Sum of Digits Always Nine


by e(ho0n3
Tags: digits
e(ho0n3
e(ho0n3 is offline
#1
Jul8-04, 04:23 PM
P: 1,370
I don't know if this has been posted already, but anwho...

If you pick any positive integer greater than 9 and subtract the sum of its digits from that number, you'll end up with a multiple of 9. How do I know it's a multiple of 9? Curiously enough, the sum of the digits of a positive integer that is a multiple of 9 is a multiple of 9. Do this enough times and you'll end up with 9.

I tried this with other number systems and it seems to hold. This is the most interesting play of numbers if found to date.
Phys.Org News Partner Science news on Phys.org
Going nuts? Turkey looks to pistachios to heat new eco-city
Space-tested fluid flow concept advances infectious disease diagnoses
SpaceX launches supplies to space station (Update)
Muzza
Muzza is offline
#2
Jul8-04, 04:30 PM
P: 696
If you pick any positive integer greater than 9 and subtract the sum of its digits from that number, you'll end up with a multiple of 9.
Let a_n...a_1a_0 = a_0 * 10^0 + a_1 * 10^1 + ... + a_n * 10^n be a number (in base 10). Since 10 == 1 (mod 9), we have that 10^n == 1 (mod 9). Thus

a_0 * 10^0 + a_1 * 10^1 + ... + a_n * 10^n == a_0 + a_1 + ... + a_n (mod 9).

Subtracting the sum of the digits, a_0 + a_1 + ... + a_n, from both sides gives the desired result.

Curiously enough, the sum of the digits of a positive integer that is a multiple of 9 is a multiple of 9.
Suppose a_n...a_1a_0 = 9k for some integer k. Then a_0 * 10^0 + a_1 * 10^1 + ... + a_n * 10^n = 9k. Taking both sides modulo 9,

a_0 + a_1 + ... + a_n == 0 (mod 9),

as required.

That's enough mathematics for today ;)
Gokul43201
Gokul43201 is offline
#3
Jul8-04, 06:07 PM
Emeritus
Sci Advisor
PF Gold
Gokul43201's Avatar
P: 11,154
If you are not familiar with congruences and modular arithmetic, consider this...

Let the number N = w + 10x + 100y + 1000z + ... (example : 574 = 4 + 7*10 + 5*100 )
Its digits are w, x, y, z, ... So the sum of the digits is w + x + y + z + ... = S, say.

So, N - S = D = (w-w) + (10x-x) + (100y-y) + (1000z -z) + ... = 9x + 99y + 999z + ... = 9 (x + 11y + 111z + ...)

So, N - S is a multiple of 9.


Now for the second part...

Consider again, N = w + 10x + 100y + 1000z + ...
We want to prove that, if the sum of its digits, S is divisible by 9, then so is the number, N, and conversely.

If S is divisible by 9, then S = 9p. Now to this, add the number D, calculated above. S + D = 9p + D. But we saw previously that D itself is a multiple of 9, so D = 9q. Hence, S + D = 9p + 9q = 9(p+q) = 9r, say. But then, S + D is nothing but N. So N = 9r, which is what we wnted to prove.

The converse is proved by starting with N = 9r and subtracting D to yield S = 9(r-q) = 9p, say.
QED

agapeflight
agapeflight is offline
#4
Apr11-09, 10:51 PM
P: 1

Sum of Digits Always Nine


Nice proof, won't the same thing work for any base a with a-1 which is always an additive generator of the residue system?

The sum of digits of any number divisible by a-1 in base a is also divisible by a-1. ????

I'm too lazy to try to crank out the proof or even find a counter to it.

My question is did you gentlemen ever see the pattern of first appearance of multiples of nine when you sum the digits of multiples of nine in order?

It is quite interesting. http://agapeflight.net/Summing Digits.xls


Anyone know of any research related to that pattern? It appears that the distribution of digit sums would be roughly lognormal after some first appearance. I did find some stuff on research into this through the wiki on Digit Sums. Thanks.
Antuan
Antuan is offline
#5
May21-09, 07:29 PM
P: 20
I agree it’s very interesting. If you’d like to see some more information on this check out a previously thread called “Dissected Number Law”. Numbers 9 and 11 are key in this sort of arithmetic phenomena.

http://www.physicsforums.com/showthread.php?t=248279
AmberCam
AmberCam is offline
#6
Jul22-09, 12:24 PM
P: 5
Quote Quote by agapeflight View Post

It is quite interesting. http://agapeflight.net/Summing Digits.xls


Anyone know of any research related to that pattern?
I'm researching that pattern - but the creator of the above link (you?) needs to take the sum of digits to completion ie

Every total should have been reduced to one root digit 9
e.g 18 then add 1+ 8 = 9
27 then add 2 + 7 = 9

Anyway it is based on mod 9

This link may be helpful:
http://www.applet-magic.com/Digitsum00.htm
JazzFusion
JazzFusion is offline
#7
Jul22-09, 01:53 PM
P: 96
It is called 'casting out nines', and it has been used for centuries by accountants, mathemeticians, and scribes as a simple 'sanity-check' on results.

Before pocket calculators were readily available, and long- column addition and subtraction were routinely done by hand, this method was used to spot check for mistakes. Ancient scribes copying manuscripts by hand would likewise assign numeric values to specific letters, add them, and do spot-checks to keep hand-copied manuscripts free of errors as well.

The earliest use of the technique I have found referenced was by Hippolytos of Rome, a bishop from the early 3rd century AD.
Mensanator
Mensanator is offline
#8
Jul22-09, 11:49 PM
P: 105
Quote Quote by JazzFusion View Post
It is called 'casting out nines'
For something really neat, go to Slot Machine Arithmetic: Casting Out Cherries
JazzFusion
JazzFusion is offline
#9
Jul23-09, 05:38 PM
P: 96
THAT is classic!! Thanks for sharing it!


Register to reply

Related Discussions
last two digits of 2^999 Linear & Abstract Algebra 10
Sum of digits Precalculus Mathematics Homework 6
last digits of 3^999 Linear & Abstract Algebra 8
Pi to 24 digits General Discussion 38
Digits of Pi General Math 7