December 1, 2011

What's so great about a batting average?

Runs, averages and wickets hardly tell the story of a player, a match or a career. We need more meaningful statistics in cricket
111

Cricket has always fancied itself a deeply statistical game, yet somehow has become increasingly bogged down in small categories of data, giving little to no thought as to how meaningful they are. Thanks in part to Michael Lewis' gift of storytelling, the word Moneyball is now not only synonymous with baseball but commonplace within business parlance to signify thinking outside current evaluation parameters. Considering the similar cadences and skill sets of baseball and cricket, it can't be long before cricket too has an overhaul of its archaic statistical processes and starts to measure what is relevant rather than simply what is easy.

That is not to say the game has completely ignored the ideas of Billy Beane and the Oakland Athletics - both Andy Flower and James Sutherland have publicly recognised that cricket is light on deep performance analysis. Flower has hired a Cambridge mathematician in an attempt to remedy that. Even so, privately held team benchmarks and key performance indicators need to not only filter to the selection table but also through to the wider cricket public. The interested spectator deserves the education and a right to form opinions on relevant data and not mere intuition.

The days of average being the leading indicator of a player's value to a team's success must surely be numbered. Perhaps its only relevance now is in attempting to compare players across generations - even then, the case made is tenuous. Cricket has changed so dramatically, it is mere superficial pub talk. We rarely recognise a player's percentage of the team's runs, let alone measure it; when wickets fall at the other end during a tight spell, who gets the recognition? A player's ability to get off strike in one-day cricket is crucial to a flowing team innings, yet scoring-shot percentage is never spoken about. The importance of bowling balls that ask a question of the batsman's length footwork is drilled in at team meetings, but never given a full statistical evaluation - particularly when it comes to deciding who gets paid what.

In Twenty20 cricket, Michael Hussey, perhaps the format's most consistent player, has his own yardstick on performance: the magic "160". It is numerically no more than the addition of average and strike rate, but it gives him a rough indicator of how good a batsman is. That is, a player who averages 60 but strikes at 100 is as valuable in his mind as some who averages 20 but gets them a better clip. It shows that you don't need to run Excel spreadsheets to find an appropriate measure.

Perhaps the first skill to break with tradition will be fielding. Catches taken certainly do not reflect someone's fielding ability or contribution to the team's performance. The Australian Argus review made direct mention of this, as did Ricky Ponting, leading up to the first Test of the summer. Perhaps we will see "runs saved or conceded" or "effective run-out percentage" flashed up on our TV screens in the coming months. We can only dare to dream.

A player's ability to get off strike in one-day cricket is crucial to a flowing team innings, yet scoring-shot percentage is never spoken about. The importance of bowling balls that ask a question of the batsman's length footwork is drilled in at team meetings, but never given a full statistical evaluation

Lewis extended his Moneyball theme in the New York Times, when he wrote about an NBA player by the name of Shane Battier, dubbing him the sport's most selfless player .When Battier is on the court, his team performs better, the opposition worse. Yet there is no hard evidence of this in the traditional, personal statistics of points scored, rebounds attained or assists given. Just like in cricket, basketball players continuously face choices between self-interest and winning. "He bats for red ink" is as ugly a tag as any to accrue in cricket, and yet the act can be seen time and time again. Wages are, after all, handed out according to averages. Someone of the ilk of Chaminda Vaas is cricket's version of Battier - content to fulfil his role as a cog in the wheel, with long, dry spells - in doing so enabling those around him to perform in dynamic bursts.

Arguably the closest measure of a player's true value to his team was devised by England's Professional Cricketers Association and has since been replicated by the Australian Cricketers Association's MVP award. While the exact algorithm is kept under lock and key, points are not only awarded for runs, wickets and catches but also for achieving such benchmarks as run- or economy rates, hundreds, volume of maidens, and even playing or captaining in winning sides.

Unsurprisingly, bowlers who can bat a little dominate the top places, but so they should - they are the most valuable commodity in the game. Think Andrew Flintoff or Chris Cairns. To put it in perspective, the top two run-scorers in the Sheffield Shield finished 9 and 13 respectively according to the MVP calculation. Bowlers certainly win first-class games of cricket and this is reflected accordingly. It may not be perfect, but it is certainly a step in the right direction. Sadly these results go all but unnoticed. The ACA could not even get a sponsor for the competition this year. For the record, Australia's most valuable player across three formats of the game was Daniel Christian.

The biggest reason why cricket is behind many professional sports on statistical evaluation is simply that the players have been public goods unavailable for hire in the open market. This has clearly been drastically changed by the privately owned franchises of the IPL. Despite the bottomless pits of money that it seems a few owners have at their disposal, and the initial irrationality that has ensued at the auctions, you don't become a wealthy businessman by buying overpriced assets and selling cheaply. It will take some time to revert to economic principles, but no one likes losing. There is also a salary cap, so "undervalued" players will always be sought after. Perhaps Australia's Big Bash League will provide a blueprint for a player's true valuation. In the IPL auction Christian was bought for a price 45 times that of Stephen O'Keefe. Both are players of similar skill sets, who have now been valued much more closely for the upcoming BBL. It won't be long before cricket too has its own Moneyball tale to tell.

Ed Cowan is a top-order batsman with Tasmania. His book In The Firing Line has just been published by New South Books

Comments have now been closed for this article

  • jay57870 on December 4, 2011, 2:51 GMT

    (Cont) However, when it comes to the question of a player's "true valuation" it's a totally different ballgame, especially in franchised sports. Moneyball can't help much, because contracts are dictated by demands of players & their agents versus what's in the coffers of the owners. As any numbers geek knows, MLB clubs have big busts too: Those deadly long-term contracts with dud stars. Another sad MLB reality: Bankruptcy is as American as baseball & Chevrolet. The Chicago Cubs & Texas Rangers have experienced bankruptcy in recent years. The storied LA Dodgers club is also in deep trouble. Perhaps MLB can learn a thing or two from the IPL with its novel auction, short-term contracts & salary caps. Remember the frugal Rajasthan Royals won the inaugural IPL trophy with the least costly player payroll! No fancy M-B model there! Importantly, how does one measure intangibles like leadership, intelligence, will-power, passion, work-ethic, staying power, etc? BBL is forewarned, Ed!

  • jay57870 on December 4, 2011, 2:41 GMT

    (Cont) Granted, batting averages don't tell the whole story. Even the hallowed "99.94" stat - as a "one-size-fits-all" metric - is overblown. As the great Don stated: "Averages can be a guide ... but are not conclusive because pitches & conditions have changed." That said, I agree that cricket stats & analysis need to be augmented with good ideas from sabermetrics - especially the "error" concept for fielding & catching efficiency. As for most valuable batting performance, I'd like to suggest Rahul Dravid as a model example. Batting effectively in pairs & building partnerships is key here. More than his 13,000+ Test runs (2nd highest after Tendulkar's 15,000+) is his striking 30,000+ partnership runs (highest) with his contribution around 42% (possibly lowest), another striking measure. Meaning he is selfless: He plays the low-key solid anchor's role while allowing strokemakers to dominate at the other end. Team first attitude is vital to success! (TBC)

  • jay57870 on December 4, 2011, 2:28 GMT

    Ed - Beware of Moneyball fever. The Hollywood M-B version is part-reality/part-fantasy. Sabermetrics - the arcane science of baseball statistics & analysis - has had some success. But the claims are often exaggerated. Case in point: In introducing the M-B system, Billy Beane & Oakland A's won 103 games & division title in 2002 with high-value bargain players. But the hidden fact: A's had 3 highly-scouted pitchers who won 194 games in 2000-03. Good pitching, like good bowling, is key to a team's success. Another reality: A's last won the World Series in 1989; but have since remained a perennial also-ran. Few of Beane's later draft picks made it. All MLB teams employ some form of M-B, but results are mixed. Many still build their core team via traditional scouting, with free agents & trades, not solely with M-B. As an early M-B adopter, the Boston Red Sox won WS in '04 & '07. This year they blew a sure playoff spot in a September collapse, possibly the worst in baseball history! (TBC)

  • HLANGL on December 3, 2011, 9:55 GMT

    I think the followings would consider all facets in deciding how influential a batsman or a bowler has been when comparing different players, especially accross different eras. It seems to be reflecting the true impact he had made in the end, not just the average which seems to the sole factor mistakenly considered by many when comparing players.

    Batting Index = (Batting Average x Batting Strike Rate x No. of Matches x Percentage Contribution to the Batting Total)/( Mean Batting Average During His Era x Mean Batting Strike Rate During His Era x Percentag of Top Order Innings)

    Bowling Index = (Bowling Average x Bowling Strike Rate x No. of Matches x Percentage Contribution to the Total Wickets x Percentage of the Top Order Wicktes Taken)/( Mean Bowling Average During His Era x Mean Bowling Strike Rate During His Era)

    This may involve lots of calculation, true, but if you need to do the comparison between different individuals, better to be complete & consider all these facets.

  • harshthakor on December 3, 2011, 5:44 GMT

    The ultimate champions challenging Bradman in his time were George Headley and Jack Hobbs.Headley outscored Bradman on wet tracks and bore the brunt of a weak batting side like no batsman in cricket history,with Brian Lara a notch behind.Lara would have been possibly the best West Indian batsman of all had he played for a champion team,with his great batting prowess.Infact upto 2009 Lara was rated ahead of Tendulkar as a test match batsman ina cricinfo analysis by Ananth Narayana.Another remarkable fact that in the 1930's Stan Mcabe who averaged 48 runs,played genuine short pitched fast stuff better than Bradman as he proved in the bodyline series.

    The batsman whose stats did them the best justice wereJack Hoobs, Gary Sobers,Graeme Pollock,Len Hutton and Sunil Gavaskar who would have mastered any attacks in any conditions.Stats of Kallis and Tendulkar are marginally inflated as the pitches are batting friendly and the bowling attacks weaker in the modern era.

  • 9-Monkeys on December 3, 2011, 4:40 GMT

    On a not dissimilar topic (how do measure greatness) this also well worth a read: www.espncricinfo.com/magazine/content/story/543468.html

  • harshthakor on December 3, 2011, 4:29 GMT

    Gandabhai ,Viv Richards would have been a genius in any era whether in Bradman's time or in the modern era.At his best even Lara or Tendulkar could not equal Viv's impact on matches or ability to turn games.Without a helmet and on the fastest of tracks he mercilessly destroyed great paceman like Lillee and Imran.In the Packer era he was arguably the best batsman since Bradman averaging 86 in the first season.Infact Viv would have thrived on the modern day flat tracks,when the bowling attacks are weaker and there is a restriction on bouncers.I would have caked Viv to average 55+ in the modern era in tests and almost 50 in the one day games.If he wished he could have broken any of the batting records.

    Ina n Inverse Tendulkar and Lara may have been greater champions had they played for teams that Viv played for .However there was a possible weakness against short-pitched fast bowling against which Viv was the supreme champion.

  • khiladisher on December 3, 2011, 3:28 GMT

    SANGAKKARA,SAMARAVEERA AND MAHELA HAVE THEIR RECORDS ONLY AT 1 CITY IN THE WORLD -COLOMBO AND IN PARTICULAR SSC GROUND-IF YOU TAKE AWAY THEIR RECORDS FROM SSC THEY AVERAGE IN THE LOW THIRTIES IN THEIR CAREERS. NOW THEY DO NOT PERFORM IN COLOMBO ALSO.

  • gandabhai on December 2, 2011, 21:08 GMT

    Sir Viv was all about dominating and swagger . Take him out of his all conqureing team and put him in Lara's or Tendulkars teams and there is no way he would have wiggled his butt walking in to bat the way he did .Sir Viv knew he was part of the most powerful cricket team the world was to witness .He had an abundance of back up .

  • BillyCC on December 2, 2011, 20:56 GMT

    @9-monkeys, I think that process is currently under way with fighters like Siddle being retained and Harris being picked when not injured and guys like Bollinger and Krejza on the outer.

  • jay57870 on December 4, 2011, 2:51 GMT

    (Cont) However, when it comes to the question of a player's "true valuation" it's a totally different ballgame, especially in franchised sports. Moneyball can't help much, because contracts are dictated by demands of players & their agents versus what's in the coffers of the owners. As any numbers geek knows, MLB clubs have big busts too: Those deadly long-term contracts with dud stars. Another sad MLB reality: Bankruptcy is as American as baseball & Chevrolet. The Chicago Cubs & Texas Rangers have experienced bankruptcy in recent years. The storied LA Dodgers club is also in deep trouble. Perhaps MLB can learn a thing or two from the IPL with its novel auction, short-term contracts & salary caps. Remember the frugal Rajasthan Royals won the inaugural IPL trophy with the least costly player payroll! No fancy M-B model there! Importantly, how does one measure intangibles like leadership, intelligence, will-power, passion, work-ethic, staying power, etc? BBL is forewarned, Ed!

  • jay57870 on December 4, 2011, 2:41 GMT

    (Cont) Granted, batting averages don't tell the whole story. Even the hallowed "99.94" stat - as a "one-size-fits-all" metric - is overblown. As the great Don stated: "Averages can be a guide ... but are not conclusive because pitches & conditions have changed." That said, I agree that cricket stats & analysis need to be augmented with good ideas from sabermetrics - especially the "error" concept for fielding & catching efficiency. As for most valuable batting performance, I'd like to suggest Rahul Dravid as a model example. Batting effectively in pairs & building partnerships is key here. More than his 13,000+ Test runs (2nd highest after Tendulkar's 15,000+) is his striking 30,000+ partnership runs (highest) with his contribution around 42% (possibly lowest), another striking measure. Meaning he is selfless: He plays the low-key solid anchor's role while allowing strokemakers to dominate at the other end. Team first attitude is vital to success! (TBC)

  • jay57870 on December 4, 2011, 2:28 GMT

    Ed - Beware of Moneyball fever. The Hollywood M-B version is part-reality/part-fantasy. Sabermetrics - the arcane science of baseball statistics & analysis - has had some success. But the claims are often exaggerated. Case in point: In introducing the M-B system, Billy Beane & Oakland A's won 103 games & division title in 2002 with high-value bargain players. But the hidden fact: A's had 3 highly-scouted pitchers who won 194 games in 2000-03. Good pitching, like good bowling, is key to a team's success. Another reality: A's last won the World Series in 1989; but have since remained a perennial also-ran. Few of Beane's later draft picks made it. All MLB teams employ some form of M-B, but results are mixed. Many still build their core team via traditional scouting, with free agents & trades, not solely with M-B. As an early M-B adopter, the Boston Red Sox won WS in '04 & '07. This year they blew a sure playoff spot in a September collapse, possibly the worst in baseball history! (TBC)

  • HLANGL on December 3, 2011, 9:55 GMT

    I think the followings would consider all facets in deciding how influential a batsman or a bowler has been when comparing different players, especially accross different eras. It seems to be reflecting the true impact he had made in the end, not just the average which seems to the sole factor mistakenly considered by many when comparing players.

    Batting Index = (Batting Average x Batting Strike Rate x No. of Matches x Percentage Contribution to the Batting Total)/( Mean Batting Average During His Era x Mean Batting Strike Rate During His Era x Percentag of Top Order Innings)

    Bowling Index = (Bowling Average x Bowling Strike Rate x No. of Matches x Percentage Contribution to the Total Wickets x Percentage of the Top Order Wicktes Taken)/( Mean Bowling Average During His Era x Mean Bowling Strike Rate During His Era)

    This may involve lots of calculation, true, but if you need to do the comparison between different individuals, better to be complete & consider all these facets.

  • harshthakor on December 3, 2011, 5:44 GMT

    The ultimate champions challenging Bradman in his time were George Headley and Jack Hobbs.Headley outscored Bradman on wet tracks and bore the brunt of a weak batting side like no batsman in cricket history,with Brian Lara a notch behind.Lara would have been possibly the best West Indian batsman of all had he played for a champion team,with his great batting prowess.Infact upto 2009 Lara was rated ahead of Tendulkar as a test match batsman ina cricinfo analysis by Ananth Narayana.Another remarkable fact that in the 1930's Stan Mcabe who averaged 48 runs,played genuine short pitched fast stuff better than Bradman as he proved in the bodyline series.

    The batsman whose stats did them the best justice wereJack Hoobs, Gary Sobers,Graeme Pollock,Len Hutton and Sunil Gavaskar who would have mastered any attacks in any conditions.Stats of Kallis and Tendulkar are marginally inflated as the pitches are batting friendly and the bowling attacks weaker in the modern era.

  • 9-Monkeys on December 3, 2011, 4:40 GMT

    On a not dissimilar topic (how do measure greatness) this also well worth a read: www.espncricinfo.com/magazine/content/story/543468.html

  • harshthakor on December 3, 2011, 4:29 GMT

    Gandabhai ,Viv Richards would have been a genius in any era whether in Bradman's time or in the modern era.At his best even Lara or Tendulkar could not equal Viv's impact on matches or ability to turn games.Without a helmet and on the fastest of tracks he mercilessly destroyed great paceman like Lillee and Imran.In the Packer era he was arguably the best batsman since Bradman averaging 86 in the first season.Infact Viv would have thrived on the modern day flat tracks,when the bowling attacks are weaker and there is a restriction on bouncers.I would have caked Viv to average 55+ in the modern era in tests and almost 50 in the one day games.If he wished he could have broken any of the batting records.

    Ina n Inverse Tendulkar and Lara may have been greater champions had they played for teams that Viv played for .However there was a possible weakness against short-pitched fast bowling against which Viv was the supreme champion.

  • khiladisher on December 3, 2011, 3:28 GMT

    SANGAKKARA,SAMARAVEERA AND MAHELA HAVE THEIR RECORDS ONLY AT 1 CITY IN THE WORLD -COLOMBO AND IN PARTICULAR SSC GROUND-IF YOU TAKE AWAY THEIR RECORDS FROM SSC THEY AVERAGE IN THE LOW THIRTIES IN THEIR CAREERS. NOW THEY DO NOT PERFORM IN COLOMBO ALSO.

  • gandabhai on December 2, 2011, 21:08 GMT

    Sir Viv was all about dominating and swagger . Take him out of his all conqureing team and put him in Lara's or Tendulkars teams and there is no way he would have wiggled his butt walking in to bat the way he did .Sir Viv knew he was part of the most powerful cricket team the world was to witness .He had an abundance of back up .

  • BillyCC on December 2, 2011, 20:56 GMT

    @9-monkeys, I think that process is currently under way with fighters like Siddle being retained and Harris being picked when not injured and guys like Bollinger and Krejza on the outer.

  • S.Jagernath on December 2, 2011, 19:23 GMT

    Batting averages these days are often skewed statistics,especially regarding the Sri Lankan cricketers.The manage to average 50+ due to the fact that they tour less than the other 7 test sides out of the top 8.Batsmen like Mahela Jayawardene,Thilan Samaraweera & Kumar Sangakkara have averages that inflated by a large amount,especially with a poor quality cricket ground like the SSC in Colombo.

  • harshthakor on December 2, 2011, 19:21 GMT

    Majid Khan averaged 39 runs,but in his peak and on bad wickets matched the class of Viv and Barry Richards.Kalicharan averaged 44 runs but was technically the best left-hander ever.Arthur Morris averaged 46 runs but was the greatest of match-winning openers in the view of Bradman just like Gordon Geenidge who averaged 44 runs was in the 1970's and 1980's.Another stalwart was Aravinda De'Silva who average around 42 runs. arguably the best batsman of his time.On his day Neil Harvey who averaged 47 runs could match even Graeme Pollock.

    The best batsman whom stats never did true justice were Rohan Kanhai,Victor Trumper,Ian Chappell,Graham Gooch,David Gower,Adam Gilchrist,Ted Dexter,Arthur Morris,Gundappa Vishwanath,Majid Khan,Gordon Greenidge ,Neil Harvey ,Zaheer Abbas,Alvin Kalicharan,Frank Worrell ,Aravinda De Silva and Mark Waugh.On their day or morally they matched the alltime best who averaged 50+.

  • 9-Monkeys on December 2, 2011, 18:44 GMT

    They say 90% of elite sport is played between the ears. In our game, with the exception of freak talents like say a Mark Waugh or a Shane Warne, the best 15-20 players in Australian cricket at any given time are pretty closely matched on talent. What elevates the better players is attitude and consistency, both mental aspects of the game. One of the failings of the Australian system following the retirements of Warne, Langer, Hayden, McGrath and Martyn - massive loses that were always going to hard to replace, has been a heavy emphasis on statistics and not the man. I now hope with the involvement of Inverarity and Marsh - deep thinkers, wonderful strategists and proven man managers, our selections will improve. Like it did in the late 80's when it elevated Marsh, Boon, S. Waugh, Hughes and McDermott, Australia needs to find the fighters, the men who will set the attitudinal standard and raise the team's performances. To do this the selectors must look beyond Sheffield Shield averages.

  • harshthakor on December 2, 2011, 15:36 GMT

    I forgot the example of Adam Gilchrist,the best match-winner of the modern era who averaged 49 runs and 60 for the first half of his career.Sangakaara averages 56 but I would always rate Gilchrist ahead as he could change the complexion of a game like no one.Viv Richards playing for a champion team infact went against him stats wise as he did not have as much of an opportunity to contribute.Herbert Sutcliffe averaged 60.83, 4 more runs than Jack Hobbs but Hobbs is always rated one class ahead as he championed bad wickets and won many more a match.Similarly inspite of Graham Gooch averaging 5 runs below Boycott who averaged 47.72,I rate Gooch ahead as he was far more exhilirating and combative.

    Had Bradman batted in Gavaskar's era he may have averaged around 70 runs.

  • BoratShah on December 2, 2011, 15:09 GMT

    I agree with what you say from a viewer's point of view. However, if fielding stats are collected, are we going to pick or drop players based on fielding stats? Do we have specialist fielders? At best, it can only be used as a tie-breaker.

  • cmw2175 on December 2, 2011, 14:07 GMT

    Another really good article Ed. I suspect that the more relevant statistics will play an increasingly large role in the game as time goes on. Although, it's only the measuring of these things that is new. Players have always had an idea about how often another player is LBW, whether he scores mainly on the off-side or leg-side, but measuring it adds weight to the arguement.

  • allblue on December 2, 2011, 13:26 GMT

    @Anirudh Singhi Thanks for your response and yes, you make a good point. I must admit I was thinking just in terms of the late order here and as your example shows such a stat could distort badly when applied to the top order. But then as the article says, there is no single definitive stat in cricket and distortions will occur, rather like a late order batsman who gets an inflated average by having a high proportion of not outs eg Jimmy Anderson. The idea for this first occurred to me a few years back when I noticed that Glen McGrath seemed to be involved in a number of substantial late partnerships where his personal contribution was relatively small, but there was no way of evaluating that benefit to the side over a career or even a series. Still, with modern database technology us cricket nuts can look forward to a whole new range of stats to pore over (and we do love a stat don't we!)

  • harshthakor on December 2, 2011, 10:59 GMT

    I wish to illustrate the following

    1.Kanhai v.Sobers -Sobers averaged 57.78 runs,Kanhai averaged 47.53 ,yet many rated Kanhai better.Rohan batted one down and at his best could make a better impact.

    Ian v Greg Chappell.-Greg averaged 53.8 against Ian's 42.4 but in a crisis Ian always performed better than Greg and oustshone him on bouncy tracks.Many placed Ian ahead.

    Mark.v Steve Waugh-Steve averaged 51 runs while Mark 42 but mark in full flow was a better batsman particularly on difficult pitches .

    Ananth Naryana made an analysis in 2009 where he considered match performances + other factors and still Walter Hammond and Greg Chappell did not make the first 10.Playing for a weaker team infact enhanced Brian Lara's rating to 2nd only to Bradman while Viv Richards was relegated to only 7th position,even behind Ponting, Dravid and Kallis.

    It proves cricket is far above mere stats or any rating programme.

  • 2929paul on December 2, 2011, 10:12 GMT

    First of all, let's wrap up the Bradman debate. This is one instance where his batting average is a true indicator of how good he was and his value to the team. He average almost 40 runs per innings MORE than any other batsman who played during his era. That is truly remarkable, even if you believe that the bowling was no good and the conditions were easy - which itself is supposition by posters here, based on statistics gleaned from this site rather than knowledge of the players of the time. Secondly, Moneyball theory would mean Tim Bresnan is an automatic selection for England's Test team. They have a 100% record when he plays, and his personal stats are very impressive. BUT, by common consent he ranks below Anderson, Broad and Tremlett as a bowler when all are fully fit. He has only previously played when one of these has been injured. So what will happen if all of them are fit? Suddenly it all becomes a subjective decision again.

  • Vinodha19 on December 2, 2011, 6:52 GMT

    as for in odi there should be a column for batsman that how many percent runs have been scored in the winning matches (for ex dravid has scored 24% runs out of india's total runs (2000-05)) In test along with the bat.avg there should be balls.faced.avg or time spend per innings these shows that how batsman is reliable during the time of match saving. And for the bowlers there should be column of two wkts list. one of the batsmen wkt and other of non-batsmen wkts(7-11) these shows there quality.

  • ashish514 on December 2, 2011, 5:51 GMT

    I disagree to most of your recommendations. If the yardstick of a good odi player is the ability to rotate strike, the sehwags, yuvrajs, gilchirists and jayasuriyas will be out of contention(Sachin & Ponting though will still be there). Only some combination of average and strike rate, something like hussey's magic number for T20's will correctly depict top performers. Then there's a paradox in the statement when it comes to bowling. If a bowler's ability to swing the ball and bowl good lengths is the right parameter despite the no. of wickets taken or matches won, why is a batting performance measured on the winning contribution and why not on his stance, backlift, distance b/w bat and pad and other technicalities. The only way bowlers can win matches is giving less runs and taking more wickets which is depicted by bowling average. And if u think Sachin is only statistics, watch the thrill on opposition's faces when they take him out, that's the best yardstick you'll ever have.

  • ARad on December 2, 2011, 5:42 GMT

    Ed Cowan was NOT WRITING ABOUT INTERNATIONAL CRICKET. There is a good reason for that. The article is about assessing VALUE IF MATCHES ARE MOSTLY PLAYED IN COMPARABLE CONDITIONS AGAINST COMPARABLE OPPONENTS IN A LIMTED PERIOD OF TIME. MLB players more or less play against the same players under similar conditions (with a few exceptions) and a single season itself may have sufficient data points to evaluate the player. It is impossible to transfer this to International Cricket, especially Test cricket. Conditions drastically change from day to day, toss matters, players change across series (No Zaheer in Eng) and the number of data points in a year are typically too few, among other factors, to reasonably COMPARE DIFFERENT PLAYERS BATTING AT DIFFERENT POSITIONS IN DIFFERENT TEAMS. (Even in IPL, we only get 14-16 'at bats' per season and each 'at bat', depending on bowler's quota mgmt and current score, would be vastly different.) That said, I would like to see people trying new metrics.

  • harshthakor on December 2, 2011, 3:26 GMT

    A very important factor is the batsman's match performances taking into account all factors-opposition,pitch percentage score of team's total score,effect on match etc.In that light Brian Lara ,George Headley,Jack Hobbs and Viv Richards. are closest to Bradman.Playing for a weak team infact gets a great batsmen a better rating ,with Lara being the best example.

    Overall Cricket cannot be defined by averages.Victor Trumper averaged only 39 run s but on treacherous wickets he was the ultimate champion.

    The era is also an important factor and in the modern era whee the pitches are far more bowler friendly and the bowling attacks weaker batsman like Viv Richards,Gavaskar and Greg Chappell may well have averaged 4-5 runs more.

    Averages will never do justice to greats like Zaheer Abbas,Ian Chappell,Rohan Kanhai,Gundappa Vishwanath,Ted Dexter,Victor Trumper or Mark Waugh who on their day matched any all-time great but averaged below 48 runs.

  • harshthakor on December 2, 2011, 3:13 GMT

    Some of the greatest batsmen of all time have not had great batting averages or aggregates.Take the example of Ian Chappell who averaged 42.42 runs and Rohan Kanhai whi averaged 47.53..Both averaged below 50 but were considered by some experts in their eras to be the best batsmen of their eras ,ahead of batsmen who averaged 8-10 more run sper inings.(eg.Greg Chappell,Gary Sobers Graeme Polock etc.)Ian Chappell was the ultimate batsman in a crisis in his time while Kanhai's batting at his best overlapped even Bradman.Another example is our own Gundappa Vishwanath a better match-winner than Gavaskar and a champion on bouncy and bad wickets who averaged around 42 runs.

    The most important factors are the runs scored in a crisis ,the nature of opposition and the pitch.However even Ananth Narayan's 2 stats analysis on cricinfo has not done full justice to great batsman,with batsman like Lara and Ponting ahead of Viv Richards and the Chappel brothers and Kanhai so low.

  • on December 2, 2011, 3:11 GMT

    @allblue i agree with you but the problem with the stat is that even though there are 10 partnerships, one player might be involved in more than 1 partnership. Lets look at at a very simple hypothectical example. let sehwag and ghambir open the innings for india. Sehwag goes for a duck at the score is 0-1. the patnership between gambhir n sehwag is 0. now sachin comes in. gambhir scores 100 while sachin scores 50. there patnership is 150. but the partnership average for gambhir is (0+150)/2 =75 but for sachin its 150 even though his contribution was only 50. So a player stats cannot be influenced by the performance of another player.

  • prashant1 on December 2, 2011, 1:57 GMT

    @Gary Williams. Frankly, I am not sure whether being in a weak team hampers the stats of a batsman. For eg. SRT had his best stats over a prolonged period through the '90s with the Indian team absolutely pathetic ,esp. away. Lara had his best period late in his career -again with the WI team at his weakest during his time. Perhaps Viv would have averaged more if he too had to be more circumspect coming in at 20/2 most times and gone about more carefully constructing more hundreds and big innings..instead of coming in at 100/2 and charged with the role of blasing away and demoralising the opposition...Almost impossible to distill individual performances from overall team dynamics (both own and opposition) in a team sport.

  • on December 2, 2011, 1:06 GMT

    I found some comment really funny here, because according to some average has no value whic is ridiculous. Can I ask a questions to all the experts here who think average does not tell a player story? How many you have seen Bradman played live? Practically very few people have seen him play in live. But why most agree he is the greatest, it's because his record and stats. Another funny thing is that people always bring Sachin in any article discussion. Come on guys give him a break, if you guys agree or not he is always be the second greatest batsman in the history of world cricket. In addition to that, I always find it interesting when people say Viv was such such such great. Average does matter and look at Viv records against two best sides he played against, Australia and Pakistan. Against both of them his test average is in 40s. In Odi his record against Pakistan is abysmal, 30. His record against England is always standout. What's does it tell you? Average does matter or not.

  • allblue on December 1, 2011, 23:50 GMT

    I think there is one missing batting stat that would be useful to know and relatively easy to compile, namely the average partnership a batsman featured in. A total consists of 11 innings plus extras, but equally it consists of 10 partnerships, and this stat would be particularly useful for determining the contribution of the lower order. For example, two number 11s both average 8, but the first has reasonable defence and sticks around for his partner, while the other has a quick flail and it's all over in a flash. The former (say) has a partnership average of 20, while the latter only 10. The first is clearly more valuable to his team's innings than the second, but nowhere is this recorded statistically.

  • Dravid_Pujara_Gravitas on December 1, 2011, 23:13 GMT

    @Alexk400, I agree with you wholeheartedly that Average is a very flawed way of looking at how great a batsman is. And to think that people want to compare the average of two batsmen (two flawed fractions) to say who is better, simply beggars belief! This great game is much more than mere numbers and flawed fractions.

  • straight_drive4 on December 1, 2011, 22:44 GMT

    great piece - love reading your articles Ed

  • Alexk400 on December 1, 2011, 22:37 GMT

    Stats are over rated. When a master (so called) batsman like sachin playing to be not out than protecting that tells the true story. Some Stats hide real worth of player by hyping them. Some stat like average of person without NOT OUT and Winning games. That tells a batsman is a key to winning a game. For example wickets per game may not tell the story if the team is losing. team can lose for so many reason but if a bowler strike every 30 ball is a most powerful stat for a bowler.There are good stats and flawed stats. batting average in current definition is massively over flawed. May be to hype bradman stats?. There is always some ugly thing hiding in every closet.

  • SBTang on December 1, 2011, 22:30 GMT

    Thanks Ed. Quality.

    Beane employed two broad strategies -- the application of non-traditional metrics to the data (eg on-base percentage) and the re-interpretation of conventional metrics (eg walks).

    Ed helpfully focuses on the former. But the latter is equally interesting if only because it appears, at first glance, to be so obvious as to go without saying. In any professional sport, it seems implausible that management would ignore well-established metrics.

    But Beane realises that that is precisely what traditionalists are guilty of -- they evaluate a young player on the basis of "what he looks like, or what he might become" rather than "what he has done".

    This dichotomy between statistical results achieved at the second-highest level of the game and aesthetics as seen through the the eyes of the establishment, permeates cricket as much as it does baseball. Nowhere is it better exemplified than in the contrasting plights of Phil Hughes and Shaun Marsh: astraightbat.wordpress.com

  • Leonb on December 1, 2011, 22:23 GMT

    @snowsnake - Bradman batted 3 times at #7 (ave 162!), 8 times at 6, 3 times at 5, 10 times at 4 (worst position 'only' averaged 53), and 56 times at 3. He batted 15 times in 4th innings averaging 73.4 with 3 hundreds and 4 50s. And wickets are now covered and cared for carefully with modern covers that mean there is little chance of damage from the elements. From 1934 to 1957 the laws allowed for protection against rain for the 24 hours prior to the match and then only the 3'6" in front of the popping crease after the start. Must have made for some difficult 4th innings chases! Finally, of the 40 occasions where the team batting 4th has the highest innings only 6 were pre 1957 and 15 SINCE 2000. I think 4th innings and 5th day batting is much easier these days

  • BillyCC on December 1, 2011, 22:20 GMT

    @Snowsnake, some interesting points you make. Here are my thoughts: Timeless tests are not a factor in this debate. Happy to be proven wrong but a batsman vs bowler situation has always been a ball-by-ball contest. You still need the right technique to bat "timelessly". Number of tests played per year is a positive and a negative. Positive for batsman in a purple patch (Dravid's recent form) and negative for batsman in a form slump (Ponting's recent form vs his five year purple patch earlier in career). Formats: the good players know how to adjust, the great players won't even need to adjust. Number of competing teams is important. In the 1980s, there were five truly strong teams and bowling attacks. These days, there are only two real competing bowling attacks (South Africa and England). Therefore, results against those two means more. I agree with you on rest days but there are special players in cricket who are "in the zone" and would prefer five straight days rather than resting.

  • nate63 on December 1, 2011, 21:27 GMT

    Great article! Love your work Eddy. The possibilities are endless, for example who would be have the best average if you weighted the bowling they faced, so the lower the bowlers average the higher the weighting of any runs scored against them. In this scenario I cant imagine Hayden stacking up against Gavaskar as an opener. What about openers for that matter, do they get extra weighting for facing the new ball? Then there are bowlers, what is the value Pontings wicket compared to Mcgrath, they both say 1 in the book right. How do account for the momentum shift in a game, for example a partnership breaking wicket or tight spell of bowling and wicket falling at the other end. There are a plethora of permutations to be discussed over a drink!

  • HLANGL on December 1, 2011, 21:25 GMT

    To me, the true greatness of a batsman should be decided by the fact that how frequently & how quickly he can impose his authority in the middle & can take control of the game. The adaptability & being able to be quickly synchronized with the match situation should be the key. No matter what the average is, whether it's 50 or 60, if he doesn't have this quality, he doesn't belong to the league of greats.In this yardstick, even a player with lesser averages would have been a greater batsman than the ones having higher averages, because of the sheer impact he's capable of bringing to the game compared to the ones with higher averages yet largely uninfluential batting. That's why Viv Richards always stood out from the rest.The players like T'kar, Lara, Ponting may have come close, but not quite there with Richards when it comes to the sheer impact. Gilchrist, J'ya, Saeed Anwar, etc. also had the same impact at their peak, but none of them could match Richards when it comes to consistency.

  • HLANGL on December 1, 2011, 21:22 GMT

    To me, the true greatness of a batsman should be decided by the fact that how frequently & how quickly he can impose his authority in the middle & can take control of the game. The adaptability & being able to be quickly synchronized with the match situation should be the key. No matter what the average is, whether it's 50 or 60, if he doesn't have this quality, he doesn't belong to the league of greats.In this yardstick, even a player with lesser averages would have been a greater batsman than the ones having higher averages, because of the sheer impact he's capable of bringing to the game compared to the ones with higher averages yet largely uninfluential batting. That's why Viv Richards always stood out from the rest.The players like T'kar, Lara, Ponting may have come close, but not quite there with Richards when it comes to the sheer impact. Gilchrist, J'ya, Saeed Anwar, etc. also had the same impact at their peak, but none of them could match Richards when it comes to consistency.

  • on December 1, 2011, 21:13 GMT

    Agreed. I've always valued a bowler higher if they can dismiss the higher order/scoring batsmen. Sorta like how the matches vs the minnows can be excluded from bowling average analyses. I've asked S Rajesh for these kinda stats. Hopefully that's possible.

  • on_the_level on December 1, 2011, 20:51 GMT

    There are lies, damned lies and statistics. My view is, just enjoy the game, and don't worry about who is the "best"!

  • LeftBrain on December 1, 2011, 20:22 GMT

    well such a thoughtful ad iteresting article. Some sese to cricket analysis after all those years of useless number crunching. if somehow cricket start using some meaningful analysis then the easy to clculate ones, I guess the first three victims would, ICC raking system, Indian batsme ad Cricinfo "stats champs"!!!!!

  • brittop on December 1, 2011, 19:53 GMT

    @ravichakra: Funny how every article gets it's "Sachin comments" regardless of what it's about. I don't know what you want from him. Sachin is a batsman whose job is therefore to score runs. If he scores a hundred, but the team don't win, is that his fault? Maybe the other batsmen didn't score enough runs or the bowlers didn't bowl well enough to get the opposition out twice. Tell me why you say his innings aren't "significant", but Dravid's and Laxman's are. It's not as though he has scored loads of 100s in a meaningless 4th innings of a match (only 3 in 4th innings, one of which won the match).

  • SnowSnake on December 1, 2011, 19:42 GMT

    It is well established that strike rate and average are inversely proportional. Also, average of middle order batsman is likely to stay higher due to not outs in drawn tests or tests where there is a win. Timeless tests, number of tests played/year, number of formats a player has played, number of competing teams, availability of technology all make comparison based on average unreliable. With Bradman, there are no videos available for criticism of this batting. Looking at number of boundaries, the fact that he started his career as 7th down batsman and played all but his last test as timeless tests does not look well for him. Also, a fact that is missing today is the number of rest days in test cricket. Batting on 5th day is difficult because there are no rest days. This was not the case in the past. It used to be 4 days of game, 5 day rest and then 6th day of game.

  • BillyCC on December 1, 2011, 19:40 GMT

    @liz1558, extremely well put. As for my opinion on greatness, I think averages tell about 75% of the story. For a batsman, other factors in greatness include: The number of hundreds vs tests played, longevity, contributions under pressure in the second innings, the quality of the opposition, contributions in wins and losses but not draws, how you bat with the tail if you are a lower order batsman, quality of partnerships you form if you are a top order batsman. What is not important in judging greatness include things like: strike rate (is already a component in many of the above factors), the support and supposed pressure of billions of fans, chances given (natural part of the game), quantity (as opposed to longevity which should matter).

  • on December 1, 2011, 19:33 GMT

    Interesting article and something that is not often appreciated is the role of those around you in helping your figures. If the bowler at the other end is leaking runs all over the park your chances of taking wickets will drop as batsmen are content to play safely against you while attacking the other bowler.

    How many wickets did Glen Mcgrath earn for his bowling partners as batsmen found it very hard to score against him so perhaps attacked the other bowlers.

    The same applies to batsmen. How much easier would it be to bat with Brian Lara in full flow knowing that even if you go 3 or 4 overs without scoring Lara at the other end was still keeping the team in the game. Relieves the presure to play the stupid shot. Also, how many middle order Australian batsment benefited from know that Adam Gilchrist at his peak was due in next.

    These are subtlies that statistics don't really get to the bottom of. The same applies in other sports such as football.

  • Dravid_Pujara_Gravitas on December 1, 2011, 18:44 GMT

    Statistics relies on numbers from controlled conditions. It's just impossible to imagine that Cricket stats are Statistics in the first place. Two different batsmen facing totally different deliveries. If two batsmen face 150 deliveries each, there will hardly be two 'same' deliveries in those two innings. What is common between those two batsmen during their respective innings, and careers on a grander scale, to compare and come to a conclusion Statistically? NOTHING! Cricket performance is not a comparative experiment or a clinical trial done in controlled conditions. Hell, clinical trials even have confidence intervals, p-values, errors etc.. Hence, Cricket stats are a joke, to even call them as Statistics. That's why I always say, cricket is much more than numbers. Not for no reason there are people like me who hold Dravid, Lara and Kallis in higher regard than the Little Master Sachin. All these are greats and none can be concluded as superior to the other based on Cricket stats.

  • on December 1, 2011, 18:20 GMT

    Maintaining a high batting average through a career does not indicate consistency. It might just mean that the player did well at the beginning. If you series averaging 100, 0, 100, 0, 100, 0, you may never have an overall average less than 50, but can hardly be called consistent. Miandad's average never went below 50. However, Bradman never averaged below 50 in any series he played in - that's much more significant.

    As far as rating players is concerned. If I were a coach I would divide runs scored not by the times given out, but by the times that the player should have been out. An easy dropped catch counts as an out. A clearly incorrect LBW counts as not-out. Marginal cases could count as a half-out. However, this is all subjective and I cannot see how it can be made into anything official.

    I would note, however, that baseball counts fielders errors and they count against the fielder and not against the pitcher. It is well-known that this is rather subjective though.

  • on December 1, 2011, 18:10 GMT

    Statistics should be further broken down on following basis.

    1. Geographical location - Lord vs. Faisalabad 2. Bowler / Batsman faced 3. Rain hit matches specially in England 4. Cloud cover 5. Wind factor 7. What position batsman playing 8. Pitch condition 9. Sunil Gavakar/ Wasim Raja played against the West Indian when they were at their peak 10. Mudassar Nazar / Tasleem Arif played at Faisalabad against Lillee both of them scored double centruries. What was that statistics. 11. Runs scored under tense condition.

  • sameer111111 on December 1, 2011, 16:39 GMT

    Awesome article. Looking just at Sachin's statistics many people would think that he would have at least 15 innings among the top 100 Wisden innings, but in reality it's zero. On the other hand there are impact players like Sehwag who would fail in 4 innings but the fifth one would be a match turning one. Moreover a Dravid or Laxman would perform when it matters the most and in an ideal world, those innings should be getting higher weightage. For too long many batsmen have been playing under ideal batting conditions with zero pressure and inflating their averages, rendering them meaningless. For example, Sachin's highest score is against Bangladesh in Bangladesh which would greatly increase his away average, but we all know how much that is worth.

  • RKB21 on December 1, 2011, 16:26 GMT

    It is also important to take in account whether a batsman scores consistently against the top 2 teams of his era. Sunil Gavaskar played well against WI during his career and Sachin has played well against Australia when they were the top team during the past 10-12 years. This is where Bradman's stats can be questioned as he only played against 4 teams (SA, WI, and IND only at home), and England (home and away) playing on the same pitches and against the same quality of bowlers.

  • Beazle on December 1, 2011, 16:14 GMT

    I saw Ken Barrington bat on a number of occasions and he was a fine player- about as good as Jonathon Trott. Yet Kenny averaged 58.67 in scoring nearly 7,000 runs. No-one- including Ken, would have dreamed of saying he was anywhere near as good a player as- say- Viv Richards who averaged "only" 50.

    In other words - averages only tell you half of the story. Do not let statistics blind you to the true soul of cricket.

  • vellupillai on December 1, 2011, 15:53 GMT

    The article inadverently points to a very valid reason for a Game like ODI's to be stopped forever,Odis only promote averages and statistics and it clearly give advantages to a certain number of players in a team (eg. openers and number 3,4)with pitches as flat as roads no wonder we have players like tendulkar who have tons of runs in ODI but if we convert that runs to success of his team its amount nothing.

  • on December 1, 2011, 14:21 GMT

    The average is likely to be higher if you are an opener in the one dayer and even more if you play mostly on subcontinent pitches. The higher middle order batsman may be able to do better than lower middle order batsmen. For example Sachin averages in mid 30s in one day matches when he batted in the middle order, but his averages is in late 40s as an opener in one day games. There is a huge advantage. Similarly Sachin did not score a single hundred until he was promoted to open the innings in one dayers, but he started scoring array of hundreds once he was promoted as an opener in the one dayers. Similarly, the batsmen batting higher in test too have an opportunity to bat longer and get good records. Again, the ones who are playing most in subcontinent will have a higher averages. Pitches all over the world these days have become more suitable for batting. So the batsmen in this generation will have higher average than their predecessors with the same ability.

  • GraemeH on December 1, 2011, 14:11 GMT

    @ansram but what if batsman 2 made his 197, 165 and 122 against "the minnows". Daryl Cullinan's average (and awesome strokeplay) didn't mean much when he faced Shane Warne.

  • GraemeH on December 1, 2011, 14:04 GMT

    Thoughtful article - I have thought about this a lot since reading Moneyball. I also remember writing an actuarial exam where we were asked to improve the batting average statistic to more accurately reflect exposure (there is a difference between 100 not out and 20 innings of 5 not out) and this is just the beginning. For bowling, I have always thought that bowlers should get more credit for the wickets of better batsmen - getting 3 tailenders is not the same as getting even 1 tendulkar or kallis or de villiers - which would make defining the better batsmen even more important. Separately, while muralitharan was truly great, his wickets hauls were helped by no-one really getting wickets at the other end.

  • 2929paul on December 1, 2011, 13:59 GMT

    Not sure what the point is about Vaughan's averages. In fact, it kind of prove's Cowan's point in his article. Batting averages are a meaningless statistic. Vaughan's true value to his team after 2003 was as a captain but I'm not sure how you measure the decisions taken in order to prove that. Perhaps the most black and white figures could be that in 2005 England beat Australia 2-1 when Vaughan was captain and in 2006/07 they lost 5-0 when he was injured and Flintoff was captain. I know there were loads of other factors to consider but the Vaughan captaincy factor was a major one.

  • on December 1, 2011, 13:04 GMT

    Robin Smith, the forgotten man of English cricket, after his first 5 innings, his average never dropped below 43! How good is that! In other words we need a stats indicator that measures consistency, above a certain baseline. Graham Gooch would not be so great then, his first 15 years of Test Cricket yielded an average of 36. It was thanks to the time he had in 1990 -1993 that he pulled it up to 42.

  • BellCurve on December 1, 2011, 13:02 GMT

    Cricket is a zero sum game. Each ball is a contest within a contest. These premises make it possible to formulate an approach to measure each player's contribution to a match. Ball-by-ball data is already available. The modeling would only take a couple of days. So I don't know why contributive metrics do not already exist? Maybe because most cricket "statisticians" are only interested in proving their own national heroes are the best and the greatest? Maybe because most cricket fans are only interested in seeing lists with their national heroes at the top?

  • on December 1, 2011, 12:57 GMT

    In Tests a batting average is probably the best indicator of the worth of a batsman, with 40 being the benchmark in the 1980s/1990s and 50 being the benchmark now. Michael Vaughan is regarded as a supreme English batsman, and had an average of 41. He played in the 2000's mostly and that would probably account for about 37 in the 1990s. If you look closer, he averaged 31 between 1999 and the start 2002 English summer. He then has a golden 6 months, against India, and then in Australia where he averaged a mammoth 76, scoring 7 centuries. However, from January 2003 until his last Test in 2008, he averaged just 36 over 54 matches! So his average was maintained by 6 great months, which he lived off for the rest of his career. Take a look at David Gower. Micky Stewart used to say his was inconsistent, but his average never dropped below 40 in 204 innings, and he played the bulk of cricket in 1980s, when an average of 40 meant something. I will continue below...

  • ansram on December 1, 2011, 12:39 GMT

    Average tells you part of the story. Here is an example. Take the following two batmen.

    Basteman 1:

    60 45 80 23 102 104 3 52 45 48 - Average = 56.2

    Batsman 2:

    0 5 197 41 9 165 122 0 5 17 - Average = 56.1

    Batsman 1 is consistant while never really peaking, while batsman 2 is inconsistant but makes it count big on his day when it really matters.

    Sachin is category 1 batsman. Sehwag, Lara etc are type 2. I would want batsmen of both types in my dream team, I would depend on type 1 to provide stability and type 2 to win me matches.

    Bradman was both!!

  • on December 1, 2011, 12:28 GMT

    In tests avaerages tells the story mostly..but in ODI .it is not..it is a mixer of average and srtike rate and actually everything...sanath jayasuriya and aravinda de silva were one of the greatest match winners in ODI.but there ODI avarages were 33 and 35..so it tells the story isnt it? who would u pick to win a ODI game..jack kallis or aravinda/?? sanath or dravid?

  • HumungousFungus on December 1, 2011, 12:20 GMT

    An excellent article again Ed, and this hits the nail on the head in many respects. An example of how statistics can be misleading: I played league cricket in Wales in the late 80's / early 90's at a time when there were no restrictions on how many overs a bowler could bowl, and the most important player in our team was a medium pace bowler (in his 40's) who, every Saturday, whatever the pitch or weather conditions, would bowl 24 remorselessly accurate overs in a row, generally uphill and into the wind. His average bowling figures would be something like 24-8-45-2. This would allow me (considerably quicker, considerably less accurate) to bowl 5-8 overs flat out with the new ball, not worry about how many I went for, and concentrate on taking wickets. Every year I'd take many more wickets than him, at a better strike rate, and a lower average. Was I a better bowler or more valuable to my team? Not a chance.

  • on December 1, 2011, 12:11 GMT

    In tests avaerages tells the story ...but in ODI they dont..sanath jayasuriya is the greatest match winner in ODI cricket upt o now..but his ODI average was 33...aravinda de silva was a agian such a great batsmen..still his average was 35..this tells the story..who would u pick...jack kallis or aravinda de silva to win a OID match....or jayasuriya or dravid? ..it tells the story

  • on December 1, 2011, 12:02 GMT

    Wonderful article Ed. It's rather funny when you think of how much cricket has gone forward during the last decade & yet the statistics remain the same. I think they tried something like 'runs saved' stat in a tri series in india (in 1996/97) which featured India-Australia-South Africa. It was interesting to see how frequently the commentators use the term "He saved 15-20 runs with his fielding" when infact the guy saves only about 4-5 runs in an innings. One problematic thing when counting ' runs saved' was you never get to know how many runs a fielder saved just by being there.( Eg- You don't take a single when you see the ball go near Ponting or you don't take two when you see that Andrew Symonds is the fielder) That alone is enough to understand how difficult it is to create perfect statistical analysis. Same goes for a bowler who gets a wicket as a result of the pressure created from the other end/ batsman who gets out after getting frustrated due to his partner.

  • waitara on December 1, 2011, 11:44 GMT

    Excellent stuff! I will look for your book. Averages and numbers are easily measurable things. These other ideas are not ... yet (except Hussey's idea, although who averages 60 in T20?). They are not as tangible, and so many different elements come into it. A batsman who scores a large percentage of his team's runs may in the next game be ordered to go out there and score quickly. If he hits 3 sixes and wins the game but is out with one run needed (and it is scored), he played an extremely valuable innings but did not score many. The intangibles, as in rugby and football, are what selectors and astute observers look for. This is why intellignet reporting is so valuable to true followers (and why cricinfo is such a great service to the game). Try being a NZer and putting up with the lack of this over there. And, off-topic, the NZ team is looking quite good, especially in batting. What a heroe Vettori truly is!

  • liz1558 on December 1, 2011, 11:43 GMT

    In test cricket the best cricketers usually have the best averages, which is why averages are still valid. And they can still be used to compare across generations since WWI, because the game is still broadly the same. The best batsmen generally average over 50 and the best bowlers in the mid to low 20s. The problem is that we've come to use stats to measure greatness,whereas older commentators rated a player and then his stats reflected that assessment. For instance Ken Barrington was never rated as a great batsmen, although Peter May was. However, Ken Barrington averaged 12 runs higher than Peter May. How do you measure greatness? Who knows - it's not entirly scientific. But I would rather have Ambrose over Flintoff, Lara over Thorpe, Viv Richards over Laxman, Warne over Swann, Sobers over Kallis and Steyn over Anderson. That's great players versus very good players, and the stats reflect the quality. But that's just test cricket - ODI and T20 are very different.

  • shivendash on December 1, 2011, 11:29 GMT

    Though average and statistics do a right job of saying who is consistent and all yet the MVP tag can never be said through it. How? Now afridi's numbers are not that highly regarded but any day afridi is a big match player in odi. But look at kallis numbers in odi... average of 45+ with bat but i dont remember a single match sans the 1998 mini world cup where kallis has won a match. Yet he had anchored it but never seen him win a single match. In that respect, a yuvraj or an afridi r better players who take control of matches. In sachin's case, I think he is just UNREACHABLE in ODIs though he is more of a kallis of odi in Test matches where he anchors but hardly win you matches. But the number of runs tells a different story.

  • spiritwithin on December 1, 2011, 11:20 GMT

    @ravichakra,sachin batting average in winning cause is much higher than many of his batting contemporaries in tests and odi's,u gave an example of 2003WC in which u said sachin scored only against minnows,let me give u stats -98 against eng(highest score by either side),36 against aus in league when no other top order indian batsman reached double figure(bowled out for 125)..52 against netherland where india were bowled out for 204(leave his score and india wud have lost that),97 against SL,97 against pakistan,which means he did scored against every team..batsman's job is to score runs,if the bowler does'nt support him or other batsman fails to contribute its not the fault of sachin...cricket is a TEAM game,u cant expect a player to make runs as well as take wickets,one individual player cant play the role of 10other players...cricinfo pls publish

  • ravichakra on December 1, 2011, 11:12 GMT

    This article should be thrown in front of the BCCI whose basis for selection is always average and batting. Do they ever consider who the opposition was? Whether the team was under any pressure when the tons of runs were scored? How easy / difficult the pitch was? What was the quality of the opposition's bowling attack? How many lives did the batter get? The reason I have put down only batting measures is simple - in India no other department is recognised. A person cannot be in the team merely because he can catch or field well. All that is required his how many runs has he scored this season.

  • ravichakra on December 1, 2011, 11:04 GMT

    @ Gupta.Ankur : In your eyes, Sachin might be a great player. That need not be endorsed by the whole world. To me, he is just one among the other 11 who are playing. To you it might seem that his records (runs scored, number of 100s, etc.) are meaningful statistics. To me, they are mere accumulation of personal milestones. A Dravid or a Laxman have contributed far more to Indian cricket or their team than a Sachin, yet got very little recognition from the cricketing fraternity. Sachin has never played for the team cause. In ODIs he cannot bat at any other position other than 1 or 2, in Tests it is always the customary no. 4 and when a few overs are left for the day, a nightwatchman strodes out to bat in his place. Why can't he open the innings in Tests when the team really needs him to? Why is Dravid always made the sacrificial lamb every time? If he is so great why is he so fussed about sight screen heights, bats, pads and all other nitty gritties?

  • ravichakra on December 1, 2011, 10:55 GMT

    @ pranavcrazyguy : Can you put down the list of performance paramaters that you think where Sachin scores heavily? He does not figure in the list of players with the most number of significant innings in Test cricket despite number of 100s being considered as one of the parameters and he has 51 of them.

  • ravichakra on December 1, 2011, 10:51 GMT

    @ nlambda : That is crap to say the least that not a single match was won by Ind where Sachin did not have a strong contribution. Just one example is the match against New Zealand where India was tottering at 21 for 4 chasing 168. Check his contribution in that match. More than half the runs in the 2003 WC scored by Sachin were against minnows - Namibia, Zimbabwe, Kenya, etc. His only noteworthy contribution was 98 against Pak.

  • ravichakra on December 1, 2011, 10:46 GMT

    Finally, an article depicting what a waste of statistical measure an average is. What should be more relevant is the number / percentage of significant innings in a player's career to judge his capabilities. Also, percentage of runs scored in team totals would be another statistic which would be relevant. How many times has a player contributed while chasing down a total in ODI or which position he has batted. For example, opening the batting in ODIs on flat pitches will definitely have more advantageous than coming down the order when the ball becomes older and does not travel / grips and turns. A clear example of how averages can be improved yet the team cause is hardly helped can be drawn from the match against SA by Ind where Sachin scored a 100 and remained no. but the fate of the match did not change or was a foregone conclusion. His 100 was a mere personal milestone. @Sudhakar86 - Ind would have never lost when Sachin played had what you said been true.

  • sonofchennai on December 1, 2011, 10:45 GMT

    Readers have become so cynical that they cant bring in Sachin in any of the cricketing articles....which actually epitomizes his greatness...no arguments..rest my case

  • BillyCC on December 1, 2011, 10:38 GMT

    The good thing in all of this is that the averages don't distort anything at the top end. If you look at the highest batting averages, the greats stack up at the extremes. Bradman rightly remains number one, with the likes of Pollock, Sobers, Headley, Hammond etc. all up there. And of the current players, the greats have the highest averages: Tendulkar, Dravid, Kallis, Ponting etc.

  • 2929paul on December 1, 2011, 10:26 GMT

    There are lies, damned lies and statistics. You can use statistics to prove whatever you want. The trick of statistics is seeing the woods for the trees and Cowan has got it spot on. I think the England fielding coach, Richard Halsall, has been monitoring every piece of fielding done in and England game since he arrived to assess and improve every England player. It was a huge part of their T20 win and helps determine where each player fields. As a coach of junior teams, I have never relied on averages as a stat on which to select my teams. Parents often cannot accept it but performance in certain circumstances within a game mean more and based on what you are trying to achieve. You cannot compare an opening batsman's stats directly with the number 5's withour putting them into context. Your quick opening bowler's first spell is different to his come back spell(s). Don't combine the figures for one set of stats. Stats aren't everything but they do help.

  • stormy16 on December 1, 2011, 9:27 GMT

    Thought provoking article and the broadcasters should cosider this - they are the one's that reach the millions and have the freedom to introduce such measures. For example in the most recent tests what is the more valuable innings - Bravos 166 V Johnsons 40. If its pure numbers its a no brainer but the relevant measure is in everything else but the number of runs. May be a bowler gets a 5fer but the oppostion has made 400+ against a bowler who got a 3fer but knocked over the top three early allowing the others to knock over the opposition for 200. Another would be is Steyns wickets at Newlans the same value as Tahir's - when the conditions are stacked in Steyns favor. The fear here of course is a mathamatical or illogical measurement that only a scientist can understand!!

  • on December 1, 2011, 9:24 GMT

    While i agree to an extent with this article, statistics can be a very useful tool particularly over a career of how good a player is, however it should be relative to others in the time that the players represented their respective teams. Stats are important, but not everything, and certainly are only useful to a limited extent if one considers a players fielding ability, leadership skills and influence. But nevertheless stats are useful and do help to provide insight into a players practical contribution to the team.

  • christy29 on December 1, 2011, 9:12 GMT

    mark ramprakash averages 26 but was a valued member of the test english side for many years.

  • Haleos on December 1, 2011, 8:58 GMT

    I have been syaing this for the long time. People coming up with all sorts of stats for players across generations. Its a plain waste of time. Well done Ed.

  • rpoduval2000 on December 1, 2011, 8:52 GMT

    Similarly I think in Test Cricket we should include how many minutes batsman occupied crease per innings as a statistic to gauge a batsman ....many a time low scoring long ininings played save test matches ..this should be definitely a indicator.....

  • on December 1, 2011, 8:29 GMT

    'Cricket has changed so dramatically, it is mere superficial pub talk'.

    Rather arrogant I feel, one thing that hasn't changed is bowling fast. The great quicks of the past forty years would still prosper today, leaving some of the modern batsmen's egos and bloated averages shredded.

    More credence should be given to winning when we look at stats but overall a list of the greatest players statistically wouldn't be greatly different to one chosen by any other sane method.

  • Gizza on December 1, 2011, 8:14 GMT

    As thought provoking and interesting as this article is, I think EJ36 is onto something. Ed Cowan is statistically quite a weak batsman. True for heavily interdependent team sports like soccer and basketball where there is so much passing and the position of your team mates is so important, more just than the final stats are important. But in cricket apart from backing up in fielding and running between the wickets every action in single whether bowling, batting or fielding. Stats reward a batsman for hitting a four instead of getting out or blocking it (average goes up, total runs and strike rate goes up). There are factors which are ignored like the pitch, how good the ball was (yorker, bouncer, good length whatever) but a more complicated baseball-like analysis can't remedy these holes. The only area which could be improved is fielding stats. Anyway, ESPNCricinfo regularly has articles and blogs on deeper stats. We don't need to see these while watching the game on the telly.

  • Gupta.Ankur on December 1, 2011, 7:34 GMT

    We really don't want to waste our time on making cricket watching a mathematical workshop?

    And people lets not debate sachin's greatness here.......we will be making fool of ourselves.......

  • on December 1, 2011, 6:59 GMT

    You can call Sachin a great of Indian cricket who averages 56 or whatever. How many times has VVS or Viru won Test matches for India compared to Tendulkar. Just as many if not more with far less opportunities. A good average does not win you a game but a good temprament and hunger to be successful. Judging someone by a number can only go so far. Hamish Marshall had never made a First Class 100 before he made one in a Test match against Australia. Paul Collingwood, Marcus Trescothik and Michael Vaughan had sub 40 FC averages however went on to average 40+ in test cricket. These are just a few examples of players with temprament to play for their countries without neccasarily figures to backup their undoubted talent.

  • Sudhakar86 on December 1, 2011, 6:50 GMT

    @Erica Herewini Few Indian players have won pressure games in few numbers but you have to note one thing. INDIA never feels under pressure when SACHIN is on crease. That is his greatness. 175 Vs Australia is the prime example. At 37 he could deliver yet more. I still believe that he is having a couple of 150 + in ODIs for sure

  • on December 1, 2011, 6:46 GMT

    What an excellent article. Imagine having stats for players like 'Pressure Sixers', 'Swing Percentage', 'Direct hits', 'Death Wickets', 'Average Partnership' etc. It will make the game so much more interesting with newer stats. Strike rate came in the 90s. A few more unspectacular things came in the past several years, but now we need modern, fast paced stats that will change the game - 'Runs stopped' for fielders would make fielders more inclined to field well. An average of 60 in tests, like that of Tendulkar or Ponting, includes many years of good form and many years of bad form. Form too can be shown statistically, and not for just long careers, but even for those with a few years of playing. I hope we see newer, better things come in the near future.

  • Agnihothra on December 1, 2011, 6:29 GMT

    To paraphrase Navjot Sidhu.. Stats are like mini skirts. What they reveal is tantalizing and what they conceal is vital.... So they prove that Daniel Christian is the most valuable player because he does all three things reasonably well... but is master of neither batting ,bowling nor fielding.. so I go again.. Lies, damned lies and Statistics....

  • pranavcrazyguy on December 1, 2011, 6:28 GMT

    Yep, I thought someone would say "Sachin" immediately.

    Ms. Erica Herewini, you will find that Sachin scores heavily on many performance parameters, in fact heavier than the other players you mentioned. Think ODI finals for example. Care to find out about Sachin's record in ODI finals vs that of the other players you mentioned? Your beliefs will go for a big toss. There are many more examples. For example, the last, deciding test of an away series.

  • SamRoy on December 1, 2011, 6:17 GMT

    Fine article. Best example, Thilan Samaraweera's average vs Aravinda De Silva's average.

  • on December 1, 2011, 6:12 GMT

    surprised not to see Ed Cowan at the Gabba against NZ - someone ran the algorithm wrong

  • nlambda on December 1, 2011, 6:09 GMT

    @Erica: hardly any match in the 1996 and 2003 world cups was won w/o Sachin making a strong contribution. It is the 'recency effect' of 2011 that Yuvraj and Dhoni are looking like bigger match winners. Until the England ODI series Dhoni's only 2011 matchwinning performance was in the WC final; he was a flop all through the league matches, QF and SF in the WC...

  • ListenToMe on December 1, 2011, 6:06 GMT

    In tests I still believe Average should be the most important factor for selecting a player into the national side. But for other formats, his strike rate, all round skills etc can also be considered. This just on the team's perspective. A player with better average means he is more consistent. That simply is a measure of his talent.

  • EJ36 on December 1, 2011, 5:49 GMT

    Suspect Ed wouldn't have written this article if his FC average was 48, rather than 38.

  • ashok16 on December 1, 2011, 5:48 GMT

    The basic moneyball concept is this: every win can be divided up as a contribution by players who are playing above the mean performance for their position/skill. (eg. player x who scored highest runs in the win may get 0.3 points while the highest wkt taker or run saving fielder will get 0.2 points until the total contribution of all the players who played above average adds up to 1.0). This was how the baseball statistician Bill James came up with the idea of figuring out who the most valuable players are. To win is the aim of every game and only contributions made towards a win matter in this analysis. It will not be too difficult for somebody to come up with an analysis for an twenty 20 league like, say last year's IPL or BBL. It might take a while to do the number crunching and will also be tricky to evaluate the fielding.

  • jonesy2 on December 1, 2011, 5:38 GMT

    brilliant again ed. the dan christian thing doesnt surprise me at all. i would imagine dave warner's fielding would put him above most players as well. superficial stats are so ridiculously over-consulted to measure performance in all sports.

  • ashish514 on December 1, 2011, 5:35 GMT

    I remember one of these recommendations being tried earlier. I don't know the exact year or tournament, but it was somewhere when Ajay Jadeja was around, and most probably in a world cup or some other one day tournament/series involving India. Whenever a fielder was shown on screen, statistics about his fielding were shown at the bottom. I am sure that no. of runs saved was one of the stats, not sure about no. of runs conceded.

  • Hareendra on December 1, 2011, 5:25 GMT

    Amazing point being raised! This is the kind of innovative thinking that is required and what I would like to see as a reader. A very challenging piece of writing.

  • dnarmstrong on December 1, 2011, 5:22 GMT

    This bloke has a future in journalism after his cricket career - love your thoughtful, well researched writing Ed. Only problem is that I didn't see any mentions of Sachin.

  • on December 1, 2011, 5:13 GMT

    @Erica Herewini - please pass on your wisdom to the Indian selectors and ask them to drop Sachin and take Yuvraj to the Aussie tour. Hell - if you were the selector, would you do that, unless you dont want to put your money where your mouth is? I dont understand why there is so much animosity towards one player?

  • CricFan24 on December 1, 2011, 5:12 GMT

    We simply have to accept that Stats are guidelines and should be treated as such . A coarse grain model has as many flaws as a fine grain model. There is no end to model refinement. And even the refinements will have flaws...and so it goes. A Lara will certainly have a greater percentage of team score than a Richards- simply by virtue of being in a weaker team. Does that make Lara the better batsman ? Lara had his best run in the last 4 years of his career - at a time when the West Indies were at their worst in his career span...There is simply no end to statistical fine tuning. You can bring out various facets of a player , but finally it is only knowledge and judgement about the game which can ascertain "true value" - a subjective value but one which most knowledgeable observers of the game will usually have a more or less unanimous consensus of.

  • Rosey86 on December 1, 2011, 5:04 GMT

    Agreed, as someone who captains a side within a club across many grades, the players who stand out to me are the players who work hard at all aspects of the game. I open the bowling and i feel it is my duty to set the tone. The pressure i build at one end over a long spell, allows my more explosive bowlers to bowl short sharp spells at the other end. The fielders put in effort all day. While stats are nice, its the effort that you cannot calculate which can tell you of a players worth. When asked by others who should be promoted a grade, i don't only think about their main attribute (batting, bowling, keeping) but how they go about their game, what they bring to a side. The hardest working cricketers are not always the most talented. Stats are a wonderful thing, but its how an individual contributes to a team effort that really matters. I hope Ed, that you get the chance to prove yourself at the highest level one day, as with Dan Christian

  • on December 1, 2011, 5:02 GMT

    time for efficiency stats............like NBA

  • TheOnlyEmperor on December 1, 2011, 5:01 GMT

    Like I have always said, Bradman is no big deal!

  • on December 1, 2011, 4:48 GMT

    Being an avid follower of baseball, I have always thought cricket needs more advanced statistics to give a more accurate description of a player. The WAR statistic (wins above replacement) take many complicated statistics and puts them together (various aspects of defense, offense, and pitching), really leaves cricket far behind.

  • on December 1, 2011, 4:47 GMT

    'Someone of the ilk of Chaminda Vaas is cricket's version of Battier - content to fulfil his role as a cog in the wheel, with long, dry spells - in doing so enabling those around him to perform in dynamic bursts.'

    Precisely.

    I agree wholeheartedly that the stats need a re-look, more so in fielding. The likes of Dravid, Jayawardene are fantastic Slip catchers, but would I rate them highly as well-rounded fielders, particularly when compared, say, to a Mark Waugh (who is probably the greatest catcher of the cricket ball I have ever seen)? I don't think so. The stats need to reflect the contribution fielders of the calibre of Symmonds, AB de Villiers, et al who can field pretty much anywhere on the field. I don't even need to bring a certain Jonty Rhodes into the debate, do I?

    And then there are the run outs. I would love to know what the stats show with regard to the run outs Ricky Ponting has created, most of the time on his own.

  • BillyCC on December 1, 2011, 4:46 GMT

    An interesting piece by Cowan; himself with a maths background at Sydney Uni I think. If we judge cricketers based on this algorithm, at first glance; guys like Bradman, Sobers, Gilchrist etc. retain their place at the top. What you may find is guys like Kallis, Shaun Pollock, Richard Hadlee, Sehwag, Steve Waugh, Allan Border and Laxman overtake the likes of Ponting, Tendulkar, Sangakkara when not keeping, Marshall, Lillee. This certainly won't be true in all the cases I've mentioned, but there will be anomalies in such an analysis.

  • McGorium on December 1, 2011, 4:12 GMT

    It is impossible to reduce human endeavour to a set of metrics, for the purpose of transparency or objectivity. Regardless of how many metrics you devise, if it is public, the people being measured will adapt such they maximize the metric. Ultimately, statistics prove nothing. Correlation does not automatically imply causation, and it requires the careful skill of a selector or manager to pick the right player, motivate and groom them. And for that purpose, averages and strike rates do the job satisfactorily. You can have averages and SRs under various conditions, and even std. dev. to capture consistency, and that's more than sufficient as a filter to weed out non-performers. Pseudo-sciences like management, economics, social sciences, etc. use stats, and look at their success-rate. Modeling is useful, but suggestive. As we saw with Wall Street, just because the quants say something doesn't mean it's always true (e.g.: look up LTCM). Lets not absolve selectors of their responsibility.

  • on December 1, 2011, 4:11 GMT

    As a statistician I have been thinking along these lines since I first read Moneyball, but I don't know where I would source the data to start a similar exercise for cricket. Billy Beane in some instances just used existing stats which were underrated, such as on base percentage. It sounds like Hussey is on the same track, combining a couple of stats to come up with a stronger one. Baseball has the concept of errors to assess fielding, so a similar thing in cricket would actually count the dropped catches more than those caught - you could also add where a no ball leads to a missed wicket.

  • on December 1, 2011, 3:51 GMT

    Yep stats wont make Sachin great. Dhoni, Sehwag, Yuvraj have won a lot more matches with lesser average ESPECIALLY PRESSURE GAMES!

  • Raju_Iyer on December 1, 2011, 3:47 GMT

    Its an old saying that Statistics are like a bikini - what they reveal is interesting but what they hide is vita. That said, I don't think performance statistics can be 100% co-related with IPL pricing auction, because other factors like a player's availability, ability to draw crowds, perform in a team setting etc. all come into play. To me, simple statistics are good enough. e.g. any batsman who has an average approaching 50 after more than 100 appearances in Tests has to be seriously good. Ditto for a bowler with over 250 scalps at around 25 runs per wicket. Similarly Strike rates and Economy raes to supplement average are good enough for ODIs. It will be sad if cricket has its own Moneyball tale to tell.

  • michaelcoffs on December 1, 2011, 3:24 GMT

    That's a fantastically thoughtful piece. And .. from a player! Please keep writing Ed. Also, searching for an algorithm that addresses the intangibles - eg "potential" (how do you measure it?) / resistance to media hype or criticism / response to sustained opposition probing and sledging etc etc. All part of cricket's rich tapestry.

  • No featured comments at the moment.

  • michaelcoffs on December 1, 2011, 3:24 GMT

    That's a fantastically thoughtful piece. And .. from a player! Please keep writing Ed. Also, searching for an algorithm that addresses the intangibles - eg "potential" (how do you measure it?) / resistance to media hype or criticism / response to sustained opposition probing and sledging etc etc. All part of cricket's rich tapestry.

  • Raju_Iyer on December 1, 2011, 3:47 GMT

    Its an old saying that Statistics are like a bikini - what they reveal is interesting but what they hide is vita. That said, I don't think performance statistics can be 100% co-related with IPL pricing auction, because other factors like a player's availability, ability to draw crowds, perform in a team setting etc. all come into play. To me, simple statistics are good enough. e.g. any batsman who has an average approaching 50 after more than 100 appearances in Tests has to be seriously good. Ditto for a bowler with over 250 scalps at around 25 runs per wicket. Similarly Strike rates and Economy raes to supplement average are good enough for ODIs. It will be sad if cricket has its own Moneyball tale to tell.

  • on December 1, 2011, 3:51 GMT

    Yep stats wont make Sachin great. Dhoni, Sehwag, Yuvraj have won a lot more matches with lesser average ESPECIALLY PRESSURE GAMES!

  • on December 1, 2011, 4:11 GMT

    As a statistician I have been thinking along these lines since I first read Moneyball, but I don't know where I would source the data to start a similar exercise for cricket. Billy Beane in some instances just used existing stats which were underrated, such as on base percentage. It sounds like Hussey is on the same track, combining a couple of stats to come up with a stronger one. Baseball has the concept of errors to assess fielding, so a similar thing in cricket would actually count the dropped catches more than those caught - you could also add where a no ball leads to a missed wicket.

  • McGorium on December 1, 2011, 4:12 GMT

    It is impossible to reduce human endeavour to a set of metrics, for the purpose of transparency or objectivity. Regardless of how many metrics you devise, if it is public, the people being measured will adapt such they maximize the metric. Ultimately, statistics prove nothing. Correlation does not automatically imply causation, and it requires the careful skill of a selector or manager to pick the right player, motivate and groom them. And for that purpose, averages and strike rates do the job satisfactorily. You can have averages and SRs under various conditions, and even std. dev. to capture consistency, and that's more than sufficient as a filter to weed out non-performers. Pseudo-sciences like management, economics, social sciences, etc. use stats, and look at their success-rate. Modeling is useful, but suggestive. As we saw with Wall Street, just because the quants say something doesn't mean it's always true (e.g.: look up LTCM). Lets not absolve selectors of their responsibility.

  • BillyCC on December 1, 2011, 4:46 GMT

    An interesting piece by Cowan; himself with a maths background at Sydney Uni I think. If we judge cricketers based on this algorithm, at first glance; guys like Bradman, Sobers, Gilchrist etc. retain their place at the top. What you may find is guys like Kallis, Shaun Pollock, Richard Hadlee, Sehwag, Steve Waugh, Allan Border and Laxman overtake the likes of Ponting, Tendulkar, Sangakkara when not keeping, Marshall, Lillee. This certainly won't be true in all the cases I've mentioned, but there will be anomalies in such an analysis.

  • on December 1, 2011, 4:47 GMT

    'Someone of the ilk of Chaminda Vaas is cricket's version of Battier - content to fulfil his role as a cog in the wheel, with long, dry spells - in doing so enabling those around him to perform in dynamic bursts.'

    Precisely.

    I agree wholeheartedly that the stats need a re-look, more so in fielding. The likes of Dravid, Jayawardene are fantastic Slip catchers, but would I rate them highly as well-rounded fielders, particularly when compared, say, to a Mark Waugh (who is probably the greatest catcher of the cricket ball I have ever seen)? I don't think so. The stats need to reflect the contribution fielders of the calibre of Symmonds, AB de Villiers, et al who can field pretty much anywhere on the field. I don't even need to bring a certain Jonty Rhodes into the debate, do I?

    And then there are the run outs. I would love to know what the stats show with regard to the run outs Ricky Ponting has created, most of the time on his own.

  • on December 1, 2011, 4:48 GMT

    Being an avid follower of baseball, I have always thought cricket needs more advanced statistics to give a more accurate description of a player. The WAR statistic (wins above replacement) take many complicated statistics and puts them together (various aspects of defense, offense, and pitching), really leaves cricket far behind.

  • TheOnlyEmperor on December 1, 2011, 5:01 GMT

    Like I have always said, Bradman is no big deal!

  • on December 1, 2011, 5:02 GMT

    time for efficiency stats............like NBA