Proposition. Let [math]a, b, c \in \mathbb{N\setminus \{0\}}[/math]. For all odd primes [math]a[/math] there exist [math]b, c[/math] such that [math]a^2 + b^2 = c^2[/math]. Namely [eqn]a\ \text{odd prime} \implies b = \frac{a^2 - 1}{2} \land c = b + 1.[/eqn]Proof. Subtract [math]b^2[/math] to yield [math]a^2 = c^2 - b^2[/math]. Factorize into [eqn]a^2 = (c + b)(c - b).\quad(1)[/eqn]Note that [math]c + b[/math] and [math]c - b[/math] are integers because theyre sums of integers.Let [math]a[/math] be odd prime. We know that the only divisors of a prime squared are [math]\{1, a, a^2\}[/math]. Thus there are only 3 cases for eq 1.(I) [math](c + b = 1 \land c - b = a^2) \lor[/math](II) [math](c + b = a \land c - b = a) \lor[/math](III) [math](c + b = a^2 \land c - b = 1).[/math]Case I. [math]c + b = 1[/math] is impossible if [math]b, c \in \mathbb{N} \setminus \{0\}[/math].Case II. This implies that [math]b = -b[/math] which is impossible for nonzero reals (may i remind you that [math]b \in \mathbb N\setminus \{0\}[/math]).So case III holds. And we know it MUST hold (means "no solution" is impossible) because of the following.From (III) derive [math]c = 1 + b[/math] and substitute it to get [math](1 + b) + b = a^2 \iff b = \frac{a^2 - 1}{2}[/math].Write [math]a = 2k + 1,\quad \forall k \in \mathbb N \setminus \{0\}[/math] (definition of odd numbers). Then [eqn]b = \frac{(4k^2 + 4k + 1) - 1}{2} = 2k^2 + 2k \in \mathbb{N} \land c = b + 1 \in \mathbb{N}.[/eqn]Note that this is not the case for a = 2. In fact there are no integer solutions to [math]2^2 + b^2 = c^2[/math].Therefore ALL odd primes [math]a[/math] generate a natural [math]b[/math] and [math]c[/math].Q.E.D.Example. If a = 13 then b should be (13^2 - 1) / 2 = 84 and c should be 84 + 1 = 85. This is true 13^2 + 84^2 = 85^2.
>>16825881For any odd n. This is because (n + 1)^2 - n^2 = 2n + 1 so every odd number is the difference between a pair of consecutive squares.
>>16825894I see I also forgot that the triples are not unique as in like (9, 12, 15) and (9, 40, 41). In the universe of positive odd numbers does the eq have exactly 1 solution IFF [math]a[/math] is odd prime?
>>16825922Sorry I meant something like [eqn]\forall b, c \in \mathbb{N} \setminus \{0\}\forall a\ \text{odd}\ (a^2 + b^2 = c^2\ \text{has exactly 1 solution} \iff a\ \text{prime})[/eqn]
>>16825922Yeah (n + k)^2 - n^2 = 2k + k^2 = k(2 + k) so if k ≠ 1, then the difference can't be prime
>>16825939*2nk , k(2n + k) you get it
>>16825939Right because c > b <=> c = b + k. Thanks Tama-anon.
>>16825938Slight nitpick it's forall positive odd a
>>16827421uhh he literally wrote >In the universe of positive odd numbers