Statistics lie

... sometimes. Not just because of Darrell Huff.

It's sometimes palpably true that stats tell a tempting but ultimately false story. Particulary the story of batting and bowling averages, which is what this is mainly about.

Not least in cricket, where stats are sometimes upheld as the only true judge of worth.

And I don't mean just:

a. when a career is so short, that stats are a hardly representative fragment
b. when a career has different sections, or develops (I think of how Gooch and Amiss started out in Test cricket. Ducks and low scores all over the place ...).
c. speaking of home vs away performances, different oppositions, etc.

The second and third of those at least can be allowed for, if the stats are intelligent enough.

No, it's cases where stats would have to be so sensitive, they would start to fall apart.

Cases in point:

1. Such as Ian Bell, or in a different sense Matthew Hayden (or name your FTB) or indeed Murali. How far do stats acknowledge pressure runs/wickets vs luxury runs/wickets?

There are a lot of "Anelkas" in cricket. Majestic when they don't need to be. Flopping regularly when most needed.

2. Danny has always praised David Sales for his contribution for the team, as against his personal glory and average. There are other cases. A batsman who grafts like hell for 5-10 runs in some cases, plays in cavalier fashion and is out for a necessary 25-30 in another, is not necessarily going to do wonders for his average. But he may win more matches.

3. WG Grace. Poor Test batting average of 32, FC av of only 39, weak by today's bracing standards. But despite the famous incident of the replaced bails, the redoubtable, pugnacious doctor was often playing for high entertainment (and stakes) on rubbishy pitches at the time, where consistent high scores were non-existent. Going back 100 years and more may be an extreme case, to make a general point ... , but do stats properly allow for different conditions ..... in different decades?

4. Lance Gibbs. A great spinner. With a Test bowling average of near 30. But: Caribbean conditions in an era when these were not greatly suited to spin success. Picking up from the bombardment of preferred pace, filling in and bowling tireless spells in the sun when others couldn't. And tying the batsmen down for others. (And yet 300+ wickets).

Don't put too much faith in stats ....

Stats tell the truth. They are what they are.

The only trouble is when people think they are something else.

Stats tell the truth of stats. ...

A lot of times stats are a waste of time

No one can convince that S Clark is a better bowler than I Sharma when the former has a far, far better average.

good thread PCS

There are lies, damned lies, and statistics.

Statistics are like Bikinis ....What they reveal is interesting ...what they hide is more interesting

Probably explains why Mudge never really gets a hang of statistics then.

Statistics partly define the luck of a player. Clark and Sidebottom's stats for example may identify them as being quite unlucky (outside issues such as dropped catches, beating the outside edge etc). However, if you actually watched them you'd see that they have the talent and that stats can therefore become misleading and can bring a player down if you've never actually seen them.


This thread is surprising after the passing of Frindall, the said Greatest Statistician. Who would likely disagree with Peter's initial post.

Basic stats will always ever be a rough starting point as they are blunt instrument.

Has anyone here read "moneyball"?

If stats told the full and accurate story, we would have no need of all the endless debates or ranking systems, and so forth.

Stats need interpretation. which introduces an unavoidable degree of subjectivity.

But it's fun...

Once again you scare me JGK. Is that the same book (about baseball or football from memory) some brokers use to play the money market?

The numbers don't lie. Someone's high score is his high score. Someone's average is his average.

It's only a lie if someone says that raw statistics are a reliable comparator of ability, but no one would say that anyway.

Bradman wrote:Once again you scare me JGK. Is that the same book (about baseball or football from memory) some brokers use to play the money market?

It is about baseball and the GM of the Oakland A's approach to picking players in the 90s (when they were regularly in the play offs even though they had the second smallest budget). In particular, it talks about the use of lesser known baseball stats (Sabermetics) to evaluate the true worth of a batsman. For instance, Kevin Youkilis, who is currently at Boston but was an All Star this year, was high on the A's list becuase of his ability to not hit. That is, he would regularly draw walks from exasperated pitchers and to some extent, getting to first base on a walk is as good as getting there with a safe hit with the added bonus that you might have soaked up a stack of an ace pitcher's maximum pitch count.

It is a fascinating book by Michael Lewis who is a brilliant writer.

You might be thinking of his other very famous book - Liar's Poker.

Liar's Poker is sitting on the desk next to me now. I might have to pinch it.

Paul Keating wrote:Liar's Poker is sitting on the desk next to me now. I might have to pinch it.

It's even better when you read it.

Actually it was definately moneyball i was thinking about. I was on some wierd seminar when i was in Alaska and there was an obscure session on identifying 'under appreciated aspects' of the market in a whole of market context and identifying those companies whose said 'under appreciated aspects' were under valued. They used the book as an example.

I only wandered in by accident because of the passing reference to sport and was, at that stage, more interested in arresting Nero than rebuilding Rome if you know what I mean.