8

$N$ is a $50$ digit number (in the decimal scale). All digits except the $26^{th}$ digit (from the left) are $1$. If $N$ is divisible by $13$, find the $26^{th}$ digit.

This question was asked in RMO $1990$ and is very similar to this question and the same as this question but it is not solved by the approach used by me whereas I want to verify my approach.

My approach:

Suppose $N=111\cdots a\cdots111$ and $N\equiv 0\pmod {13}$

Now $N=10^{49}+10^{48}+\ldots+a10^{24}+\ldots+10+1=(10^{49}+10^{48}\ldots+10+1)+(a-1)10^{24}$

$N=\dfrac{10^{50}-1}{9}+(a-1)10^{24}$

Now $10^{12}\equiv 1\pmod {13}\Rightarrow 10^{24}\equiv 1\pmod {13}$ by fermat's little theorem.

Thus $(a-1)10^{24}\equiv (a-1) \pmod{13}\Rightarrow \dfrac{10^{50}-1}{9}\equiv 1-a\pmod{13}$ since $N\equiv 0\pmod{13}$

$10^{24}\equiv 1\pmod{13}\Rightarrow 10^{48}\equiv 1\pmod{13}$ or $10^{50}-1\equiv -5 \pmod{13}$

Now $10^{50}-1\equiv -5\pmod {13}\Rightarrow 9(1-a)\equiv -5\pmod{13}$

$a=3$ clearly satisfies the above conditions

$\therefore$ The $26^{th}$ digit from the left must be $3$.

Please suggest what is incorrect in this solution and advice for alternative solutions.

THANKS

Devansh Kamra
  • 1,951
  • 1
    $111111$ is divislbe by $13$ so only consider 25th and 26th digits – Math Lover Sep 01 '20 at 09:25
  • 1
    @Peter Yes, but by then they have already multiplied away the denominator; note how $4$ became $36$. – Arthur Sep 01 '20 at 09:26
  • @MathLover But we can also consider the residues of $\frac{10^{50}-1}{9}$ and $10^{24}$ mod $13$ to get the solution. – Peter Sep 01 '20 at 09:29
  • 1
    @Peter absolutely we can and OP's is the right way to do it. I was just making a comment that if we know $111111$ is divisible by $13$, we can ignore first $24$ and last $24$ $1's$. – Math Lover Sep 01 '20 at 09:32
  • 2
    @Peter If we want to add $10^{50} + 10^{49} + \cdots + 10 + 1$, then that becomes $\frac{10^{51} - 1}9$. Turns out, however, that that's not what we want to add. So yes, the exponent in the fraction should've been $50$, but only because the original sum was not the one we were after, not because of an arithmetical error. – Arthur Sep 01 '20 at 09:32
  • @MathLover In fact , this gives the solution immediate and we can easily show $$10^6\equiv 1\mod 13$$ with Euler's criterion , or easier by noting $13\mid 1001$ – Peter Sep 01 '20 at 09:38
  • Also a source – Clemens Bartholdy Jun 12 '21 at 23:13

4 Answers4

2

$10^{50}$ is a 51-digit number. And in a 50-digit number, the digit 26th from the left is represented by $10^{24}$.

Other than these two mistakes, I find your approach entirely reasonable. And if they were looking for a 51-digit number, with all except the 25th digit from the left being $1$, then it would've been correct too.

Edit: After having corrected these two off-by-one errors, the solution looks fine.

Arthur
  • 204,511
2

Another way is to use the trick from Wikipedia (that doesn't solve your solution)

Taking $N$ from the right, and applying the sequence $(1, −3, −4, −1, 3, 4)$ as instructed on the page (multiply the digits from the right by the given numbers in sequence), we get

$0$ for the 6 first digits from the right ($1-3-4-1+3+4=0$), repeating the sequence, $0$ up to digit 24 (from right), we still have $0$

Then, the next $6$ are our $a$ and $5\times 1$, or $$a-3-4-1+3+4\\=a-1$$

We did $30$ digits, $20$ to go. The next $18$ will give $0$, the last $2$ give $1-3$, thus the whole sum is $$a-1-2=a-3$$ The only digit that would have $a-3\equiv 0\pmod {13}$ is $$\bbox[5px,border:2px solid #ba9]{a=3}$$

Déjà vu
  • 974
1

After your editions, your approach is correct. Here's an alternative one:

The number $N$ consists of $24$ ones followed by the two digits $1a$ (the $2$-digit number $10+a$) followed by another $24$ ones, so with the number $M$ consisting of $24$ ones, $M:=\sum_{k=0}^{23}10^k=\frac{10^{24}-1}{9}$, we have $$N=M\cdot10^{24+2}+(10+a)\cdot10^{24}+M$$ Since $13$ is a prime, from Fermat's Little Theorem we know that $10^{12}\equiv1\pmod{13}$, and it follows that $13\mid(10^{12}-1)(10^{12}+1)=10^{24}-1=9M \Rightarrow 13\mid9 \lor 13\mid M$. Obviously, $13\nmid 9$, so $13\mid M$.

Now, if $13\mid N$, it follows that $13\mid (10+a)\cdot10^{24}$, and since $13\nmid10^{24}$, it must be $13\mid10+a$. Since $0\le a\le9$, it must be $a=3$.

Wolfgang Kais
  • 1,430
  • 8
  • 7
0

There's several tricks you can use but mostly they are similar to yours.

A famous well known trick is that as $1001 = 13*7*11$ so your number, $N$ is divisible by $13$ if and only if the $N- 1001*10^k$ is divisible by $13$ and so we can remove any pairs of $1$s if there are $3$ spaces apart. So we can get rid of the $1$ and $4$ one, the $2$nd and $5$ one, and the third and $6$th ones to get rid of the first $6$ ones ($111111\div 13 = 8547$ BTW). We can repeat that $4$ times to get rid of the first $24$ ones, and do it to the end to get rid of the last $24$ ones to and up with $11111...11d111.....11$ is divisble by $13$ if and only if $1d00000....000= (10+d)\times 10^{24}$ is.

Now $1001 = 13*7*11$ so $100\equiv -1 \pmod 13$ so $10^{24} = 1000^{8}\equiv (-1)^8\equiv 1 \pmod {13}$. So $(10+d)\times 10^{24}\equiv (10+d)\times 1\equiv 10+d \pmod {13}$ so if this is divisble by $13$ we must have $d = 3$.

That was tedious.....

We could also do, by Fermat's little Theorem $10^{12} \equiv 1 \pmod {13}$ so $10^{12}- 1 =999999999999 \equiv 0 \pmod 13$ so $13$ divides $999999999999 = 9\times 111111111111$ and so $13$ divides $9$ or $111111111111$ so $13|111111111111$ and we do similar to above to get $(10+d)\times 10^{24}$ and as $10^{12} \equiv 1$ then $10^{24} \equiv 1$ and $10+d\equiv 0$ so $d = 3$.

.....

Or we could realize the remainder of $10\div 13$ is $10$. The remainder of $10^2 \div 13$ is $9$ and so on, and these must eventually cycle through. Just list them all: $10 \equiv 10; 10^2\equiv 9; 10^3 \equiv 12 \equiv -1$. SO $10^4\equiv -10\equiv 3$ and $10^{5}\equiv -9\equiv 4$ and $10^6\equiv 1$ and then it repeats. And add them all up. (In groups of $6$ whe get $\sum_{k=0}^5 10^k \equiv 1+10 + 9+(-1)+(-10)+(-9) \equiv 0$ so $13|111111$)

All of theses are more or less the same idea and lead to the conclusion $d=3$.

fleablood
  • 130,341