A few months back, I did a comprehensive study of Test allrounders. In it, Shakib Al Hasan emerged as the top allrounder, a whisker above Garry Sobers. This article became one of my most discussed ones in recent times. Quite a few of those who wrote to me were not comfortable with giving Sobers second place. However, an equal number were happy that the all-round skills of Shakib, a player from a less prominent team, were recognised.

In addition, there were many emails suggesting minor and major tweaks. The range of responses was staggering and these covered every aspect of analysis in both width and depth. I realised I would have to do a comprehensive overhaul of the analysis. I have now finalised the revised Test allrounders' tables with inputs from various contributors. The changes are summarised below. Also, using the changed parameters as the overall base, and accounting for the nuances of ODIs, I have worked out a list of the best ODI allrounders.

Test Allrounders Analysis
Let me first cover the revised basis for the Test allrounders' analysis. There are four groups of parameters. A move away from longevity-based measures is a feature of the revised analysis.

1. Weighted Batting Average (20 points): There was no doubt about the importance of this solid batting measure. The maximum permissible 20.0 points are given for a WBA of 55.0; the figure is proportionately lower for lower WBA values.

2. Bowling Average (20): There was also no doubt about the importance of this wonderful bowling measure. This also has a weight of 20.0 points, which is given for a bowling average of 20.0 (it is proportionately lower for higher bowling averages).

3.1. Share of team values (15): This recognises the player's contribution to the team. In order for this measure to be comprehensive, it takes into account runs scored, wickets taken and balls bowled. This ensures that allrounders who carried differing load levels are recognised. No one type of allrounder is favoured. A workload of 20% is allotted the maximum points, 15.

3.2. Consistency Index (15): This is the most important new measure added, replacing the concepts of Impact Index and High Impact Index used earlier. Many people felt uncomfortable with the black-or-white treatment of the impact indices, and the double counting, so I have split players' careers into three-Test blocks, with the last block being three, four or five Tests long. I then looked at what the player achieved in each block. (This is somewhat similar to the E-Runs measure - Equivalent Runs - explained later.)

I considered a wicket as the equivalent of scoring between 25 and 31 runs, depending on when the Test was played. Then I determined whether that was a successful block from an allrounder's point of view. A block should satisfy three criteria to qualify as successful: 300 E-Runs, 100 runs and six wickets (for a three-Test block, and extrapolated, as required, for the last block). This allows the players flexibility to handle lean batting or bowling phases, and for them compensate batting lows with wickets and bowling lows with runs, while maintaining an overall contribution.

The percentage of successful blocks is considered; 15 points are allocated for an 80% success rate. A perusal of the related numbers will show you how effective this index is. If an allrounder like Imran Khan or Jacques Kallis did not bowl at all in a few Tests, they might lose a little, but then they were not playing those matches as allrounders.

4.1. Delivery of base components per match (10): This is an important new measure based on the average contributions of runs, wickets and catches/stumpings per match. Everything is brought to an E-Runs basis: each wicket is valued at X runs, where X is the exact RpW value across the allrounder's career. It is low for matches played before the First World War, and reasonably high for the first decade of this millennium, and the range is 25.5 to 33.7 for the set of allrounders considered. Each catch/stumping is valued at 40% of the value of a wicket. The weight for this measure is 10 points, which is allotted for an average of 210.

The next two parameters relate to Performance Ratings.

4.2. Best Match Performance (5): The best all-round performance gets the player a share out of 5 points. The best here is the famous Ian Botham masterclass in the Jubilee Test in Bombay in 1979-80. This performance gets nearly 5 points; the others get proportionately lower for lower values. This is represented in the tables below by the RPP-Best column.

4.3. Average Performance per Test (5): This recognises sustained performance across a career. The weight for this measure is 5 points, which is allotted for an average of 700 rating points per match. In the tables, this is the RPP/T column.

4.4. Career length - Years/Tests (10): This was specially requested by many readers. They wanted the ratings to recognise players who performed at the highest level for many years and played many Tests. This is the only non-performance measure, relating rather to a player's longevity. The allocation of points is based on a combination of years and Tests. Kallis, with a long career on both counts, gets the highest value, ten points.

There may seem to be an overlap between 4.1 and 4.3; 4.1 is based on raw, unadjusted, basic runs/wickets/catches per Test - all brought to an E-Runs basis. There is no context there. On the other hand, 4.3 is based on a contextual, complex calculation covering the batting and bowling performances. It incorporates pitch quality, bowling/batting quality faced, innings status, support received, location/result, relative team strengths, target before the team, quality of wickets, and so on. As such, while seemingly there is an overlap, these are two different measures.

A simple comparison: Brian Lara's unbeaten 400 is 247 runs more than his unbeaten 153, going by 4.1, while the 153 gets 150 more rating points than 400 under 4.3. Also, fielding isn't accounted for in 4.3. In summary, 4.1 covers the standard player delivery of the basics, while 4.3 accounts for the outliers and extraordinary performances.

Now the selection criteria, which are the same as in the earlier article. For my first level of selection, I included players who have scored over 1500 runs and taken 50 or more wickets, or scored over 1000 runs and taken 75 or more wickets. This got me a substantial number of players. To avoid including bowlers who have played many Tests and thus accumulated enough runs (like Anil Kumble, Shane Warne, Chaminda Vaas, Stuart Broad, et al), I have set a limit of 22.5 as the lower cut-off for WBA (a batting average of around 25). And to avoid classifying Chris Gayle, Mark Waugh, Asif Iqbal, et al as allrounders, I have excluded those who took less than a wicket per Test. There was a temptation to keep 40 as the upper limit for the bowling average, but I have resisted that since it would mean the exclusion of a few players who are normally perceived as allrounders - prominent among them Ravi Shastri and Carl Hooper. A total of 51 players qualified. All teams, barring Sri Lanka, are represented. This is understandable considering that only Vaas comes closest to being an allrounder for Sri Lanka.

With these revised parameters, there has been a slight shift in the table positions. First, let us have a look at the revised table of top allrounders.

Fittingly, one of the greatest players of all time, Sobers, is at the top of the revised Test table. He has a reasonable edge over Shakib. This despite the fact that Shakib has the edge on most of the newly introduced measures. Sobers has a substantial lead in the Batting group and a slight lead in the Ratings Performance group. Shakib has an edge in the other two groups but just falls short. However, just as Sobers deserves all the accolades for finishing first, Shakib deserves recognition for continuing to give Sobers such a tough fight. Those who question Sobers' lead in batting should remember that he is, inarguably, one of the five best players to have touched a bat, irrespective of the basis used. Perhaps if I do this analysis in a couple of years again, Shakib might move to the first place, since he could potentially improve some of his numbers.

Imran Khan, Botham, Richard Hadlee and Keith Miller, the next four on the table, will be in anybody's list of top five allrounders ever. Imran, Hadlee and Miller are strong in the bowling department, while Botham and Hadlee excel in the Rating performance areas. Hadlee is ahead in the contribution-per-Test group. Each of these four has their own spheres of excellence, and that comes out clearly.

The great pre-war allrounder Aubrey Faulkner deservedly comes next. He leads narrowly from Trevor Goddard and Kallis. R Ashwin gets a well-deserved place in the top ten.

A study of the percentage values is revealing. Sobers and Kallis get nearly 25% of their points from the WBA component. Imran, Hadlee, Miller, Ashwin and Shaun Pollock are bowling allrounders. Shakib, Goddard, Vinoo Mankad, et al derive their strengths from the "Contribution and Consistency" group. Finally, Botham, Hadlee, Ashwin, Pollock and Kapil Dev are strong in the fourth group. That is how it should be. Almost all the 20 players featured are true allrounders, who are likely to shine in different areas.

Comparing the previous table with the current table, I have summarised the changes below.

Sobers and Shakib have exchanged places at the top. Imran moves up five places, Botham three. Faulkner moves down a little. Jack Gregory drops like a stone. Miller and Chris Cairns drop slightly. Ashwin and Kallis exchange places. Kapil moves up. Tony Greig, Ravindra Jadeja and Andrew Flintoff have maintained their respective places.

The table of base values is above. Here are some of the interesting values.

- The high WBA of Sobers, which stands at 51.3. Kallis' WBA is close to 50. Only five players have WBAs over 40.

- The sub-25 bowling average of Hadlee, Imran, Miller, Ashwin, Pollock and Jadeja.

- Shakib's high Runs-Wkts-Balls share of 18.7%, followed by Faulkner, with 18.2% and Mankad, with 18.1%.

- Shakib's Consistency level of 73.7% (he is followed by Dwayne Bravo, with 69.1%, and Miller, with at 66.7%). Also, note the relatively low percentage values for the players I call "specialist allrounders" - like Pollock, Kapil, Hadlee, Kallis, etc.

- The very high Base components (Runs/Wkts/Catches) per match of Ashwin, Shakib, Botham, Sobers, Hadlee and Imran. All have contributed 180 or more E-Runs per Test.

- Botham's once-in-a-lifetime performance in Bombay, fetched him 2034 Rating points. He scored 114 and took 6 for 58 and 7 for 48 in the match. The next one, some distance away, is Mushtaq Mohammad's 1798. Mushtaq scored 121 and 56 and took 5 for 28 and 3 for 69 against West Indies in Port-of-Spain in 1976-77. Next on this list is Botham's 1981 Headingley performance.

- Hadlee's average rating points per match of 666, very closely followed by Ashwin, with 665 points. Sobers and Imran have over 600 points per match.

This Ratings values chart is offered with no additional comments. The high values have already been outlined in the previous chart. The corresponding point values can be located in this table.

ODI allrounders analysis
The ODI analysis is more or less based on the Test parameters. However, new parameters have been added to incorporate relevant metrics, such as strike rates. To distinguish the Test tables from the ODI tables, I have kept the ODI maximum limit as 1000. The cut-off values are much simpler here - 75 wickets and 2000 runs; 57 players qualify.

1.1. Weighted Batting Average (125): There was no doubt about the importance of this solid batting measure. This has a weight of a maximum of 125 points, which are given for an adjusted WBA of 50, and proportionately lower for lower values. The adjustment is by period.

1.2. Batting Strike Rate (125): The maximum points here are allotted for an adjusted strike rate of 125. The adjustment ensures that the low strike-rates during the early years of the format are pushed up and vice versa for latter years.

2.1. Bowling Strike Rate (BpW-125): There was also no doubt about the importance of this wonderful bowling measure. This too has a weight of 125 points maximum, which is given for an adjusted bowling strike rate of 30 BpW; it is proportionately lower for higher values.

2.2. Bowling Accuracy (RpO-125): This is given as much importance as the balls-per-wicket measure; 125 points are given for an adjusted RpO of 3.5.

3.1. Share of Team values (125): This recognises the share of relevant team values. In order for this measure to be comprehensive, I have included runs scored, wickets taken and balls bowled. This ensures that allrounders who carried differing load levels are recognised. This does not favour one type of allrounder. A workload of 30% is allotted the maximum points.

3.2. Consistency Index (125): I split players' careers into four-ODI blocks, with the last block being four, five, six or seven matches long. I then checked what the player achieved in each block. This is somewhat similar to the E-Runs metric. This time I considered a wicket as the equivalent of between 25 and 31 runs, depending on the period when the ODI was played. Then I determined whether that is a successful block from an allrounder's point of view. The block should satisfy three criteria to qualify as successful one: 180 E-Runs, 60 runs, and two wickets (for a four-ODI block, and extrapolated, if required, for the last block). This gives the player flexibility to compensate for batting lows with wickets, and bowling lows with runs, while maintaining an overall contribution. The factor being considered is the percentage of successful blocks; 125 points are allocated for an 80% success rate. A look at the related numbers will show you how effective this index is. If some allrounders did not bowl at all in a few ODIs, they might lose a little, but then they were not playing as allrounders.

4.1. Important Tournaments - Batting (50): This measure and the one following it cover performance in the latter stages of tournaments. In general, the Super Eights, Super Sixes, quarter-finals and later stages of ICC trophies, and finals of lesser tournaments, are considered. Since the number of such matches varies considerably between players, an average performance per match is used. Fifty points are given for 60 runs per match.

4.2. Important Tournaments - Bowling (50): For bowling, 50 points are allotted for 1.8 wickets per match.

The next two relate to the Performance Ratings.

5.1. Best Match Performance (50): The best all-round performance gets the player a share out of 50 points. The best is the famous Aravinda de Silva masterclass in the 1996 World Cup final, which got 151 rating points. This performance gets nearly 50 Best Match Performance points. Other performances get proportionately lower for lower values.

5.2. Average Performance per ODI (50): This recognises sustained performance across a career. The weight for this measure is 50 points maximum, which is allotted for an average of 65 points.

5.3. Career length - ODIs played (50): This parameter is to recognise players who have performed at the highest level for many years and played many ODIs. This is allotted a maximum of 50 points. Sachin Tendulkar, with a long career of 463 ODIs, gets the highest value here, with the benchmark being 500 matches.

Shakib has proved that his close second place on the Test allrounder table is no fluke, and that he is the leading allrounder in the world now across formats. He tops the ODI allrounder table by a comfortable margin. As true allrounders do, he does not lead in terms of the base measures, but he is in the top quartile for both batting and bowling. He substantiates this with excellent performances in the other measures and has finished a deserving first. But for his ban in 2019, he might have finished even higher. However, he has come back with a bang and his recent performances have been excellent.

A surprise placement in the second position. Not many will have expected Flintoff to finish second. He has achieved this high place through a table-topping combined value of the two basic groups (Batting and Bowling): he has 350 points in these two groups, and that, with his very solid achievements in the other groups, fetches him second place. He is very close to the top in the bowling group. Kallis has had a more balanced all-round career in ODIs than in Tests and this has resulted in his well-deserved third position.

Viv Richards in fourth place? Those who are surprised should know that he has taken 118 ODI wickets to supplement his outstanding batting skills. The bowling numbers are in the lower quartile but his other achievements more than make up for those. Shane Watson comes next and that should not be a surprise since he was a genuine allrounder and a very good fourth bowler. Kapil gets a well-deserved sixth place, aided by the significant upward tweaking of his batting strike rate.

Lance Klusener might seem a surprise at No. 7, but a look at his numbers, especially bowling, shows they are almost identical to those of Imran, who appears next. Sanath Jayasuriya's ninth position should not surprise anyone, since he took over 300 wickets, and nor should tenth place for Gayle. These two were nearly proper allrounders in ODIs.

I am extremely happy with the 11th position of Shahid Afridi, whose figures are mind-boggling - over 8000 runs at a near-120 strike rate and nearly 400 wickets. Tendulkar just misses out on featuring in this table. He finished 18th.

Looking at the percentage of rating points, Richards is strongly batting-centric, as is Afridi. Flintoff, Wasim Akram and Pollock are strongly dependent on bowling points. Shakib is very strong in the contribution/consistency group and in performances in important matches - where Akram and Klusener are also strong. Finally, Afridi makes his presence felt in the other group, through his rating performance achievements.

As I did for the Test allrounders, I am going to look at the highlights of the table above.

- Richards has a high WBA of 44.3 while Kallis clocks in at 38.1. At the other end, Akram has a low WBA of 14.8.

- Afridi has an excellent strike rate, of 118. Kapil's already high strike rate, during an era of staid scoring, has been elevated to 111 after adjustment for period. As was Richards' 103. Jayasuriya clocks in at 93.

- Flintoff has a low BpW value of 34. Akram is close behind at 35. Dwayne Bravo, not featured on the table, shares the lowest value with Flintoff. Among the spinners, Mohammad Nabi required 48 balls for every wicket he took.

- Pollock had the lowest RpO of 3.8. The only other bowler with a sub-4.0 RpO is Nabi, with 3.9. Richards, Jayasuriya and Cairns were quite extravagant, conceding over five runs per over.

- Shakib's contribution percentage (Runs/Wickets/Balls) was a very high 26.6. He is the leader on this measure, by a mile. Richards, with dominating batting numbers, provided 22.4% of his team's work load. At the other end, Akram and Afridi had a contribution percentage of around 17.

- Shakib's consistency is amazing. Out of 53 blocks, he met the tough requirements in 39 blocks, leading to a success rate of 73.6%. Flintoff achieved a figure of 65.7% and Kallis 67.1%. Akram was successful in only 18% of blocks, since most times he did not meet the batting standards.

- In important matches, Richards averaged 42 runs and Tendulkar 47. Botham took around 1.6 wickets per match in such games. Tendulkar, of course, played a lot of such matches - 111, to be precise.

- The best ever all-round performance was de Silva's 3 for 42 and 107 not out in 124 balls in the 1996 World Cup final, which fetched him 153 rating points. Next comes Afridi's magnificent performance of 7 for 12 and 76 off 55 balls against West Indies in Providence in 2013, which fetched him 146 rating performance points. He is closely followed by Richards, whose 189 not out off 170 balls and 2 for 45 at Old Trafford in 1984 fetched him 141 rating points.

- Shakib averaged 61 rating points per match. Botham follows with 57 points per match.

As in Tests, this rating values chart is also offered with no additional comments. The high values have already been outlined in the previous chart. Readers may locate the corresponding point values in this table.

Conclusion
Sobers is now, deservingly, placed at No. 1 on the Test allrounders' table. He performed a triple role - premier batter, seam bowler and spinner - in most Tests he played. Not to mention his outstanding fielding skills. However, Shakib has been outstanding in two formats, and there is no doubt that he is the premier allrounder of the past four decades or so.

Calling for an all-time XV
In 2013, I conducted an exhaustive readers' poll to determine a group of 15 players to be considered for an all-time World team. There was excellent response and the results were very insightful and interesting. I now call for submissions again since new contenders have emerged, as also new measures for selection. You can email your entries through one of three routes with the subject "All-time XV - 2021".

- Send an email to my personal mail id, if you have it
- Send an email to the email id at the bottom of this article
- Send an email to the Talking_Cricket group, more on which is below.

When sending in your XV, provide your name, place of residence, and your list of 15 players (no more, no less). The team must be an all-terrain one. A manager/coach is optional. If you send multiple entries from one email id, I will consider the last one sent. Thus, you have the opportunity to change your selections. You don't have to justify your selections; I prefer short emails. I will include this message in the next two articles. I will write a summary article, which will probably be published in January. The entry that matches the final selection or comes closest to it will be acknowledged.

- eight batters/allrounders
- one wicketkeeper
- four pace bowlers
- two spinners

Talking Cricket Group
Any reader who wishes to join the general purpose cricket-ideas-exchange group of this name that I started last year can email a request for inclusion, giving their name, place of residence and what they do.

Email me your comments and I will respond. This email id is to be used only for sending in comments. Please note that readers whose emails are derogatory to the author or any player will be permanently blocked from sending any feedback in future.

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