2
$\begingroup$

Is it true that, for any Pythagorean triple $4ab > c^2$?

So this came up in a proof I was working on and it seems experimentally correct from what I've tried and I would imagine the proof is similar to proving,

$$ab < \frac{c^2}{2}$$

The idea I have for this approach is,

$$4ab > c^2$$ $$4ab > a^2 + b^2$$ (Then maybe something with the triangle inequality?)

So is this statement true (it seems to be), and how can I prove it?

A counter-example would also be acceptable.

$\endgroup$
3
  • 1
    $\begingroup$ I don't think it will be true if the triangle is long and thin. $\endgroup$ Dec 3 '19 at 19:34
  • $\begingroup$ I'll do some more testing with that in mind thanks $\endgroup$
    – PMaynard
    Dec 3 '19 at 19:35
  • $\begingroup$ A counter example would also be acceptable $\endgroup$
    – PMaynard
    Dec 3 '19 at 19:36
3
$\begingroup$

$15^2 + 112^2 = 113^2$, but $4 \cdot 15 \cdot 112 \le 113^2$.

$\endgroup$
1
  • $\begingroup$ Thanks! I really didn't want it to be true because it would have resolved another question I was working on. None the less greatly appreciative. $\endgroup$
    – PMaynard
    Dec 3 '19 at 19:41
1
$\begingroup$

If $a^2 + b^2 = c^2$ then

$c^2 < 4ab$ is the same thing as claiming $a^2 + b^2 < 4ab$

This is true if and only if $a^2 - 2ab + b^2 < 2ab$ which is to say

$(a-b)^2 < 2ab$

There shouldn't be any reason that should be true in general. If let $m = $ average/midpoint of $a,b$, that is, $m = \frac {a+b}e$ and $e = $ the radius of the interval $a$ to $b$, that is $e = |m-a| =|b-m| = \frac{|a-b|}2$ then we have

$(a-b)^2 < 2ab \implies$

$4e^2 < 2(m-e)(m+e) = 2(m^2 - e^2)\implies$

$3e^2 < m^2$

Which utterly need not be true!

Of course finding a pythagorean triplet where $a= m\pm e$ and $b=m\mp e$ are integers so that $a^2 + b^2 = 2(m^2 + e^2)$ is not merely and integer but a perfect square may put a wrinkle in finding a counterexample.

But not an insurmountable wrinkle.

We need, for a counter-example, $3e^2 \ge m^2$ which just require $e$ be significantly large, which is utterly irrelevent (it would seem) to $2(m^2 + e^2)$ being a perfect square.

... But to find a counter example:

.....

Back to square 1: We can find a Pythogorian triple $a^2 + b^2 = c^2$ by letting $b$ be any odd integer and if $b^2 = 2a+1$ so $c^2 = (a+1)^2$ ... retrosolving by letting $a = \frac {b^2-1}2$.

So we want a counter example of $c^2 = a^2 + b^2 = a^2 + 2a + 1 = (a+1)^2 \ge 4ab$ or substitutint $a = \frac {b^2-1}2$, we want a case where

$(\frac {b^2+1}2)^2 \ge 4\frac{b^2 -1}2b$.

Surely we can find a counter example.

$(b^2 + 1)^2 \ge 8(b^2-1)b$

$b^4 + 2b^2 + 1 \ge 8b^3 -8b$

$b^4 -8b^3 +2b^2 - 8b + 1 \ge 0$

yeah that can be solved...

(Obviously as $b\to \infty$ the $b^4-8b^3 + 2b^2 - 8b + 1\to \infty$ so there is some $K$ where $b > K$ will always have $b^4 -8b^3 +2b^2 - 8b + 1\ge 0$.)

Taking a sledgehammer and letting $b > 8$, say, $b=9$ and $a = \frac {b^2 -1}2=40$ so $40^2 + 9^2 = 41^2> 4*9*40=36*40$

Yeah... it's not true.

(Actually our sledgehammer was fairly precise. $b^4 -8b^3 +2b^2 - 8b + 1$ has only two real solutions; one between $0$ and $1$ (far closer to $0$ than to $1$) and the other between $7$ and $8$. So of primmitive triplets where $b$ is odd and $b^2 = 2a+1$. $a = 40, b=9, c=41$ is the smallest primitive triplet (and thus smallest of all triplets) where $c^2 > 4ab$.)

( so $c^2 < 4ab \iff \frac {c}{\gcd(a,b,c)} < 41$)

$\endgroup$
2
  • $\begingroup$ Thanks! Also I wish I could give +2 for the sledgehammer as it makes a way to construct a counterexample! Certainly a useful tool $\endgroup$
    – PMaynard
    Dec 3 '19 at 20:30
  • 1
    $\begingroup$ Actually $a=40$ and $b = 9$ is the smallest primitive counterexample but all with $8< b< a$ will fail. $\endgroup$
    – fleablood
    Dec 3 '19 at 20:32

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.