For natural numbers $n\ge m$, let $n\underset{m \text{ times}}{\underbrace{!!!\dots!}}=n(n-m)(n-2m)(n-3m)\dots$ where all factors are natural numbers (we exclude $0$ and negative factors).


What is the units digit of $1!+2!+3!+4!!+5!!+\dots+k\underset{\left \lfloor \sqrt{k} \right \rfloor \text{ times}}{\underbrace{!!!\dots!}}+\dots+1992\underset{44 \text{ times}}{\underbrace{!!!\dots!}}$? ($\left \lfloor \cdot \right \rfloor$ denotes the floor funtion).

My Attempt (Is wrong as Peter Foreman commented below):

Consider the first $9$ terms:



Each of the remaining terms includes at least on factor that ends with $0$. Therefore, the each term ends with $0$.

Hence the units digit of the given expression is equal to the units digit of the sum of the first $9$ terms. So, $1$ is the units digit of the given expression.

Peter Foreman said: "$17!!!!=9945$". This showed me that my attempt is wrong. Thanks Peter Foreman.

Any help would be appreciated. THANKS.

  • $\begingroup$ @PeterForeman ,, yeah , you are right that I am wrong. $\endgroup$ Aug 22 '19 at 8:25
  • $\begingroup$ It's also going to be non-zero whenever the number of ! is a multiple of 5 and n isn't. $\endgroup$
    – user694818
    Aug 22 '19 at 8:38
  • $\begingroup$ @MatthewDaly Right. So can not we solve this problem by ordinary methods? $\endgroup$ Aug 22 '19 at 8:42
  • $\begingroup$ I'd write a program that spits out the first hundred unit digits and stare at it for a while. The pattern will be easier to prove when you know what it is. $\endgroup$
    – user694818
    Aug 22 '19 at 8:51
  • $\begingroup$ It's $1$. (I don't know how to prove it without computation) $\endgroup$
    – Vepir
    Aug 22 '19 at 9:58

When $k \geq 25$ and the floor part $p$ of $k^{0.5}$ is coprime with $10$, the unit digit of $k’=k! \ldots !$ is $0$ (there is an even number in $k,k-p$ and one divisible by $5$ in $k!!!!!$).

When $k \geq 25$ and $p \wedge 10=2$, there is going to be a number divisible by $5$ in $k,k-p,k-2p,k-3p,k-4p$, and $k’$ is divisible by $5$ and congruent to $k$ mod $2$, so the unit digit of $k’$ is $5$ if $k$ is odd and $0$ if $k$ is even.

When $k \geq 25$ and $p \wedge 10=5$, then $k(k-p)$ is even, so $k’$ is even. The congruence mod $5$ is trickier: $k’$ is congruent to $k^r$ mod $5$, where $r$ is the number of factors in he product (ie $r-1$ is the floor part of $(k-1)/p$, so $r$ is either $p$, $p+1$ or $p+2$).

When $10 | p$, as above, $k’$ is congruent to $k^r$ mod $10$.

Note that everything depends only on the unit digit of $k$, $p$ and $r$: when $p$ is set, and $p \wedge 10=2$, the sum of any four $k’$ corresponding to consecutive $k \geq 25$ vanishes mod $10$.

When $p,r$ are set and $p \wedge 10=5$, the sum of any five $k’$ corresponding to consecutive $k \geq 25$ is always divisible by $10$.

When $p,r$ are set and $p \wedge 10=10$, the sum of any ten $k’$ corresponding to consecutive $k \geq 100$ is congruent to $3$ mod $10$ if $4|r$ and $5$ mod $10$ otherwise.

So all we need now is time for processing all integers from $1$ to $1992$.

  • $\begingroup$ Why the downvotes? $\endgroup$
    – Mindlack
    Aug 22 '19 at 9:33
  • $\begingroup$ Can you please clarify what do you mean by the notations $k'$, $p \wedge 10=10$,...? Thank you very much. $\endgroup$ Aug 24 '19 at 10:52
  • $\begingroup$ $\wedge$ is the gcd. $p$ is the floor part of $\sqrt{k}$. $r-1$ is the floor part of $(k-1)/p$. $k’=k(k-p)(k-2p)\ldots(k-(r-1)p)$. $\endgroup$
    – Mindlack
    Aug 24 '19 at 11:08
  • $\begingroup$ So this problem is not solvable unless we find the units digit of all $1992$ terms, right? Is not there any method to solve it? $\endgroup$ Aug 24 '19 at 11:19
  • $\begingroup$ The answer outlines many possible shortcuts, but why would you expect there to be an instant solution without any calculation? $\endgroup$
    – Mindlack
    Aug 24 '19 at 13:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.