10

If we have a right triangle then the inradius is equal to $$r=\frac{a+b-c}2,$$ where $c$ is the hypothenuse and $a$ and $b$ are the legs.

This formula is mentioned in various places and it can be useful both in geometric problems and in problems on Pythagorean triples.1

Question: How can one derive this formula?


1It is stated on Wikipedia (in the current revision without a reference). Some posts on this site where this equation (or something closely related) is mentioned: If the radius of inscribed circle in a right triangle is $3 cm$ and the non-hypotenuse side is $14cm$, calculate triangle's area., Prove the inequality $R+r > \sqrt{2S}$, In a Right Angled Triangle., How do I find the radius of the circle which touches three sides of a right angled triangle?, Is there a way to see this geometrically?, Range of inradius of a right Triangle.


I will mention that I would be able to derive this myself in some way. (And some of the posts linked above in fact include something which basically leads to a proof.) Still, I think that it is to have somewhere this nice fact as a reference. And I wasn't able to find on this site a question specifically about this problem. I can post an answer myself - but I wanted to give others an opportunity to make a post first.

ACB
  • 3,713
  • 1
    One possible way: let $I$ be the incenter and suppose that incircle touches legs $AC$ and $BC$ at points $P$ and $Q$, respectively. Then, $PIQC$ is a square, so $r=PI=CP$. – richrow Oct 19 '21 at 10:48
  • A sketch is outlined here https://artofproblemsolving.com/wiki/index.php/Inradius – ə̷̶̸͇̘̜́̍͗̂̄︣͟ Oct 19 '21 at 10:59
  • In the last example among your links, it is neatly proved (by Intelligenti pauca) (see the edit). Can you see that he says $c=a+b-2r$? – ACB Oct 19 '21 at 11:01
  • @ACB Yes, I have explicitly mentioned in the question that some of the links give a proof. Still, I considered posting question like this useful. 1. It is about this specific formula. 2. The answers collect various proofs. 3. If somebody is searching for a proof of this formula on MSE, the title "Is there a way to see this geometrically?" won't be the first guess for a post where to find something like that. – Martin Sleziak Oct 19 '21 at 11:40
  • Another way: we know area of right triangle is $ab/2$ and we know $A = r \cdot s$ where is $s$ is sub-perimeter. So $r (a+b+c) = ab$. Multiplying both sides by $(a+b-c)$, we get $r \cdot (2ab) = ab (a+b-c)$ – Math Lover Oct 19 '21 at 11:42
  • @MathLover Why not expanding that comment into an answer? – Martin Sleziak Oct 19 '21 at 11:53
  • @MartinSleziak I did not check the links you shared so wasn't sure whether this answer was already given earlier or not, did not want to duplicate – Math Lover Oct 19 '21 at 11:58
  • Another related post: https://math.stackexchange.com/q/2274268/947379 – ACB Feb 03 '22 at 09:43

9 Answers9

12

enter image description here

Let the tangent points be $A'$, $B'$ and $C'$ labelled in the usual way. Then, since tangents from a point to a circle have equal length, $CB'=r=CA'$.

Therefore, for the same reason, $$AB'=AC'=b-r$$ and $$BA'=BC'=a-r$$ And since $AC'+BC'=c$, we get $a-r+b-r=c$ and hence the result.

David Quinn
  • 34,121
  • 1
    +1. One can readily show that, for any triangle, the tangent segments from $A$, $B$, $C$ always have lengths $a'=(-a+b+c)/2$, $b'=(a-b+c)/2$, $c'=(a+b-c)/2$, respectively. (Just solve the system $a=b'+c'$, $b=c'+a'$, $c=a'+b'$.) When there's a right angle, the inradius is necessarily equal to the corresponding tangent segment. – Blue Oct 21 '21 at 14:37
  • After noticing that it is the same picture, I wondered whether the images was taken from Wolfram MathWorld article on Right Triangle. (Basically by accident I have noticed the same picture also in another post on this site.) – Martin Sleziak Oct 27 '21 at 08:59
  • 1
    @Martin Sleziak Probably - I just grabbed the first one I could find – David Quinn Oct 27 '21 at 09:16
  • As mentioned in my comment below the question, the image in this post is very comprehensive. – ACB Oct 27 '21 at 14:04
6

image

$$\large \Delta=\color{green}\blacktriangle+\color{red}\blacktriangle+\color{blue}\blacktriangle$$ $$\frac{ab}2=\frac{ar}2+\frac{br}2+\frac{cr}2$$ $$ab=(a+b+c)r$$ $$ab(a+b-c)=(a+b+c)(a+b-c)r$$ $$ab(a+b-c)=[(a+b)^2-c^2]r$$ $$ab(a+b-c)=(\underbrace{a^2+b^2-c^2}_{0}+2ab)r$$ $$ab(a+b-c)=2ab\cdot r$$ $$r=\frac{a+b-c}2$$

ACB
  • 3,713
4

Triangle square

By the squares method, $$r^2+r(a-r)+r(b-r)=r\,\dfrac{a+b+c}2,$$ $$2(a+b-r)=a+b+c,$$ $$\color{green}{\mathbf{r=\dfrac{a+b-c}2.}}$$

Then $$R+r=\dfrac c2+r=\dfrac{a+b}2\le\sqrt{ab\mathstrut}\,=\sqrt{2S}\,.$$

4

Here's an approach that I don't think I've seen before (but don't doubt exists in the literature):

enter image description here

Blue
  • 75,673
3

$$\begin{align*}r&=\frac{\Delta}s\\&=\sqrt{\frac{(s-a)(s-b)(s-c)}{s}} \\&=\sqrt{\frac{({a+b-c})({a-b+c})({-a+b+c})}{8s}} \\&=\sqrt{\frac{({a+b-c})[c^2-({a-b})^2]}{8s}} \\&=\sqrt{\frac{({a+b-c})(2ab)}{8s}} \\&=\sqrt{\frac{({a+b-c})}{2}.\frac{\Delta}{s}} \\&=\sqrt{\frac{({a+b-c})}{2}.r}\end{align*}$$ $$\implies r=\frac{a+b-c}2,$$

Asher2211
  • 3,406
  • 10
  • 31
3

Splitting the right-angled triangle into 3 triangles of bases $a$, $b$ and $c$ and the same height $r$(the in-radius)yields $$ \begin{aligned} \frac{r a}{2}+\frac{r b}{2}+\frac{r c}{2} &=\frac{a b}{2} \\ r(a+b+c) &=a b \\ r &=\frac{a b}{a+b+c} \cdot \frac{a+b-c}{a+b-c} \\ &=\frac{a b(a+b-c)}{a^{2}+b^{2}+2 a b-c^{2}} \\ &=\frac{a+b-c}{2}\quad \left(\because c^{2}=a^{2}+b^{2}\right) \end{aligned} $$

Lai
  • 20,421
1

We can consider the triangle formed by $(a, 0), (0, b), (0, 0)$ in the Cartesian Plane and proceed to find the point of intersection of the angular bisectors.

One angular bisector is clearly the line $y = x$.

For the other one, assume that the angle at (a, 0) to be $\theta$. The angular bisector is $y = -\tan \frac{\theta}{2}x+ a\tan \frac{\theta}{2}$.

The point of intersection these lines is $$ \left(\frac{a\tan\frac{\theta}{2}}{1 + \tan\frac{\theta}{2}}, \frac{a\tan\frac{\theta}{2}}{1 + \tan\frac{\theta}{2}}\right) $$ which means that the inradius is just $\frac{a\tan\frac{\theta}{2}}{1 + \tan\frac{\theta}{2}}$. Starting from $\tan\theta = \frac{b}{a}$, it is easy to show that $\tan\frac{\theta}{2} = \frac{c - a}{b}$.

$$ \begin{align} \frac{a\tan\frac{\theta}{2}}{1 + \tan\frac{\theta}{2}} &= \frac{a}{2}\tan\theta \left(1 - \tan \frac{\theta}{2}\right) \\ &= \frac{b}{2} \left(1 - \frac{c - a}{b}\right) \\ & = \frac{a+b-c}{2} \end{align} $$

Priyatham
  • 2,607
1

Labeling in Linked Wiki was incorrect. Vertex $C$ should have been more suitably labeled/placed at the vertex containing the right angle. Vertices A,B,C ordering should correctly be:

enter image description here

We have in-radius $r$

$$r=\frac{\Delta}s =\sqrt{\frac{s(s-a)(s-b)(s-c)}{s^2}}=\frac{a+b-c}{2}$$

on simplification using the cut & pasted image above

enter image description here

also depicts a correct relation for in-radius $r$ as can be verified by algebraic cross-multiplication leading to Pythagoras thm of the right triangle:

$$ (a+b)^2-c^2= 2ab \to c^2= a^2+b^2. $$

Narasimham
  • 40,495
1

\begin{align} r &=\frac{\Delta}{s} \\ &=\frac{\frac{ab}2}{\frac{a+b+c}2} \\ &=\frac{ab}{a+b+c} \\ &=\frac{\frac{(a+b)^2-c^2}{2}}{a+b+c}\\ &=\frac{(a+b-c)(a+b+c)}{2(a+b+c)}\\ \Rightarrow r=\frac{a+b-c}2 \end{align}

RiverX15
  • 888