it seems that in rogue finance, money laundry and rogue business some start to invent data that respect that law to look good.
the trick to detect that fake is to use the same rule for the correlation between data, which might be meaningless, but will respect the low. the idea of that law is that physical values don't care about the unit you use, and have no preference for round numbers. whatever are the unit, the coefficients, the statistics of first digits should be the same, and the onlyway to keep the scale invariance is having a log law for digit probability... 2012/3/16 Xavier Luminous <[email protected]> > Off the top of my head I'd like to mention that Benford's Law is > particularly good at rooting out cheaters. Basically, the most > significant digit from a sets of naturally occurring data tends to > follow a well known power law distribution. This is true for things > like lengths of rivers, street addresses, amounts entered on your > taxes, etc. > > I know they use this in voting already, but I'm not sure exactly how. > Would be interesting to see how this works out in this particular > case. > > > On Fri, Mar 16, 2012 at 5:03 AM, James Bowery <[email protected]> wrote: > > The reason I'm posting this to vortex-l is that of all the candidates, > the > > only one that represents a serious threat to establishment science is Ron > > Paul. > > > > The basic story is that a "signature" of "vote flipping" has turned up -- > > and the beneficiary in every case of this signature has been Mitt Romney. > > This analysis, if validated, could trigger the collapse of the Soviet, > er, > > American Empire. > > > > > http://www.ronpaulforums.com/showthread.php?363915-We-NEED-more-hands-on-deck.-Significant-evidence-of-Algorithmic-vote-flipping > . > > > > The first message is a good synopsis of the current arguments. > >
