It must be verified that for all $x \in \mathbb{Z}$ it holds that $x^2 - 1 \bmod{8} \in \{0, 3, 7\}$. First some definitions. Using the following theorem a definition for $\bmod$ is provided:
Theorem. For all $a \in \mathbb{Z}$ and $d \in \mathbb{N}$ unique integers q, r exist satisfying $a = q \cdot d + r \wedge 0 \leq r <d$
The integers $q$, $r$ correspond to the quotient and remainder, respectively, of $a$, these being defined as:
$a = (a / d) \cdot d + a \bmod d \wedge 0 \leq a \bmod d < d$
With $q = (a/d)$ and $r = a \bmod d$. To verify the claim I use induction. I use a brute force approach where I search for all numbers $x$ such that $x^2 - 1 \bmod 8 = 0$, $x^2 - 1 \bmod 8 = 3$ or $x^2 - 1 \bmod 8 = 7$. First I search for all $x$ such that $x^2 - 1 \bmod 8 = 0$. If $x^2 - 1 \bmod 8 = 0$ then 8 is a divisor of $x^2 - 1$, i.e.: $8 \mid x^2 - 1$. Now I search for some other $x$ such that 8 is a divisor of $x^2 - 1$. Let $f(x) = x^2 - 1$ and $8 \mid f(x)$ then $8 \mid f(x + a)$ for some $a \in \mathbb{N}$. I compute this a:
$8 \mid f(x) = 8 \mid x^2 - 1 \implies 8 \mid (x + a)^2 - 1 = 8 \mid x^2 + a^2 + 2ax - 1 = 8 \mid x^2 - 1 + a^2 + 2ax$
The implication holds for e.g.: $a = 4$ since:
$8 \mid x^2 - 1 + a^2 + 2ax \implies 8 \mid x^2 - 1 + 16 + 8x = 8 \mid f(x) \wedge 8 \mid 16 + 8x = true$
So it can be concluded that if $8 \mid f(x)$ then $8 \mid f(x + 4)$. In fact it may be concluded that if $8 \mid f(x)$ then $8 \mid f(x + 4k), \forall x \in \mathbb{Z} \wedge k \in \mathbb{N}$, since:
$8 \mid (x + 4k)^2 - 1 = 8 \mid x^2 - 1 + 16k^2 + 8kx = 8 \mid f(x) \wedge 8 \mid 16k^2 + 8kx$
So it holds that $8 \mid f(x) \implies 8 \mid f(x + 4k)$. By inspection one finds that $8 \mid f(1) = 8 \mid 1^2 - 1 = 8 \mid 0 = true$ and so $f(x) \bmod 8 = 0$ for all $x \in \{1, 5, 9, 13, ...\}$. Since $f(x)$ is symmetric it holds that $f(-x) = f(x)$ and so it follows that $f(x) \bmod 8 = 0$ for all $x \in \{-1, -5, -9, -13, ...\}$. Upon further inspection one finds that $8 \mid f(3) = 8 \mid 3^2 - 1 = 8 \mid 8 = true$ and so it follows that $x^2 - 1 \bmod 8 = 0$ for all $x \in \{3, 7, 11, 15, ...\}$ and all $x \in \{-3, -7, -11, -15, ...\}$
Next I search for all x such that $x^2 - 1 \bmod 8 = 3$, or equivalently $x^2 - 4 \bmod 8 = 0$. Using the same approach I find that $x^2 - 1 \bmod 8 = 3$ for all $x \in \{2, 6, 10, 14, ...\}$ and all $x \in \{-2, -6, -10, -14, ...\}$. Finally $x^2 - 1 \bmod 8 = 7$ for all $x \in \{0, 4, 8, 12, ...\}$ and $x \in \{0, -4, -8, -12, ...\}$. And so one concludes that for all $x \in \mathbb{Z}$ it holds that $x^2 - 1 \bmod{8} \in \{0, 3, 7\}$.
This question comes from a course in discrete mathematics in computer science. I feel that my approach is overkill and that I'm doing something wrong and that there must be some cleverer way of solving the problem. If anyone can help with a better, cleaner, approach, or point out errors, it will be greatly appreciated :).

- 1,191
-
3Why the down vote? – J. W. Tanner Nov 25 '22 at 00:07
-
2True, no reason to downvote the question. Follows community standards. – lafinur Nov 25 '22 at 00:42
-
3It's getting ridiculous when every question is a "duplicate" of some related question. I think Bill is misusing his gold badge. – Suzu Hirose Nov 25 '22 at 05:52
-
1@Suze It should not be surprising that a large site like this over 12 years old already has almost all methods of answering common exercises. – Bill Dubuque Nov 25 '22 at 09:43
-
1See the linked dupe(s) for many different approaches, e.g. eight methods are in this answer, some of which generalize widely. See also the manu "Linked" questions in the sidebar of the dupes. – Bill Dubuque Nov 25 '22 at 12:55
4 Answers
First of all, if $x^2 - 1 \equiv y \pmod 8$, then $$(x+4)^2 - 1 = (x^2 - 1) + 8(x + 2) \equiv y \pmod 8;$$ that is to say, $x^2-1$ and $(x+4)^2 - 1$ have the same remainder upon division by $8$, for all integers $x$.
Hence it suffices to test $x \in \{0, 1, 2, 3\}$. We have $$\begin{align} 0^2 - 1 &\equiv 7 \pmod 8, \\ 1^2 - 1 &\equiv 0 \pmod 8, \\ 2^2 - 1 &\equiv 3 \pmod 8, \\ 3^2 - 1 &\equiv 0 \pmod 8. \end{align}$$
This concludes the proof.

- 135,869
-
Please strive not to post more (dupe) answers to dupes of FAQs, cf. recent site policy announcement here. – Bill Dubuque Nov 25 '22 at 04:50
$$x^2-1=(x-1)(x+1)$$ If $x$ is even, $x-1$ and $x+1$ are both odd. The possibilities are then $7\times1$, $1\times3$, $3\times5$ and $5\times7$, which gives you $3$ or $7$.
If $x$ is odd, $x-1$ and $x+1$ are both even, and one of them is divisible by $4$, so the remainder is always $0$.

- 11,660
-
1At least read the answer before claiming it's a duplicate. This approach is different from all three other answers posted. – Suzu Hirose Nov 25 '22 at 05:49
-
Arguments like this have already been posted hundreds of time here, e.g. above is already in the 3rd proof in this answer in the 1st dupe as well as many other places (follow the "Linked" questions there and/or search with approach0.xyz $\ \ $ – Bill Dubuque Nov 25 '22 at 13:00
Since $x^2-1\bmod8$ only depends on $x\bmod8,$ its possible values are obtained by taking 8 consective values for $x,$ e.g. $x\in\{-3,-2,-1,0,1,2,3,4\}.$
$(\pm3)^2-1\equiv0\bmod8.$
$(\pm2)^2-1\equiv3\bmod8.$
$(\pm1)^2-1\equiv0\bmod8.$
$0^2-1\equiv7\bmod8.$
$4^2-1\equiv7\bmod8.$
Your observation that $(x+4)^2\equiv x^2\bmod8$ was clever but not really worthwile.

- 34,650
-
Please strive not to post more (dupe) answers to dupes of FAQs, cf. recent site policy announcement here. – Bill Dubuque Nov 25 '22 at 04:50
- If $x = 4 k$ with $k \in \mathbb{Z}$ then : $$x^2 - 1 = 16 k^2 - 1 = 16 k^2 - 8 + 7 \equiv 7 mod 8$$
- If $x = 4 k + 1$ with $k \in \mathbb{Z}$ then : $$x^2 - 1 = 16 k^2 + 8 k + 1 - 1 = 16 k^2 + 8 k \equiv 0 mod 8$$
- If $x = 4 k + 2$ with $k \in \mathbb{Z}$ then : $$x^2 - 1 = 16 k^2 + 16 k + 4 - 1 = 16 k^2 + 16 k + 3 \equiv 3 mod 8$$
- If $x = 4 k + 3$ with $k \in \mathbb{Z}$ then : $$x^2 - 1 = 16 k^2 + 24 k + 9 - 1 = 16 k^2 + 24 k + 8 \equiv 0 mod 8$$

- 1,733
-
Please strive not to post more (dupe) answers to dupes of FAQs, cf. recent site policy announcement here. – Bill Dubuque Nov 25 '22 at 04:50