What is Digg?Sponsored by Activision
Band Hero view!
guitarhero.com - The biggest event music event of the year is now in your living room.
49 Comments
- skimitar, on 10/12/2007, -0/+2I remember reading about this years ago in New Scientist (unfortunately, they no longer have a copy of the article on their site - it was called the "Power of One") - that's not a criticism BTW - i dugg it - numbers never get old.
Another good write up (though more technical) is at http://mathworld.wolfram.com/BenfordsLaw.html
I had access to a large numeric database at the time through work and tested this following the NS article - sure enough, the log probabilities held. Confirming something for yourself is always fun (i.e. I really needed to get a life) - ryan_merket, on 10/12/2007, -1/+2In other words, if you want to cheat on your taxes be sure to use numbers that start with a 1 NOT a 5.
Thanks for the tip. - volatilegx, on 10/12/2007, -0/+1Difficult to understand why until you read the Dow Jones Industrial Average example at the bottom of the page.
- Morieris, on 10/12/2007, -0/+1in binary, all the numbers start with 1.
ooh, spooky - jeremymerrill, on 10/12/2007, -0/+0whoops, site got digg-bombed.
- hugotheschmugo, on 10/12/2007, -0/+0Try the coralcdn cache
http://www.rexswain.com.nyud.net:8090/benford.html - thewebguy, on 10/12/2007, -0/+0actually i see what i was missing. it takes 20,000 times 500% to reach 100,000, but it takes 100,000 times 200% to get to 200,000. there may be more time spent between 20,000 and 100,000 than there is between 10,000 and 20,000, but the time spent per ten thousand from 20,000 and 100,000 is far less than that first 10k.
lol - hugotheschmugo, on 10/12/2007, -0/+0Or the google one
http://72.14.203.104/search?q=cache:X7k61spjqVgJ:www.rexswain.com/benford.html+&hl=sv&gl=se&ct=clnk&cd=1&client=firefox-a - gwjc, on 10/12/2007, -0/+0the wiki link... it is pretty kewl.. dugg:
wolever, on 10/12/2007, -0/+0I've mirrored it at http://wolever.net/benford/
This is served off my DSL connection though, so if more people woulden't mind taking a bit of the hit it would be nice. - 4850lu73, on 10/12/2007, -0/+0digg effect in action yet again
- Stoutlimb, on 10/12/2007, -0/+0I went to this site, and Siteadvisor marked it as a "red" site, meaning that it had trojans and/or spyware on the site. Here is the link to the Siteadvisor analysis... http://www.siteadvisor.com/sites/rexswain.com Apparently they rated it as a "6" on the nuisance meter.
- thewebguy, on 10/12/2007, -0/+0i don't know if i get it, once you get to 20,000, you have to gain 500% to get to the next number that starts with 1 (being 100,000)
- inactive, on 10/12/2007, -0/+0crazy stuff
- MikeKnoop, on 10/12/2007, -0/+0Yea you'll need some braincells to understand the logic. By the way, has anyone noticed the square picture animated gif underneath the "who blogged this"? What is that?
-Mike - semiotix, on 10/12/2007, -0/+0This is comment number 18.
Ooooh, it works! - PercyKittenz, on 10/12/2007, -0/+0I don't find this too statistically interesting. It's the nature of numbers that lower numbers are easier to come by than higher numbers in any finite situation.
Let's say you're looking at house numbers from the first 300 pages of the phone book. Given that you're also only looking at addresses that have at least four digits as the article states is a requirement for this to work. Okay. You're going to find that the first digit is one 15% more often than it's two because there are fewer streets that are long enough to get to the 2000s. You have to have a equal sample for it to mean anything.
To put it another way, looking at the area of rivers, smaller rivers are more common than larger rivers, so you'll find more rivers in the sample in the 1000's than in the 2000's, and even fewer in the 3000's and so on. These aren't just random numbers we're talking about. - dirtyfratboy, on 10/12/2007, -0/+0@MikeKnoop
I think that "square picture animated gif" is called an "ad." - bmw@, on 10/12/2007, -0/+0Re rtphokie's "crap" comment: not so ...
http://mathworld.wolfram.com/BenfordsLaw.html - leviathan3k, on 10/12/2007, -0/+0stoutlimb:
EICAR is a test virus. It's specifically meant to be detected without actually having any sort of dangerous payload. And as it's the only thing that seems to be on their site, it's harmless. - hugotheschmugo, on 10/12/2007, -0/+0Trying to quote Wikipedia again, pretty good text I think:
- bmw@, on 10/12/2007, -0/+0Whoops! Sorry, skimitar ... just saw that your ref beat me.
- dognose, on 10/12/2007, -0/+0hmm 20% of their figures begin w/ 1
and, why is there no zero? - akatrito, on 10/12/2007, -0/+0i read this, its pretty intriguing. i find it interesting to know that numbers arent random that they fit a certain pattern.
- StephnDolenc, on 10/12/2007, -0/+0great story; a bit redundant at times, but not a problem
- seventoes, on 10/12/2007, -0/+0Thats really weird... thats like the only thing having to do with numbers that ive ever read without being forced... Kinda of a cool theory...law... how did that become a law? Its just a probability thingy..
- RichMorin, on 10/12/2007, -0/+0I decided I had to check this out for myself, so I tallied the first digits from the file sizes on my Mac OS X system. It convinced me...
http://www.cfcl.com/~rdm/weblog/archives/000942.html - capajc, on 10/12/2007, -0/+0Damn cool.
- prousa, on 10/12/2007, -0/+0"Let's say that the Dow goes up at a rate of about 20 percent a year. That means that it would take five years to get from 1 to 2 as a first digit."
Wait. 5 years at 20 percent a year to go from $1,000 to $2,000? Anybody who has taken basic math knows it's not ... it works like this:
Initial -- $1,000
1st year -- $1,200 or 1,000x1.2
2nd year -- $1,440 or 1,200x1.2
3rd year -- $1,728 or 1,440x1.2
4th year -- $2,07.36 or 1,728x1.2 and then we've broken $2,000. - jasqwerty, on 10/12/2007, -0/+0I've fudged numbers before and I really never felt a need to go in the middle. Why do all these fraudsters like the 5s and 6s?
- madmathmatician, on 10/12/2007, -0/+0"in binary, all the numbers start with 1.
ooh, spooky"
lol, math saves the day again ;) - Trjn, on 10/12/2007, -0/+0""Let's say that the Dow goes up at a rate of about 20 percent a year. That means that it would take five years to get from 1 to 2 as a first digit."
Wait. 5 years at 20 percent a year to go from $1,000 to $2,000? Anybody who has taken basic math knows it's not ... it works like this:
Initial -- $1,000
1st year -- $1,200 or 1,000x1.2
2nd year -- $1,440 or 1,200x1.2
3rd year -- $1,728 or 1,440x1.2
4th year -- $2,07.36 or 1,728x1.2 and then we've broken $2,000."
I see 5 years.
The initial year is year 1, instead of year 0.
Although I'm pretty sure most people who have bothered with more advanced maths would have started counting from zero as well, thats how I initially did it too, had to go back and look again. - Otto, on 10/12/2007, -0/+0It's interesting, but what the article doesn't mention is that Benford's "law" has been found, many times, to not actually apply in all situations.
- 3rdMEgo, on 10/12/2007, -0/+0really interesting
+digg - trogdoor, on 10/12/2007, -0/+0"in binary, all the numbers start with 1.
ooh, spooky"
except 0. - whitehatlurker, on 10/12/2007, -0/+0@dognose "and, why is there no zero?"
If you counted numbers starting with a zero, you'd have to conclude they all start with zero, we just leave off leading zeros for numbers 1 or greater. That's somewhat outside the scope of the problem.
The original article doesn't do a good explaining the phenomenon. See the Wolfram site for a better description. - inactive, on 10/12/2007, -0/+0I put money that the person who posted this owns a mac!
***** libs! - OBKenobi, on 10/12/2007, -0/+0I think Benford is nuts.
- Al_Bean, on 10/12/2007, -0/+0I submitted this link a while ago, its a link to a 15 minute audio piece from the BBC about 1 (talks about the fraud detection methods + more)
http://digg.com/science/1_(one)_-_Literally,_the_most_popular_number_ - dalle, on 10/12/2007, -0/+0Great article, I like these kind of stuff, although I suck at math. :-D
- klamathvx, on 10/12/2007, -0/+0Whoa, before I hit digg, this was at 1111 diggs.
- FrostyTheDMan, on 10/12/2007, -0/+0Something tells me that the number 1 comes up on the logarithm that tells us how many times Benford (or any mathematician) gets laid... still better than mine though (lil self deprecation in case the IRS mathematicians are reading my post).
- cvrti5, on 10/12/2007, -0/+0Interesting, but as a mathematician I am not impressed by the explanation.
- AgentLars, on 10/12/2007, -0/+0"i read this, its pretty intriguing. i find it interesting to know that numbers arent random that they fit a certain pattern."
Actually, I don't think it is saying that numbers aren't random per se, but that SOME numbers aren't random and you can spot the potential fakes by the fact they may have been changed to seem random. Then maybe I've missed the point too. I only read it 1 time, which seems to be the best number... - hugotheschmugo, on 10/12/2007, -1/+0The following, quoted from Wikipedia,
- Timbero, on 10/12/2007, -1/+0pretty cool story. and also 7 years old ;)
- kkaabboomm, on 10/12/2007, -1/+0is it sad that i learned about this in one of my stats classes here at gatech?
- pkulak, on 10/12/2007, -1/+0Actually, if you're going to cheet on something, don't make up the numbers yourself; you can't do it. Use street addresses from the phone book, or something.
- inactive, on 10/12/2007, -3/+0wow, that was crap
The Digg Toolbar for Firefox lets you Digg, submit content, and keep track of Digg even when you're not on the Digg site. Download the official 
© Digg Inc. 2009
— Content created and posted by Digg users is