I feel it is ridiculous when people argue otherwise, you must always allow for the context that the runs where scored, or wickets taken because it is an important variable. Not everyone gets to bat or bowl for extended periods on true wickets or flat pitches or snakepits or dustbowls. Unfortunately it is difficult to average these variables. It is not wrong to judge by runs/innings out, it is ridiculous to get religiously fanatical and believing greatness from these incomplete figures.
We have to allow for empirically undetermined facts.

It was no less numerically-advantaged a man than Walter R. Hammond who, in the opening paragraph of his delicious /Cricket My World/, declared that "figures can't lie but liars (as we know to our cost!) can figure."

You are right. Statistics are not everything. Your ranking on the Test Championship does as it effectively measures you performances over a period of time. Say what you want about Anderson, but England are still number 5 on that table. That says enough about their players performances.

Well done!!!
Its the same with almost all international players except a few Australians.
Really, someone like a Sehwag or a Jaffer score a ton per test series and scores a huge one[mostly on the worst pitch] and comes out with pretty impressive series average when he has contributed almost nothing to mhis team's success[a glaring difference being his performance at Galle]

Interesting piece Mark....your quote
'That is why I am always deeply suspicious
of contributions to cricket debates which
effectively say “X’s career average was 35
and Y’s was 40 and that proves it”
Also your comment about comparing across
eras is true. That is why the Lara V Tendulkar
is always so tantalising and controversal.
Both played most of their careers at the
same time i.e. played against the same teams,
same pitches and same frequency. A common
observer of the facts and stats as they stand
will see that Lara has 77 more test runs than
Tendulkar but played 19 less test matches
and more importantly 12 less innings.
Doesnt this suggest that Lara was more
successful batsman (i.e. runs per match &
runs per inning)???But the same observer will
also find Tendulkar's average being 1.35
better than Lara's. You will notice the not out's.
Tendulkar 25, Lara 6. Dont these facts suggest
Tendukar was the more consistent player???? Tendulkar
also has 5 more tons than Lara but this is weighed up
against the 12 more innings he has played.

Individual player stats will never show some important
information. The teams the players played in.
Tendulkar has batted in a team with three other players
averaging over 50. Lara was the only player with a 50
average in the team for many years.

Does that imply that Lara had more or less pressure than
Tendulkar? Indian cricket and west indian cricket come
with completely different pressures.

Both players hold records in test cricket. Tendulkar's
record of most tons is testament to his longevity as
much as his skill. He will soon take the most runs
record also, another record based on longevity.

The most runs in test cricket record apart, Lara's
records are based on one-off feats of brilliance.
i.e. 400*, 375, most runs in an over, 100 before lunch.

I await Tendulkar's retirement when we can finally
compare, judge, crit, and admire both players and
perhaps come to a decison on who was the greatest.

You are both right and wrong.
Statistics can have tremendous value in analysing sport. In baseball, modern usage of statistics has increased the standard and the understanding of the game to unparalleled heights; in football and similar sports, the poisson and skellam distributions model the game extremely
accurately.
However, to look further at baseball's example, the statistics revolution necessitated the analysis of old statistical measures (such as the batting average), and where found lacking, the removal thereof. Baseball's batting average was found to have little correlation with how valuable a player actually was, so better statistics were adopted.
So here we see the two sides of the argument. The arguing against statistics by many cricket fans is overwrought: the only statistics against which they should be raging are those which do not fulfill their purpose. Ironically, it is statistical analysis, not rhetoric, which will explain which these are.

Batting averages, by definition,smooth out the troughs and peaks, and should be viewed only as an indicator of someones performance over time. If you combine it with his Standard Deviation,( a measure strangely not used by a statistics obsessed cricket fan)which tells you how consistently close he was to his overall average,(ie, how consistent he was) you have a better indicator of his overall quality.But these are only indicators:most would consider Sehwag one of the current greats, with an average over 50.But the fact is that though he got a triple- and a double-century in his last two series, he only once made 50 in the remaining innings.So he has a great average, but a high Standard deviation pointing out his unreliability.

Statistics or no statistics, i have to agree with Mike's comments on Ian Bell. Bell appears to have this goldenboy image in certain sections of the media, with many seeming to believe that his place in the team should be sacrosanct due to his wonderful technique. Well, Smith's technique is not what one would describe as orthodox, but how many runs has the SA captain scored in comparison to Bell? And more to the point, when has Ian Bell knuckled down and produced an innings when his team really needed it? 199 on a road at Lords? So what. He couldn't even finish the job off yesterday when England were almost certain of victory and batting on another decent wicket. The simple fact of the matter is that Ian Bell doesn't appear to have the mental strength at the moment. Bring in Rob Key, or even Ramprakash. Key especially deserves another run in the England team, especially after this season.

Mark Twain was right on the mark when he (allegedly) said "There are three type of lies: lies, damned lies and statistics."

Almost no matter what you want to 'prove' you can find statistics to back up your case. It's not hard to do and (partly for that reason) it's one of the most common ways that stats is mis-used.

Specifically in the context of cricket, I believe statistics should only be used in a relatively minor supporting role in the discussion of how good/bad/indifferent a player/team is/was. They can only ever be a rough guide, suggestive of what might lie beneath, and should always carry along with them a list of assumptions that were made in compiling them. To do anything that stands on statistically sound theory would take far more time to understand than most people have. That's why there are professional statisticians in this world!!

I agree with R.Nayan in using "Standard Deviation" more often as a cricket statistical tool. Ed Smith in his book "What sport tells us about life:Bradmans average..." uses it to great effect.
There could also be a column intorduced about the quality of the wicket when the score was produced. In some kind of index factor, therfore a score on a difficult wicket would have some kind of factor of 2. Or is there a more sophisticated statistical tool for this?

Stats, indeed, rarely tell the full story. There were a load of people here in Colombo who always told me, a few years ago, that "Murali is better than Warne because he took his wickets in nearly 10-15 fewer tests". However, this statistic is a almighty lie to Warne's capabilities because, even though Warne played more test, he bowled less overs per match because he had great bowling partners, unlike poor Murali, who never had any bowling partners (till Mendis' arrival), and so he ploughed through an average of 70 overs per match!

Another crazy one was a comparison of a superb opening batsman, Krish Srikkanth and of a middle order 'wall', Rahul Dravid. One person said that, since Dravid and Srikkanth had the same strike rate, they were equally good. What a right royal joke! Srikkanth played in the times when a strike rate of 70 is equal to today's 100!

Following on from my earlier post when i compared how stats do not tell the full story regarding Lara and Tendulkar i have found another comparison that stats suggest are very similar but we all know it's not quite that simple.
One current batsman, another from 70/80's. Both have played 125 test matches, although the current batsman has played 2 more innings hence 124 more runs to date. One has a top score 270 the other 236* Very close huh? The modern batsmen has close to a three run better average due to 10 more not outs than the batsman from the past. At this point we may be swinging towrds the modern batsman as the superior, until we find the 70/80's batsman has 34 tons compared to 25! What stats dont tell us is that Gavaskar opened the batting for Indian for many years and faced the most hostile fast bowling ever while Dravid batted at three or four and was surrounded by other batsman with 50 averages.

Do not live by stats alone!

Well, statistics don't lie and as someone has pointed out earlier averages are smoothed curves over the years. But we still need statistics to prove our points because the ultimate comparisons are the runs and wickets for decision making in cricket which are expressed in absolute nos. I guess the authors grudge is against simple measures like averages and strike rates not against statistics. Statistics are merely observations whereas inferences are our own based on our understanding of observed scenario. So I guess we shall not call statistics as lies but our understanding of statistics as poor which needs further reading to make the right use of it. Cricket is game based on nos. and we need statistics to analyze that. Ironically author himself used statistics to make his point home

I agree with R. Narayan's suggestion regarding the use of standard deviation to help judge the consistency of a player. I would love to see some stats regarding who the most 'consistent' player was using this kind of analysis.
On the other hand, I also don't think this would help determine who a better player was in any sense. Can we really say that a batsman who scores 45, 55, 49, and 51 in four innings is a better batsman than someone who scores 180, 0, 0, 20? Although they have the same average, who is more likely to have won a game for their team? Doesn't that depend on the situation? Also, doesn't every team need both kind of players?

You are right! Mark Twain bellowed, "There are lies, damned lies and then statistics!".

I agree totally with that viewpoint since 99% of all statistics is totally worthless anyway!

nice to see someone with a perspective!

Eddy, i see you've compared Gavaskar with Dravid. I think it would be fairer to compare Gavaskar with Tendulkar, And I still think Gavaskar would be adjudged the greater. He faced the greatest bowling attack of all time and THRIVED on it!!
As for Averages being indicative of worth, I think Runs Per Inning (RPI) would be better to gauge what to expect from a player walking out to bat. If a player has an average above 54 but scores 48 per inning, you should not expect 50 runs from him! Not Out's do not give you extra runs!

@ Eddy . It's ironical that you say dont live by stats alone yet in illustrating that Lara is a better cricketer than Tendulkar you point out, that although Lara has played less matches than Tendulkar he has more runs.(Unless I'm mistaken thats a stat). Now let me just say this. A man ,Mr. A, scoring 40,42,43,45,42,40,49 and 39 in eight innnings to me has played better cricket than Mr. B who has scored 10, 12, 21 ,2,19 ,400 in six innings. Yes,Mr.B has played fewer innings and has more runs yet in my book he is not the better cricketer. Think about that Eddy.
In addition,(just to indulge Mr. Eddy ), when asked recently who was the better cricketer Shane Warne arguably the greatest spinner of all time stated that it was Tendulkar ; Granted ,you may , in turn, contend that Murali stated Lara was the best batsman he has bowled too , however I consider Warne to have the better cricketing brain (His leadership during the IPL supports this) and so his views to me are more substantial.

I would agree that the statistics can't be used in isolation: the context needs to be taken into account. On Murali v Warne, I would observe that Murali took a significant proportion of his wickets against Zimbabwe and Bangladesh, Warne played at most one or two matches against these countries.
On Lara v Tendulkar, it is very difficult to use statistics to separate the two. Certainly Lara had higher peaks in terms of 400*, 375 and 277, but Tendulkar may have been slightly more consistent. While the second half of Lara's career was spent in a struggling side, it is also important to remember that until Dravid and Ganguly emerged in 1996, Tendulkar spent the first 7 years of his career (when India didn't play much Test cricket, which means that he may have lost out in terms of runs and performances) in a struggling side as well.
My take on batting averages above 50, which are quite common now, is that it is important to remember that bowling attacks are a lot weaker these days.

Yes, statistics can be useful but its how you get them and when that matters!! It may be cruel but Tendulkar's mass of runs in ODIs doesn't mean much without winning the world cup. He had one chance and flopped, the other batsman who had their chance and scored 100s when it mattered on the big stage were Lloyd, Richards, Aravinda de Silva, Ponting and Gilchrist. Only Richards averages higher than Tendulkar but the others all scored when it mattered. Averages don't mean anything when it comes to lifting the trophy.

It is extremely funny that people still talks about tendulkar's consistency. For me Lara was above tendulkar even in terms of consistency. 2, 8, 1, 8, 2, 5, 55,3, 20, 32,248 * these are the scores of Sachin tendulkar during 2004 where his 248 against Bangladesh managed his average for the whole year. Lara may have 400 against harmission, flintoff, simon jones, Hoggard, but even sachin have not out innings like 248, 201 against minnows.

No matter how big inning lara's 400 is it still counts for just one not out. If lara had 21 not out like Sachin tendulkar his average would have been around 57 but if sachin had 6 not out like Lara his average would have been not so impressive 49.
A not out may look good on batsman , but its more runs that batsman scores is better for the team.

Statistics is a tool, and not a result. A wise man would use statistics to understand different aspects of a player. Greatness cannot be measured... not by numbers or any other thing, it has to be experienced. e.g. an openers 60 odd runs in 50 balls against superfast bowlers to set the tone of the innings can be better than a middle order batsman's run a ball (not out) century built on top of that, even though the middle order batsman's average against the same opponent and on the same pitch and in the same era is far better than the opener. Statistics don't lie, but they don't tell the entire truth... And entire truth is well beyond numbers. Time people stopped using numbers of any sort as the only tool to measure greatness or lack of it.

@ 'Steven' im guessing youre the same guy who wrote in another blog last week 'I would prefer to have an always reliable Chanderpaul on my team than an inconsistent great player,like say a Lara' ? Is there no end to your Lara bashing?? If Lara was so inconsistent why does he average 52.88? You cant use the not outs as he only had 6 (6,48,13,153,80,400) only half were scores over 50!
In fact if these scores were not notouts (as in runs per inning) he'd still average 51.52. Thats higher than Tendulkar, Ponting, Dravid, Kallis, Sehwag...etc
Your arguement is that Lara scored big hundreds followed by small scores. Well he also scored 48 fifties (one less than Tendulkar and the 6th highest ever). i just don't get your argument Steven. If i could choose between a run of scores 110,10,20,10,100 OR 45,50,40,40,35,40 (both the same total) theres only one winner.

Thanks for all the comments.

I think statistics are very useful, but they only measure what they measure. In order for them to be useful to an argument, you have to show why the number is of relevance and, if using them to compare players, demonstrate that the basis of comparison is fair.

In my view, the career average is hardly ever relevant to any interesting question, and even if it might be, it is often not a fair basis on which to compare two players with markedly different career patterns.

And, as Tushar so eloquently says, numbers can rarely capture the true value of a performance in the context of a particular game.

Going by stats alone, then Vinod Kambli (remember him any one?) must be the greatest Indian batsman. Enjoy what you see and who makes your heart flutter at the moment. Do not look back to see how they stack up in one's career. Afterall, a 300+ on the flattest of pitches, that does not earn a win for the team shouldn't count much. Ask Lara (his dragged out 400, just for the world record sake), Gayle (his 300), Sehwag (his 300 against SA recently)...drats these stats. Means nothing in my book. Enjoyed the hammering Laxman/Tendulkar (that too playing only on the on side)/Dravid/Ganguly handed out to Waugh's team (and NHussain's england and M.Vaughn's england). Now that will be in my memory for ever. Not what their career stats are. Who cares what they scored on flat pitches and ofcourse against B'desh and Zimbabwes (and more recently WI) of this world?

I believe statistics are useful but you cannot use them at high level. To have valid comparisons, you have to remove certain data points that dont fit the analysis.
Completely agree with Graeme's post that Tendulkar's mass of runs don't mean much without winning the World Cup. If you remove Tendulkar's ODI centuries made on Asia subcontinent and against mediocre attack, you realize that his record is actually quite poor in comparison to Lloyd, Richards, Aravinda de Silva, Ponting and Gilchrist. This explains why he has not been able to win the World Cup.

Bla bla bla. thanks. last time i read this blog. i gave it three tries. the first time, kallis isn't a great instead of just giving him credit for being a better batsman than england has had for years and a better bowler than a lot of the crap they have had in their attack the last 15 years. the following week, graeme smith's 154 not out is basically useless because he isn't regarded by you as a fluent batsman. nevermind if your hero chanderpaul isn't that fluent either. and now, you're backing up your arguments with a lot of hot air of statistics. statistics are a true measure of what you're looking for. if you want to know who has the best career average, statistics will give that to you. doesn't mean that the one player helped his side out of more holes, no. but it does show you that the one performed better in his career against the teams he played. butcher's career average shows that he was simply a below average player.

Nice column and observation, but out of that piece only one thing is proven that none of the English men apart from KP has a world class record (ie batting average of 50 + over the carrier or bowling average of less then 25) over the last 15 to 20 years and by all accounts they are and have been an average side all along and have won nothing but an ASHES series at their home GROUND . So it proves the point if they would have players (atleat 3 to 4 ) with 50 plus averages or bowling averages of less then 25 or 25 plus centuries or 400 plus wickets,and not have results accordingly then YOU COULD HAVE A CASE but ............... STATS DONT LIE check out Aussies RECORDS Both individual and TEAM as well

@ Veritas......i agree with most of your comments,
i was only comparing Drav and Sunny because they
have played the same number of games and very similar
innings, i definatley wouldnt compare them as batsmen.
@ Steven...lets not go down the 'he said he was better
route' i.e. warne saying tendulkar was better or
Murli/ McGrath saying Lara was better, really its just
opinions that could go either way. But what i would
say Steven is that your observation that a batsman
'Mr. A, scoring 40,42,43,45,42,40,49 and 39 in eight
innnings to me has played better cricket than Mr. B
who has scored 10, 12, 21 ,2,19 ,400 in six innings.
Yes,Mr.B has played fewer innings and has more runs
yet in my book he is not the better cricketer' Doesnt
hold water. Scoring mid-forties again and again will
not help you win a match or be considered great just
as scoring 10's or 20's will have the same effect.
But a few 10's and 20's followed by a 100 or 200 is
something else.

I'll make a bet with you. I bet most readers of this
blog would rather score 10, 10, 20, 100, 10, 200 (350)
in 6 innings than 49, 40, 46, 37, 55, 32, 46, 45 (350)
in 8 innings.

Big scores often win/help to win matches. 40's and 50's
hardly ever!

We all look for tons scored and scores that win or set-up
wins as judges of players perfomances.
Give me a ton and a double ton in 6 innings (3 tests)
anyday over nice litte knocks that dont do much for the team.

Check this out as a case in point. Ali Cook recently
scored 60,18,60,76,9,39,67 against SA in 3 tests.
England lost the series.

In the same series Ash Prince scored 101,9,149,39,2,4,24.
SA won the series.

think about it steven.

@Eddy- Did Windies win their series against England when Lara scored his 400 not out after a series of low scores . I think not. Why? Cause he was not consistent during that series. Similarly, after Bell scored his 199 and later failed to make any meaningful contribution during the rest of the test series against South Africa;did England win that series? You already answered that.Lets try a different approach .( Maybe you might get what I'm saying). A team with 7 consistent batsmen scoring at least 45 runs every match will more than likely lose less matches than a team who has players who score one big hundred every four innings. Don't get me wrong scoring hundreds is important but CONSISTENCY (at least 40 runs) is more important . If a player is consistent at scoring big hundreds then great. Case in point- West Indies would not lose that many matches if they had at least 3 more Chanderpaul(not a big 100 guy)like batsmen, than a player who scores one big hundred every six inning.

I think an interesting thought is highlighted in your argument about Jimmy andersen and Mark Butcher. You rightly pointed out the strange extremes in which ther careers have progressed. Well, statistics is exactly the same. They can be highly useful and highly useless in any given situation depending on what the context is. But, over a reasonably long career, definitely, it comes out as a useful tool to judge performance. I'll give you a statistic... 99.94.. few of us have ever really seen Don bat, but we accept him as the greatest batsman this planet has seen on the sheer awe of that number. In some ways, Jacques Kallis' achievements should be considered in the same light. A cricketer who has scored 30 centuries and 200-odd wickets surely is up there with the greats of the game. At the same time, there are others, who have made it cuz of their sheer competence to play in pressure situations. I think both these set of players deserve to be remembered as greats.

@ Steven...again. I know where your going with your
points of view but im not sure your really understand
cricket or batting.

You failed to comment on my point about Prince and Cook.

It's quite amazing if you look at the facts.
Prince scored 328 runs in the series.
Cook scored 329 runs in the series.

Prince score a ton in the first test and the second test
and set up the wins. The remainder of he innings he totaled
78 runs from 6 innings!!

Cook was far more consistent scoring 4 50's in 7 innnings
but only his efforts in the last test made any difference.

By your opinion Cook is the more consistent, more runs,
more 50's.

i think i can rest my case.

To say you'd have Chanderpaul
over Lara says it all.

You believe Chanderpaul is a more consistent player
than Lara was...let me see.

Lara scored a 100 every 6.8 innings and a 50
every 4.8 innings.

Chanderpaul has a 100 every 10.2 innings and
a 50 every 3.9 innings.

So the fact suggest that Lara and Chanderpaul
will both score a 50 around 4 or 5 innings but
Lara scores more frequently in the ton stakes.
Hence my point, 100's are more vaulable to
a team than 50's.

ps, your comment about the 400* has been discussed
ad nauseum. It was more about not losing the
match and the series 0-4 than trying to win.
Also i think it was important that the record
was re-taken by a 'great' (Hayden is very good)
player(Like previous players Sobers,Hutton, Bradman)
against a proper team (not ZIM!).

The 400* is often seen by many non-west indian fans
as an innings of selfishness and meaningless.
This record was all the Windies fans could shout
about at the end of a dreadful series and decade.
To a true cricket fan will know how difficult it
is to score 100, let alone a 200 or 300 or 400*

Staistics caqn decieve but they don't lie. About the comparisons like Sachin and Lara. Their stats are too close for someone to just stand head and shoulders among the next one. (Bradman is the only exception). A true comparison can be made when the 2 players play under the same condition. Ex - Tendulkar v/s Dravid or Pointing v/s Hayden. its ironic that the record of all 4 of them are not that apart. They are all modern test greats.
Ifs and buts matter a lot, so a realistic comparison is almost impossible. Example So far Poiting has made only one run in 6 inning he last played in India and shane warnes average was 50+ v/s india in test. So had shane warne or pointing been born in india would they be the great players, who knows but there is question mark. Similarly what if tendulkar was born in south africa etc would he still have that 55 average that made him great. But still statistics do tell a story especially if a carrer has spanned over many years.