THE CORDON HOME

BLOGS ARCHIVES
SELECT BLOG
October 3, 2008

Allrounders

The best Test all-rounders: a follow-up

Anantha Narayanan

The main article received a high number of quality responses. A number of useful suggestions were provided and the exchanges were conducted without the acrimony that had been the feature of earlier exchanges. I thank the readers for this trend and hope that this will be continued in future.

The final table has been slightly modified based on tweaks in the Performance parameter as suggested by Hamish and Kartik. Details at the end.

I had mentioned that I would do an in-depth article on selected all-rounders incorporating more measures, as suggested by readers. Some of the relevant points mentioned by readers are summarised below.

  1. Take care of an all-rounder being the first/second bowler (e-g., Imran/Hadlee) vs fourth/fifth bowler (e-g., Sobers/Kallis).
  2. Take care of quality of wickets captured.
  3. Take care of quality of bowling faced by the all-rounder in batting.
  4. Equalise Batting and Bowling to a greater extent.
  5. Take into account Bowling strike rates.
  6. Reduce the weight for longevity measures.
  7. Take into account Fielding data.
  8. Take into account Captaincy data.

I have considered all these requests and incorporated whatever is possible and quantifiable. The response to these points are given below in the same order.

  1. This has been taken care of in a round-about manner. I have determined the % of Balls bowled by the bowler to the Balls bowled by the team. This will clearly give an idea of the bowling importance of the bowler. In addition the % of wickets captured by the bowler to that captured by the team is also deternined to do a more informed determination of the bowler status.
  2. Has been done by assigning the dismissed batsman's batting average to each wicket captured by the player. Capturing the wicket of Tendulkar will add 54.23 to the kitty while dismissing Zaheer Khan will add only 11.77 to the sum. Fair enough.
  3. Has been done by assigning an opposite team's actual weighted bowling strength, based on balls bowled and bowling average, to each run scored by the player. Take three centuries scored by Kallis. The 115 against Zimbabwe will be weighted with a value of 46.47, the 105 against Pakistan will be weighted with a value of 35.40 and the 139 against Bangladesh will be weighted with 59.98. Cannot do more.
  4. Has been done in different ways.
  5. Has been done. The longevity weightings have been halved.
  6. Not done since the Bowling Strike rate is a constituent part of the Bowling Average. I am aware that separating the Bowling Accuracy and Bowling Strike Rate and doing a differential weighting of the two measures will change the dynamics of the calculations. However I see no reason for doing the same separately since the Bowling Average is a universally accepted figure encompassing both. If Sobers has higher bowling strike rate, he makes up for it partly with his accuracy and this is already reflected in his relatively high Bowling Average. Doing the Strike Rate in addition will penalize players doubly.
  7. Not done since an all-rounder analysis should take only Batting and Bowling. Fielding does not become part of a player's all-round abilities. If a fielder fields at first slip, he would get a number of catches. If he fields at long-on, he will get very few catches.
  8. Not done because a player's captaincy ability does not add to his all-round abilities. It is an opportunity given to him which he does very well (as Imran did) or poorly (as Botham did). However what about Hadlee who never captained New Zealand. It was not because he did not have captaincy skills, rather it was because NZ board did not want to overload him. He cannot be penalised for this. Similar situation exists with Kallis/Miller et al.
The criteria for selecting the elite group of all-rounders is the following.
  • 2500+ runs.
  • 25.00+ Batting average.
  • 100+ wickets.
  • Bowling average less than 40.
13 players qualify. Out of these, Vettori is excluded since, surprisingly, his bowling average is higher than his batting average and that is not the sign of a world class all-rounder. Hadlee also has a sub-30 batting average but he makes up with an outstanding bowling average. The following 12 all-rounders qualify.
No Player          Runs   Avge    Wkts   Avge

1.Sobers G.St.A 8032 57.78 235 34.04 2.Kallis J.H 9761 55.46 240 31.23 3.Imran Khan 3807 37.69 362 22.81 4.Miller K.R 2958 36.97 170 22.98 5.Botham I.T 5200 33.55 383 28.40 6.Pollock S.M 3781 32.32 421 23.12 7.Hadlee R.J 3124 27.17 431 22.30 8.Kapil Dev N 5248 31.05 434 29.65 9.Cairns C.L 3320 33.54 218 29.40 10.Flintoff A 3494 32.35 206 32.21 11.Goddard T.L 2516 34.47 123 26.23 12.Greig A.W 3599 40.44 141 32.21

All these players have the figures to qualify as all-rounders. They could all bat at no.7 and above and would feature in all their teams as one of the top 5 bowlers.

However since this list is primarily a post-WW2 list, a separate analysis is done for all the all-rounders who have played their games before 1948. This list is shown at the end.

The following parameters have been used with the weightings allotted.

1. BAT - Runs Scored (7.5 points):

The weighting has been reduced by 50% to 7.5 points. One point per 1500 runs and limited to 7.5 points.

2. BAT - Batting Average (10 points):

The range is from Sobers (57.78) to Hadlee (27.17). The point allocation ranges from 0.0 for Batting Average of 25.00 to 10.00 for Batting Average of 60.00.

3. BAT - Bowling Quality faced (10 points):

The range is from Botham (34.38), a surprise indeed, to Miller (43.63), indicating the average quality of bowlers during 40s-50s. Botham's number puts paid to the wrong presumption that he was a flat-track-bully and did not do well against good quality bowling.

The point allocation ranges from 0.0 for Bowling quality value of 50.00 to 10.00 for Bowling quality value of 30.0.

4. BAT - % of Team Runs scored (5 points):

The range is from Sobers (15.78%) to Pollock (7.68%), both understandable. The point allocation ranges from 0.0 for TRS % of 0.00 to 5.00 for TRS % of 16.0.

5. BOW - Wickets captured (7.5 points):

The weighting has been reduced by 50% to 7.5 points. One point per 60 wickets and limited to 7.5 points.

6. BOW - Bowling Average (10 points):

The range is from Hadlee (22.30) to, not surprisingly, Sobers (34.04). To a great extent Sobers has made up his poor strike rate with his accuracy. The point allocation ranges from 0.0 for Bowling Average of 40.00 to 10.00 for Bowling Average of 20.00.

7. BOW - Wickets Quality (5 points):

I have always felt that Flintoff consistently captiured top order wickets. This is more than borne by these numbers. The range is from Flintoff (34.85) to Cairns (29.29). Incidentally 84 of Flintoff's 206 wickets (over 40%) are of batsmen with Batting Averages exceeding 40.0.

The point allocation is from 0.0 for Wkt quality of 25.0 to 5.0 for Wkt quality of 35.0.

8. BOW - % of Team Balls (5 points):

As stated already this is a clear indication of the player's importance to the bowling attack. Kallis is clearly the lowest with a % of Team Balls bowled value of 15.22, less than a sixth, indicating that he was, at best, South Africa's fifth bowler. On the other hand, Imran Khan bowled 27.27% of his team balls making him to be the top bowler.

The range is from 0.0 for Team Balls % of 10.0 to 5.0 for Team Balls % of 30.0.

9. BOW - % of Team Wickets (5 points):

This is another indicator of the all-rounder's place in the bowling attack. The previous one indicates the effort put in. This indicates the results. Kallis is clearly the lowest with a % of Team wickets captured value of 16.59, again less than a sixth, indicating that he was, at best, South Africa's fifth bowler. On the other hand, Imran Khan has captured 37.07% of his team wickets making him to be the top bowler. It should be noted that the team wickets are accumulated only in matches where the all-rounder bowled. This is to take care of the 7 matches in which Imran played as a pure batsmen.

The range is from 0.0 for Team wickets % of 10.0 to 5.0 for Team wickets % of 40.0.

10. ARF - All Round performances in Tests (10 points):

This is a great measure of the individual match performances. I have changed the criteria to 100+ runs & 4+ wickets for P1 performances and 75+ runs & 3+ wickets for P2 performances. These are AND conditions and not OR conditions. Since completing the previous article I have realised that this is a true measure of the all-rounder's contributions in individual Test matches since he contributes heavily to the team performance by excelling in both batting and bowling.

0.5 points are alloted for each P1 performance and 0.25 points for each P2 performance.

Sobers is the undisputed leader in this category. His performances are listed below. This table will outline his greatness. In 93 Tests he has excelled with P1 level all-round performances in 11. He has done P1/P2 level performances in 24 Tests (more than 25%).

Sobers G.St.A
P1 1960 0490 Eng Win (3+1) 4 wkts & 145 runs (145+  0)
P1 1960 0491 Eng Win (3+2) 5 wkts & 141 runs ( 92+ 49)
P1 1962 0526 Ind Win (4+1) 5 wkts & 153 runs (153+  0)
P1 1962 0529 Win Ind (0+5) 5 wkts & 154 runs (104+ 50)
P1 1966 0608 Win Eng (5+3) 8 wkts & 174 runs (174+  0)
P1 1966 0610 Ind Win (3+2) 5 wkts & 103 runs ( 50+ 53)
P1 1968 0629 Eng Win (1+3) 4 wkts & 113 runs (  0+113)
P1 1968 0636 Win Eng (3+3) 6 wkts & 247 runs (152+ 95)
P1 1969 0646 Aus Win (2+3) 5 wkts & 126 runs ( 13+113)
P1 1971 0685 Win Ind (2+2) 4 wkts & 187 runs (178+  9)
P1 1972 0695 Win Nzl (4+0) 4 wkts & 177 runs ( 35+142)

P2 1958 0448 Win Pak (1+2) 3 wkts & 132 runs ( 52+ 80) P2 1961 0506 Win Aus (5+0) 5 wkts & 85 runs ( 64+ 21) P2 1963 0546 Win Eng (0+3) 3 wkts & 154 runs (102+ 52) P2 1965 0584 Win Aus (3+0) 3 wkts & 93 runs ( 69+ 24) P2 1965 0588 Win Aus (2+2) 4 wkts & 87 runs ( 45+ 42) P2 1966 0605 Win Eng (0+3) 3 wkts & 161 runs (161+ 0) P2 1966 0607 Win Eng (4+1) 5 wkts & 97 runs ( 3+ 94) P2 1966 0609 Win Eng (3+0) 3 wkts & 81 runs ( 81+ 0) P2 1968 0643 Win Aus (4+0) 4 wkts & 86 runs ( 19+ 67) P2 1969 0654 Win Eng (2+1) 3 wkts & 79 runs ( 29+ 50) P2 1971 0684 Win Ind (3+0) 3 wkts & 112 runs ( 4+108) P2 1971 0686 Ind Win (1+2) 3 wkts & 132 runs (132+ 0) P2 1973 0726 Win Eng (3+0) 3 wkts & 95 runs ( 21+ 74)

To view all players' performances click here.

11. ARF - Average Runs scored / Wickets captured per Test (5 points):

This rounds off the individual Test all-round performances. I total the Runs and 25 times the Wickets and divide the sum by the number of Tests played. This is a clear parameter of consistency and all-round delivery. This is slightly biased in favour of the bowling since an average of 100 runs per Test has been achieved only by one batsman while 4 wickets per Test by 48 bowlers.

The range is from Hadlee (161.6) to Greig (122.8).

The point allocation is from 0.0 for ARF value of 100 to 5.0 for ARF value of 160.0 and above.

Based on these calculations the top all-rounder list is given below.

No Player           Bat     Bow     A/R    Total

1.Sobers G.St.A 26.47 14.64 12.74 53.86 2.Botham I.T 17.32 20.91 9.12 47.34 3.Imran Khan 15.11 25.90 5.72 46.73 4.Hadlee R.J 12.54 26.85 7.22 46.61 5.Kallis J.H 25.28 13.97 7.27 46.52 6.Pollock S.M 13.57 24.39 4.62 42.58 7.Cairns C.L 15.52 16.36 7.84 39.73 8.Kapil Dev N 13.80 21.12 4.10 39.01 9.Miller K.R 12.07 19.28 6.00 37.35 10.Goddard T.L 14.65 17.30 4.68 36.64 11.Flintoff A 14.52 16.86 4.89 36.27 12.Greig A.W 18.69 12.40 5.09 36.17

As expected Sobers is comfortably on top, this time with a cushion of nearly 14%. He is the undisputed leader in Batting and All-round performances and has acceptable Bowling values. What has really tilted the table in his favour is his consistent delivery of all-round performances. No one else has matched him.

Botham comes next and is just ahead of Imran Khan. He has acceptable fogures in all three areas and this has helped him move to the second place.

Then three greats occupy the next three positions and these are closely bunched. This trio is led by Imran Khan and followed by Hadlee and Kallis. They are separated by fractions of a point and should really considered as a group. Imran Khan and Hadlee excel in Bowling and have reasonable Batting figures. Kallis has an excellent batting figure and just about reaches the Bowling mark.

My suggestion to the readers is not to split hairs among these three. A slight change in parameters will move them up or down. I would say that let us take Imran as the first among these three equals. For that matter I would suggest that we take Botham as the first among the four equals since the four all-rounders are within a single % point of each other.

Then come Pollock, Cairns and Kapil Dev. Those who query Cairns' higher position than Kapil must understand that Cairns has better Batting and Bowling average figures than Kapil Dev. If people refer to Kapil's taking wickets on the Indian soil, one can point out Cairns' scoring runs on those difficult New Zealand pitches. Cairns is again an under-rated all-rounder.

Sobers is the best batsmen, followed by Kallis. Hadlee is the best bowler, closely follwed by Imran Khan. Sobers is the comfortable leader in the Test match performance area, followed by Botham.

To view the parameter values and the calculated Index values for these 12 all-rounders click here.

A look at the pre-World War all-rounders:

Using Hammond's last Test during 1948 as a cut-off, I have selected all-rounders of the pre-WW2 era, lowering the cut-off to 1000 runs and 50 wickets. The calculations have also been slightly adjusted to provide for the lower cut-off levels, lower batting averages and lower bowling averages. The table is given below.

No Player         Runs   Avge  Wkts   Avge  Index

1. Faulkner G.A 1754 40.79 82 26.59 38.99 2. Tate M.W 1198 25.49 155 26.16 34.31 3. Gregory J.M 1146 36.97 85 31.15 32.47 4. Hammond W.R 7249 58.46 83 37.81 30.15 5. Noble M.A 1997 30.26 121 25.00 29.30 6. Armstrong W.W 2863 38.69 87 33.60 26.16 7. Rhodes W 2325 30.19 127 26.97 26.02 8. Kelleway C 1422 37.42 52 32.37 23.68 9. Woolley F.E 3283 36.08 83 33.92 22.71

Faulkner, the great South African all-rounder is the leader. He is the one all-rounder who would have clearly walked into the all-time all-rounder table. Incidentally Faulkner gets an unadjusted valuation of around 34.5 points. The two Australian all-rounders, Alan Davidson and Richie Benaud are missing from the two tables. They were two genuine all-rounders who would have graced any era.

Final tweak on performance parameter:

This is a slightly revised final Table based on two very valid suggestions made by Hamish and Kartik re performances.

Hamish has suggested that since there is a clear skew, in favour of the batsmen, in fixing the P1 and P2 criteria as 100r/4w and 75r/3w, it should be changed to either P1=100r/5w... or P1=75r/4w. This is a valid suggestion, based on sound statistical analysis and the P1-P2 limits have been changed to 80runs/4wkts and 60runs/3wkts respectively. The point allocation has been changed to 0.5 and 0.3 to reflect the narrowing of gap between P1 and P2.

Kartik has made a valid suggetion to penalize the very poor performances. This makes sense since virtual non-contribution has to be recognized as we do the above-average contributions. I have taken a cut-off of 40 runs (equivalent to 1w/20r) as the point below which a test is deemed to be a failure for an all-rounder. Suitable adjustment has been made for treating tests like the one Imran played as a batsman. Equivalent runs are taken to compensate for wickets. The penalty has been fixed at 0.15, which is half the P2 value and is quite fair.

The revised table is given below.

No Player           Bat     Bow    A/R    Total

01.Sobers G.St.A 26.47 14.64 13.74 54.86 02.Botham I.T 17.32 20.91 11.07 49.29 03.Imran Khan 15.11 25.90 7.37 48.38 04.Hadlee R.J 12.54 26.85 8.72 48.11 05.Kallis J.H 25.28 13.97 7.17 46.42 06.Pollock S.M 13.57 24.39 6.37 44.33 07.Kapil Dev N 13.80 21.12 5.35 40.26 08.Cairns C.L 15.52 16.36 8.19 40.08 09.Miller K.R 12.07 19.28 6.55 37.90 10.Goddard T.L 14.65 17.30 5.18 37.14 11.Flintoff A 14.52 16.86 5.34 36.72 12.Greig A.W 18.69 12.40 5.24 36.32

As I had mentioned earlier there are no major differences. The gap between Sobers and others has come down slightly. The only significant move is that Kapil Dev and Cairns have exchanged places. The other minor change is that there is now a larger gap between Hadlee and Kallis so much so that I would consider the second best as the trio Botham, Imran and Hadlee with Botham as the first amongst equals.

This can be taken as the final table.

To view the complete table with all values, please click here.

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

RSS Feeds: Anantha Narayanan

© ESPN Sports Media Ltd.

Posted by Harsh Thakor on (January 28, 2011, 11:10 GMT)

Sorry Ananth,I again praise your efforts and basic accuracy where Sobers remains as the king and Imran is not overrated.However I feel morally Keith Miller was the most complete allrounder after Sobers.If you analyse his era and oppostion and his brilliant performances Miller was morally higher than Botham ,Imran and Kallis.Botham benefited from the absence of great players in opisng sides in the Packer era while Imran or Kalis hardly championed the cause silmuntaneously with ball and bat.Kallis is the best batting allrounder after Sobers ,but lacks the match-winning flair of Botham,Imran or Miller.I wish Keith Miller could be re-analysed as well as Kapil Dev,who received little asssiatnce from docile Indian wickets and hardly had any bowling support.Kapil,remember,unlike Botham was outstanding against the West Indies.

Posted by Harsh Thakor on (January 28, 2011, 11:08 GMT)

Ananth,with Kalli's recent brilliant performances is he morally closer to Gary Sobers as an all-rounder?Several fans rate him ahead but they ignore Gary's phenomenal all-round peformnaces in test matches and series.However one has to consider that Kallis has hardly had the opportunity in recent yaers to prove his bowling with so many bolwers dominating the South African attack.In a crisis Kallis has atleast proved Sobers equal,even if he lacks Sober's match-winning flair.Kallis is less talented but posesses the temperament and consistency of the all-time greatest-Sobers.

Posted by Harsh Thakor on (November 7, 2010, 7:50 GMT)

Godd work ,Ananth but it statistically overrates Richard Hadlee.The closeness between Botham and Imran is well reflected in the work as well as the overall superiority of Sobers who won matches consistently with bat and ball unlike Kallis.Imran was the greatest all-rounder from 1981-1987 but never championed the winning cause equally with bat and ball,unlike Sobers or Botham at his peak from 1977-1982.Infact when he was a great batsman from 1988-1992 he hardly exceleed with the ball .Botham's career suffered after 1982 and had he mantained that conistency he could well have become the equivalent of at Tendulkar to allrounders.

Kapil Dev is much better than statistically depicted if you remember the games he turned with bith ball and bat and that he bowled brilliantly on docile sub-continent tracks.Perhaps a seperate list should be made of great batting allrounders and great fast -bolwing allrounders ,which may be fairer.

Posted by Harsh Thakor on (November 7, 2010, 7:40 GMT)

Sorry,Ananth,but i really wished to come back yo you after changing my mind after the latest E.S.P.N selections.The most important factor is the criteria at choosing the greatset all-rounder.

In stats even if Kallis is the best he has never changed the complexion of a match both with ball and bat, as much as Gary Sobers. In that light Ian Botham was a champion between 1977-1982 being the best match-winner after Sobers.Imran Khan again in Stats maybe the best of the great 4 from 1981-1987 in his era but was never a great batsman and greatfast bowler in the same period and unlike Ian Botham he could not win a match with both ball and bat although he was the best match-winner at his peak with the ball.

Stats does injustice to Kapil Dev,who was morally very cloe to Botham and Imran,as he bowled on placid tracks.

Overall ,I admire this analysis as it confirms Gary Sobers as the king being the best match-winner consistently with bat and ball.

Posted by TRUEOMMENT on (October 25, 2008, 10:05 GMT)

who says kallis is not supposed to be an allrounder?he is and will be the best allrounder in cricket's history.the day is not far when he will be as good as sachin(though he is,but the so called 'sachin fan' people dont agree).he will leave a not achieveable mark of the only cricketer to score over 13000 runs aswell as bagging 350 wickets in both forms of the game(or even more)i am not saying because i am a fan but as always for losers 'truth is always bitter' however i dont mean to criticize sachin or lara..they are legends but wait..only time will tell...for me and you you all he's gonna be 'THE BEST ALLROUNDER BASTMEN'IN HISTORY

Posted by Nadeem M on (October 25, 2008, 9:09 GMT)

Continued ...

Jubilee test: Eng vs India, Bombay, 1979-80

In another analysis, BB Mama (Indian statistician) took the ratio of batting and bowling averages. Again, an interesting comparison. There was another analysis that took many parameters, and divided the players into three categories: batsmen who bowled well (Sobers), bowlers who could bat (Hadlee), and true all-rounders.

But your analysis certainly takes the cake! Thanks very much for your efforts.

Posted by Nadeem M on (October 25, 2008, 8:59 GMT)

Excellent analysis. Very fascinating. I just went through it ... not sure if you're still scanning for comments, but here are some, nonetheless.

I grew up in Pakistan, and was a regular reader (and contributor!!!) of the Cricketer (Pakistan) monthly. Gul Hameed Bhatti had once compiled a list of all rounders, using 1,000 runs + 50 wkts as criteria, and allocated one point per run, 20 points per wicket and 15 points per catch. Take the total points in the career, divide it by the tests played, and you get "performance points" per test. Sobers topped that list by miles, followed by Botham.

Another point: at the risk of committing a sin by questioning Imran's performance: how many tests did he win by his batting alone? None. He saved many, but didn't win any. During his best all-round performance (vs India at Faisalabad in 82-83: 117 + 11 wickets), he was the 4th centurian for Pakistan. Compare that with Botham's Jubilee test: 114 and 13 wickets. 2nd highest score: 49.

Posted by luke on (October 17, 2008, 19:22 GMT)

i can certainly appreciate the time and effort put into this as i have done similar exercises myself (usually to prive my theories - which, btw, i always manage to do with a tweak here and criteria adjustment there:)). However, different people faced different bowlers so unless each bowler is assigned rating (who decides comparative ratings over different eras...and how??) and then batsmen are evaluated accordingly and equally so with bowling - who are they bowling to?

btw, if you do remove the tests against zim and BD then how do other allrounders figures change and those who did not have a chance to bowl against them who were there zims and BDs (didn't NZ take years to win a test - even if they were 'fairly' competitive they must have been whipping boys...how about the others??)

in general...a good attempt and an interesting 1/2 hour read. cheers.

Posted by John Clark on (October 17, 2008, 8:16 GMT)

Just remove all of the 'tests' against Bangladesh and Zimbabwe. These matches de-value test cricket.

Kallis' figures than become: Bat. Av. 51.8, Bowl. Av. 34.0.

In addition, you are comparing players not yet retired (and hence before their end-career decline) with players whose careers are fully over, including their end-career decline, which for most players is significant. You need to project Kallis' final career totals, including an end-career decline, before comparing to retired players. [[ Ananth: John, you have made two points. One is well-made and the other not a valid suggestion. This is the not-so-good one. Let me give an example from way back. Hammond's 336 was made against a New Zealand attack, so poor that it would make the current Ban/Zim attacks seem lethal. Let me give below the career (???) figures of the six NZL bowlers. Freeman:1@ 169.00 Newman: 2@127.00 Dunning: 5@98.60 Babcock: 16@38.12 Page: 5@46.20. Weir: 7@29.86 What do here. In reality Hammond's 336 was worth, at most, a century. The point is that weak and very weak teams have been there always and we have to accept these teams and go on.

]]

Posted by shafiq on (October 17, 2008, 5:33 GMT)

The truth is we pakistanis cann't see Imran other than NO1 coz he is the king of the hearts. Anyways have you considered captaincy? Good work anyways!

Comments have now been closed for this article

ABOUT THE AUTHOR

Anantha Narayanan
Anantha spent the first half of his four-decade working career with corporates like IBM, Shaw Wallace, NCR, Sime Darby and the Spinneys group in IT-related positions. In the second half, he has worked on cricket simulation, ratings, data mining, analysis and writing, amongst other things. He was the creator of the Wisden 100 lists, released in 2001. He has written for ESPNcricinfo and CastrolCricket, and worked extensively with Maruti Motors, Idea Cellular and Castrol on their performance ratings-related systems. He is an armchair connoisseur of most sports. His other passion is tennis, and he thinks Roger Federer is the greatest sportsman to have walked on earth.

All articles by this writer