, just to get an idea.
Sobers-100, Kallis-98, ..... Wasim Akram-39.
Hadlee-100, Imran-98, Miller-97,........Hooper-45.
Then if we add the two numbers we will get an excellent idea of the all-rounder standing. This will not be the usual 1-2-3 and adding but the exact measures.
Pl send a test mail to the given direct mailid so that I can establish direct communication with you.
Ananth:
]]

Great analysis. All rounders are the knights of cricket - they give the team so much flexibility. Good to see Keith Miller and Aubrey Faulkner get their due, although I am sad that Alan Davidson, one of my favourites, did not make the cut. keep it up.
[[
Davidson just about missed it. 1328 runs.
Ananth:
]]

Regarding those who think Imran's batting numbers are inflated because of not outs, please remember that the Pak team of the 80s was second best in the world because Imran was marshaling the tail. Without Imran, we have now come to know the famous Pakistani batting collapses of today. Pakistan drew a lot of matches particularly because Imran would bat 5 - 6 hours to save a match. His batting stands against the Windies in that legendary 9 match series are an example of that, as are his lower order partnerships with Qadir, among the most prolific in the history of cricket. As for the relative number of his low centuries, remember, this is a man who declared Pakistan's innings when he was 92 not out. He only cared about winning.

Ananth, great stuff as always. One question though. Why would you consider batting and bowling averages to be qualitative? As a consultant who uses these four-quadrant graphs frequently (and think that your idea of using these is brilliant, btw), I would submit that a qualitative assessment is one where you cannot quantify a score (something like a scale of high, medium, low could be considered qualitative). Batting and bowling averages have numbers associated with them - by definition they *quantify* a player's performance on this dimension. While I have no quibbles with where you place the players (allowing that is virtually impossible to consider extenuating circumstances for each player like Kapil's bowling avg suffering due to his playing on the subcontinent etc) - I don't think your distinction of qualitative vs. quantitative assessments holds water. They are both solidly quantitative - and are no less valuable for that.
[[
Sharad
Funny thing is that after sending the article for publication, I looked up Qualitative and Quantitative in Wikipedia. The definitions were not what I had in my mind and I had the same misgivings you have. I realised that "Quality" as defined there was not what I had in mind.
Perhaps, "Performance" to denote the Averages and "Numeric" to denote RWpT might have been more relevant.
Ananth:
]]

Ananth, an interesting thought occurs to me. Rather than rely on the numbers alone, why not use a ranking method. It doesn't eliminate continent-to-continent differences, of course, but it does help neutralize the fact that run scoring was much harder before covered wickets etc.

The idea is that you take a given player (using the criteria you already picked), and his batting average, and rank that relative to *all* players including specialists who meet a minimum qualification criterion (e.g. 10 matches) during that player's career. Then rank by quartiles or deciles. Now, we have somewhat normalized the data by era. Do the same for bowling. This would I think be quite fascinating. One could of course do the same for RPT and WPT.

I would expect the batting allrounders - Sobers etc. to rank 80th+ percentile in batting and perhaps 30th in bowling, and perhaps vice versa for Imran or Wasim. Thoughts?
[[
Sharad
Pl see response to Mike.
Ananth:
]]

Nahim,

Great idea - although for this kind of graph I would think that outward bowed curves representing efficiency frontiers would be more appropriate. In other words, if you are a bowling-centric allrounder, one could imagine improving batting quite a bit with little impact on bowling; but as the batting average pushes further right, the resultant impact on bowling becomes increasingly significant.

Actually, this discussion about z factors and rankings reminds me that you did a fine analysis of the fifteen great bowlers earlier on.
I expect you are doing a similar analysis of the 15 great batsmen. Once you are done with that, I would like to see where Kallis and Sobers rank in the list of all time batting greats. With Hadlee at number 3, Imran at 6, and Akram and Pollock in the top 20 bowlers, I would like to see whether Sobers and Kallis will rank in the top 10 batsmen of all time, and whether we can qualitatively rank their bowling skills and vice versa for Hadlee, Imran and others. The question then will be, of course, how the all time great batting all rounders stack up versus the all time great bowling all rounders. Perhaps adding the an overall combined ranking of both those departments for Kallis, SObers, Imran, Hadlee and Pollock will give us an idea of where they are? (eg. 2 + 120 = 122 for Hadlee)
Too bad Miller was not included in that earlier analysis, though.
[[
That had a tough cut-off. So Miller did not qualify. However the analysis can always be re-done with lower cut-offs.
Instead of adding the ranks, I can even do the type of adding computed measures as indicated in my reply to Mike.
Ananth:
]]

Ananth
Alan Davidson doesnt qualify? While I havent seen him play at all, heard and read a lot about him

The one outstanding statistic amongst all-rounders in my view is Imran Khan's record as Captain. In 48 Tests as captain (only 4 Tests less than Bradman's entire career), Imran made 2408 runs at 52.7, and took 187 wickets at 20.26. This includes 45 wickets at 14.86 against West Indies. It's the closest any cricketer has come to being superman.

Ananth,
1)
Kallis should be referred to as “Kallis the Kollosus”…Unfortunately he suffers from the “Dravid/Steve etc. Syndrome”…His “style” or lack of relative to some other contemporary greats tend to reduce the true appreciation of his gifts. If he had played in a bygone era his numbers would have spoken for themselves.
For eg. I don’t recall too many ppl referring to Bradman’s “style”, “flamboyance”, “flair”, “artistry” etc etc….He was simply a Run Machine…and his sheer numbers drown everyone else.

2)
Since you are using BCGs , a Business/management “tool” – Here’s and interesting analysis connecting the stockmarket to India’s ODI performances.

http://www.indianexpress.com/news/stock-markets-decline-when-india-loses-odis-says-study/693648/

I realise that this type of analysis would be slightly out of your jurisdiction but it still is a good indicator of what ODIs and Tendulkar mean to India.

Interesting analysis - though i feel the quadrant names, unintentionally give a slightly unfair sense. I understand it but not calling Kallis and Sobers top allrounders takes a bit of sheen off.
A suggestion - an all rounder contributes with both ball and bat. Suppose you take all the tests of a player - see the runs he scored wrt runs his team scored(say he scored x% of the teams runs) and ditto for wickets taken(say he had y% of the wickets taken. Then a composite would be Avg (x,y) which would give the all rounder's value.
One could also decide if a player is bowling/batting centric by cut offs on x and y independent of each other. The problem with the above analysis is that Sobers is made a batting centric allrounder because his batting is so brilliant. But he took 235 wickets in an era when the world record was 309. If he took, say 15% of WI's wickets, he's a bowling centric allrounder as well. What those cut offs should be, i don't know- perhaps the data would show that.

[[
Z-Scores for Test all-rounders
Anirban had mentioned the value of Z-scores and I had requested him to do the Z-scores work and send me the document. It is a very interesting one and the final graphs are more or less similar to the averages based one I have posted.
I have uploaded the document for the readers to download and view.
The link is given below.
http://www.thirdslip.com/misc/all-rounders-z-scores.docx
My thanks to Anirban.
Ananth:
]]

Deservingly Gary Sobers is at the top.Kallis may match him statistically but when placed in true light Sobers is incomparable.No all-rounder has so consistently swung the fortunes of a match with bat and ball.

At his peak Ian Botham was the best match winner of all if you study the period from 1977-1982.The manner he won the 1980 Jubilee test at Mumbai and the 1981 home Ashes is legendary.Imran was more consistent than Botham and the best from 1981-87 ,but could not at his best turn the complexion of a match with both ball and bat like Botham could.Imran was mainly a match-winning bowler and became agood batsman late in his career.Infact in the light of ball and bat together Kapil Dev had more prowess.Richard Hadlee,is rated too high as he again was mainly a fast bowler and became a good batsman late in his career.

Kallis is an all-time great but has not consistently been a match-winner with both ball and bat.Had Keth Miller played more he could well have reached the top.

Hi Ananth,
A few bloggers I noticed bagged Flintoff, whilst his career stats are not mind blowingly special, the fact is like Botham before him they totally bossed an individual Ashes Series (unfortunately). So perhaps to put some perspective an an analysis on a 3 Test or more series in isolation may show who had the greatest impact on a Test series ever. I suspect Bothams '81 Ashes series would be hard to beat.
Great article & can't wait for the MVP analysis.
Cheers
AG

I agree that Hadlee was the best fast bowler amongst all-rounders and could well make the all-time test 11 as an ideal partner to Marshall, and Akram or Lillee.

However readers must be explained why Kallis is not in the same level as Sobers,despite of being the champion statistically.The point is how many matches can an allrounder win with ball and bat or perform outstandingly.At his best Botham was Sobers greatest challenger while amongst fastbowling alrounders Imran Khan was the king.The Ian Botham of 1977-82 and the Imran Khan of 1981-87 were the greatest challengers to Sobers with Keith Miller very close.In certain phases if you consider his performances against the mighty West Indies Kapil Dev was almost in that bracket.Had he played for astronger team on helpful pitches he may well have joined Botham and Imran.

My best in order-Sobers,Botham,Imran,Kallis,Miller,Kapil Dev,Hadlee,Tony Greig,Pollock,Mankad,Procter.

Great Analysis Ananth.. I, however, feel that such analyses do not do proper justice to players. Without taking into account the teams that they were playing for, or the quality of support players (batting or bowling) they had, such comparisons can cause grave misunderstandings or misinterpretations.

I am not sure how a Botham would have fared if he was playing with a very weak bowling around him (did Flintoff last in this era, despite being hailed as the 'new Botham'?). Would Imran have lasted so long if he didn't have the fantastic pace battery around him?

This post is written a little after the article was done, but I've been analyzing Players & Stats trying to get a better understanding of past great cricketers & your articles are excellent over many years! This analysis though discounts a few Batting All-rounders from the past C.Macartney, S.McCabe, T.Dexter & others, these Bowled nearly 100 Balls per Match played (similar to M.Mohammad). Looking at these & other very good Batsman it really does increase their 'Value'. A worth that I'm very impressed by (I'd have liked to have been an all-rounder) in players, now & in times past!