A number theory problem by Lucas Tan

Consider a random positive integer x x . Subtract x x by its digit sum. [E.g. digit sum of 27 = 2 + 7 27 = 2 + 7 ]

Which is the smallest integer, aside from 1, that always divides this new integer?

For fun: Let the answer be α \alpha . Prove that for any multiple of α 2 \alpha^2 , the digit sum is always a multiple of α 2 \alpha^2 .

a) Prove by using Series

b) Prove by using Induction

c) Prove by using Modulo

3 7 5 11 2

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

5 solutions

Parth Sankhe
Dec 10, 2018

Let x = a + 10 b + 100 c + 1000 d . . . x=a+10b+100c+1000d...

The number concerned here would be a + 10 b + 100 c . . . ( a + b + c . . . ) = 9 b + 99 c + 999 d . . . a+10b+100c... - (a+b+c...)=9b+99c+999d... , which is always divisible by 9, hence always divisible by 3.

It's essentially the same hahah :)

Try out the "for fun" part of the question too!!

Lucas Tan - 2 years, 6 months ago
Lucas Tan
Dec 9, 2018

Let x = a l l r 1 0 r a r x = \sum_{all r} 10^r a_r , where a r , r Z a_r, r \in \mathbb Z , a r [ 0 , 9 ] a_r \in [0, 9] and r [ 0 , ) r \in [0, \infty) .

x a r x - \sum a_r

= 1 0 r a r a r = \sum 10^r a_r - \sum a_r

= ( 1 0 r 1 ) a r = \sum (10^r - 1)a_r

Assume 1 0 n 1 10^n - 1 is a multiple of 9 9 .

1 0 n + 1 1 10^{n+1} - 1

= 10 ( 1 0 n ) 1 = 10(10^n) - 1

= 10 ( 1 0 n 1 + 1 ) 1 = 10(10^n -1 + 1) -1

= 90 k + 9 =90k + 9 , where k Z k \in \mathbb Z 1 0 n + 1 1 \Rightarrow 10^{n+1} -1 is a multiple of 9 9 .

\therefore 1 0 n 1 10^n -1 is a multiple of 9 9 for all n n . (proven)

\therefore x a r = ( 1 0 r 1 ) a r x - \sum a_r = \sum (10^r - 1)a_r is a multiple of 9 9 for any positive integer x x . (proven)

Since the number is a multiple of 9 9 , it must be a multiple of 3 3 .

For fun:

a. Let x x from the above solution be a multiple of 9 9 x = 9 m \Rightarrow x = 9m , where m Z m \in \mathbb Z

Hence 9 m a r = 9 k 9m - \sum a_r = 9k

a r = 9 ( m k ) \sum a_r = 9(m - k)

\therefore the digit sum is a multiple of 9 9 .

b. Consider 9 n = 10 a 1 + a 0 9n = 10a_1 + a_0 , where a 0 , a 1 Z + a_0, a_1 \in \mathbb Z^+ . Assume a 0 + a 1 = 9 k , k Z + a_0 + a_1 = 9k, k \in \mathbb Z^+

For n + 1 n + 1 ,

9 n + 9 9n + 9

= 10 a 1 + a 0 + ( 10 1 ) =10a_1 + a_0 + (10 - 1)

= 10 ( a 1 + 1 ) + ( a 0 1 ) =10(a_1 +1) + (a_0 - 1)

( a 1 + 1 ) + ( a 0 1 ) = a 1 + a 0 = 9 k \therefore (a_1 + 1) + (a_0 -1) = a_1 + a_0 = 9k

To disprove a 1 + a 0 9 k a_1 + a_0 \neq 9k , consider n = 1 n = 1

9 ( 1 ) = 10 ( 0 ) + 9 9(1) = 10(0) + 9

a 1 + a 0 = 0 + 9 = 9 a_1 + a_0 = 0 + 9 = 9

a 1 + a 0 = 9 k \therefore a_1 + a_0 = 9k \forall 9 n = 10 a 1 + a 0 9n = 10a_1 + a_0 by induction.

From 9 k = a 1 + a 0 9k = a_1 + a_0 , let a 1 = 10 b 1 + b 0 a_1 = 10b_1 + b_0

9 k = 10 ( b 1 ) + ( a 0 + b 0 ) \therefore 9k = 10(b_1) + (a_0 + b_0)

a 0 + b 0 + b 1 = 9 k \Rightarrow a_0 + b_0 + b_1 = 9k' , where k Z + k' \in \mathbb Z^+

By induction, for all multiples of 9 9 , the digit sum is a multiple of 9 9 ( b 1 \because b_1 can be expressed as 10 c 1 + c 0 10c_1 + c_0 and so on).

c. Recall that 1 0 n 1 10^n - 1 is a multiple of 9 9 \forall n n .

Let 9 k = a l l r 1 0 r a r 9k = \sum_{all r} 10^r a_r , where a r , r , k Z a_r, r, k \in \mathbb Z , a r [ 0 , 9 ] a_r \in [0, 9] and r [ 0 , ) r \in [0, \infty) .

For any integer α \alpha ,

α ( 1 0 n ) \alpha (10^n) = α ( 1 0 n 1 + 1 ) = \alpha (10^n -1 +1)

= 9 m + α , m Z = 9m + \alpha , m \in \mathbb Z

α \equiv \alpha mod 9 9

9 k 0 \because 9k \equiv 0 mod 9 9 ,

1 0 r a r 0 \sum 10^r a_r \equiv 0 mod 9 9

a r 0 \sum a_r \equiv 0 mod 9 9

a r = 9 m \Rightarrow \sum a_r = 9m' , where m Z m' \in \mathbb Z

\therefore the digit sum of a multiple of 9 9 is a multiple of 9 9 .

Nishant Ranjan
Oct 19, 2019

Let's say we have a two digit no. with digits a and b. 10a + b - ( a+b) = 9a-9b = 9(a-b) Note that we can always factor out 9. So 9 is definitely a factor. Since 9 is a factor, we can be sure that 3 is also a factor as 9= 3*3. So we can be sure that 3 is the smallest factor other than 1. Note that this method works for all numbers no matter the number of digits.

Shubham Maurya
Dec 18, 2018

Never thought this also happens..ah nature has its own interesting patterns

Is this an explanation?

Costumed Creeper - 2 years, 2 months ago
Jesse Li
Dec 12, 2018

The divisibility rule of 3 is that the sum of all its digits is divisible by 3. If you take the sum of a number's digits and subtract it again, you will end up with 0, which is divisible by 3.

Sorry, could you explain how you know that the digit sum of a random integer is a multiple of 3, such that you canuse the divisibility rule of 3? Also, how do you always get 0 after subtracting off a random integer with its digit sum? Do clarify these 2 questions.

Lastly, as you were saying, 0 is divisible by 3, then is 0 not divisible by 2? By this logic, 2 should be the answer.

Lucas Tan - 2 years, 6 months ago

Log in to reply

0 is divisible by all numbers except 0. The divisibility rule of two is that the last digit is even, NOT that the digits sum is even. I never said that the sum of any integer's digits is divisible by three, but the digit sum divided by three will leave the same remainder as the original number divided by three, so you can substitute the digits sum for the original number. When you subtract the digits sum again, you get 0.

Jesse Li - 2 years, 5 months ago

Log in to reply

Ahh I think I understand what you mean now. Is this what you are trying to say?

Let x = 1 0 r a r x = \sum 10^r a_r

x b x \equiv b mod 3 3

Digits sum, a r b \sum a_r \equiv b mod 3 3

Hence x x - (digits sum) 0 \equiv 0 mod 3 3

Is this what you are trying to say? If so, could you do me a favour? I'm interested in seeing the proof for the divisibility rule of 3 3 , where the remainder obtained when a random integer is divided by 3 3 equals to the remainder obtained when its digits sum is divided by 3 3 :)

Lucas Tan - 2 years, 5 months ago

Log in to reply

@Lucas Tan That's kind of hard to explain, and honestly, I don't quite understand the proof either, but briefly, any power of 10 divided by 3 leaves a remainder of 1.

Jesse Li - 2 years, 5 months ago

Log in to reply

@Jesse Li It's okay, I understand the proof.

Note that r r mod 9 r 9 \equiv r mod 3 3 for any r Z r \in \mathbb Z , 9 = 3 2 \because 9 = 3^2 .

From my proof, I have shown that 1 0 r a r a r 10^r a_r \equiv a_r mod 9 9 .

Hence, 1 0 r a r a r 10^r a_r \equiv a_r mod 3 3 .

x x - (digits sum)

= 1 0 r a r a r = \sum 10^r a_r - \sum a_r

( a r a r ) \equiv (\sum a_r - \sum a_r) mod 3 3

0 \equiv 0 mod 3 3

Hence a random integer subtracted off its digits sum is a multiple 3 3 .

Lucas Tan - 2 years, 5 months ago

Let the number be xy. 10x+y-(x+y)=9x. So it is divisible by 9 so by 3

Rajputana Rupa Singh - 2 years, 4 months ago

Log in to reply

This only works for a 2-digit number. The proof you have is incomplete and insufficient.

Lucas Tan - 2 years, 4 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...