Benford's law

[scooterscot]

It's smeging disturbing.

I've tried testing some random data sets from bank statements to historical values, nearly every time the leading digit '1' comes up 30% of the time.

What's going on?

 

[malvolio]

It works after the decimal point as well. Many years ago an American programmer made a lot of money by realising the actual mid-point for rounding was not 50 cents but roughly 33 cents. He wrote a programme that reported the rounding based on 0,5 but calculated it at 0.3 and put the balance in his own account.

 

[barrydidit]

It works after the decimal point as well. Many years ago an American programmer made a lot of money by realising the actual mid-point for rounding was not 50 cents but roughly 33 cents. He wrote a programme that reported the rounding based on 0,5 but calculated it at 0.3 and put the balance in his own account.

That was Superman 3 you're thinking of.

 

[malvolio]

That was Superman 3 you're thinking of.

In 1968...? Don't think so.

 

[Contreras]

It's smeging disturbing.

I've tried testing some random data sets from bank statements to historical values, nearly every time the leading digit '1' comes up 30% of the time.

What's going on?

For business/personal current account transactions it's nearer 40% here!

Digit		Business	Personal	Predicted
1		38.7%		42.9%		30.1%
2		16.7%		20.0%		17.6%
3		8.9%		10.9%		12.5%
4		8.9%		7.0%		9.7%
5		9.5%		6.6%		7.9%
6		4.2%		2.1%		6.7%
7		5.4%		2.9%		5.8%
8		2.4%		3.8%		5.1%
9		5.4%		3.8%		4.6%

Sample size	336		1535

 

[scooterscot]

For business/personal current account transactions it's nearer 40% here!

Digit		Business	Personal	Predicted
1		38.7%		42.9%		30.1%
2		16.7%		20.0%		17.6%
3		8.9%		10.9%		12.5%
4		8.9%		7.0%		9.7%
5		9.5%		6.6%		7.9%
6		4.2%		2.1%		6.7%
7		5.4%		2.9%		5.8%
8		2.4%		3.8%		5.1%
9		5.4%		3.8%		4.6%

Sample size	336		1535

Perhaps there's a fraud in progress.

 

[OwlHoot]

It's smeging disturbing.

I've tried testing some random data sets from bank statements to historical values, nearly every time the leading digit '1' comes up 30% of the time.

What's going on?

Could it be that your annual salary is £15000, and your monthly pay packet is £1005 ?

 

[Contreras]

Perhaps there's a fraud in progress.

I was just thinking that too.

 

[MicrosoftBob]

It works after the decimal point as well. Many years ago an American programmer made a lot of money by realising the actual mid-point for rounding was not 50 cents but roughly 33 cents. He wrote a programme that reported the rounding based on 0,5 but calculated it at 0.3 and put the balance in his own account.

Was that before or after Office Space ?

 

[northernladuk]

And for all the other thick bastards like me that don't have a clue what's happening here...

Benford's law - Wikipedia, the free encyclopedia

Interesting comment in there to help fend off HMRC for the less scrupulous of you though..

Accounting fraud detection[edit]

In 1972, Hal Varian suggested that the law could be used to detect possible fraud in lists of socio-economic data submitted in support of public planning decisions. Based on the plausible assumption that people who make up figures tend to distribute their digits fairly uniformly, a simple comparison of first-digit frequency distribution from the data with the expected distribution according to Benford's Law ought to show up any anomalous results.[14] Following this idea, Mark Nigrini showed that Benford's Law could be used in forensic accounting and auditing as an indicator of accounting and expenses fraud.[15] In practice, applications of Benford's Law for fraud detection routinely use more than the first digit.[15]