Let $ABC$ be a scalene triangle. $I$ is incenter. Common point of inscribed circle and $BC$ is $E$. $AF$ is angle bisector. If circumcircles of $ABC$ and $AEF$ meet at $A$ and $D$, then prove $∡ADI=90°$. MY TRY: I chased angle. And got little value of result. I have found out $∡CDE=∡EDB=∡BAF=∡FAC$. And other than this I can only tell $ABCD$ and $ADEF$ are circumcircle.

enter image description here

  • 1
    $\begingroup$ Where exactly is the point $F$ on the angle bisector of $\angle B$ $\endgroup$ – Anas A. Ibrahim May 27 at 6:19
  • $\begingroup$ It is AF. I mean AF is angle bisector my bad sorry. F is at BC $\endgroup$ – Ualibek Nurgulan May 27 at 7:56

Something is definitely wrong with the question.

Checking out:

  1. $\triangle ABC$ be a scalene triangle.

  2. $I$ is incenter.

  3. Common point of inscribed circle and $BC$ is $E$.

  4. $BF$ is angle bisector.

  5. Circumcircles of $\triangle ABC$ and $\triangle AEF$ meet at $A$ and $D$.

The image:

enter image description here

clearly illustrate that $\angle ADI$ is not anywhere near $90^\circ$.

| cite | improve this answer | |
  • $\begingroup$ oh sorry my bad AF is angle bisector $\endgroup$ – Ualibek Nurgulan May 27 at 7:53
  • $\begingroup$ Hello how did you draw the picture. What kinf of application di you use $\endgroup$ – Ualibek Nurgulan May 27 at 8:03
  • $\begingroup$ @Ualibek Nurgulan: Yes, with $AF$ as a bisector it looks OK. I added a corrected image to he question. $\endgroup$ – g.kov May 27 at 8:14
  • $\begingroup$ @Ualibek Nurgulan:the picture was prepared using Asymptote, "Asymptote is a powerful descriptive vector graphics language that provides a natural coordinate-based framework for technical drawing. Labels and equations are typeset with LaTeX, for high-quality PostScript output." Also, you can check out this forum. $\endgroup$ – g.kov May 27 at 8:17

Denote by $k_0$ the circumscribed circle of triangle $ABC$ and by $k_1$ the circumcircle of triangle $AEF$, where by assumption $k_0 \cap k_1 = \{A, D\}$. Extend the angle bisector $AF$ until it intersects the circumcircle $k_0$ of $ABC$ into the second point $L$ on $k_0$, the first being $A$. Then $L$ is the midpoint of the arc of $k_0$ between the points $B$ and $C$ that does not contain point $A$, because $AL$ is the angle bisector of angle $\angle \, BAC$. Therefore $LB = LC$. After some very simple angle chasing, one can show that $\angle\, LBI = \angle\, LIB$, which means that the triangle $BLI$ is isosceles with $LB = LI$. Thus we have that $$LB = LC = LI$$ Construct the circle $\omega$ with center $L$ and radius $LB$. Then the three points $B, \, C,\, I$ lie on $\omega$.

If you perform inversion with respect to $\omega$, the circle $k_0$ is mapped to the line $BC$ and in particular the point $A$ is mapped to the point $F$. However, both points $A$ and $F$ lie on circle $k_1$, which means that circle $k_1$ is mapped to itself under the inversion in $\omega$ (and is in fact orthogonal to $\omega$). Since $k_0 \cap k_1 = \{A, D\}$ their image under the inversion with respect to $\omega $ is $BC \cap k_1 = \{F, E\}$ which means that the point $E$ is mapped to the point $D$ under the inversion and the points $D, E$ and $L$ are collinear.

Now, consider circle $k_2$ circumscribed around triangle $EFI$. Since $IE \, \perp \, BC$ we see that $\angle\, IEF = 90^{\circ}$, which means that the center $O_2$ of $k_2$ is the midpoint of segment $IF$ so $O_2$ lies on the angle bisector $AL$ and thus the points $L, \, O_2,\, I$ are collinear. Hence circle $k_2$ is tangent to circle $\omega$ at point $I$. Under inversion in $\omega$, the circle $k_2$ is mapped to the circle $k_3$ passing through the image points $I, \, A, \, D$ of points $I, \, F, \, E$ respectively, and $k_3$ is also tangent to $\omega$ at point $I$. Hence the center of $k_3$ must be collinear with the the centers $L$ and $O_2$ of $\omega$ and $k_2$, which lie on the angle bisector $AL$, so the center of $k_3$ also lies on $AL$ and therefore the center of $k_3$ lies on the segment $AI$. The latter fact however means that $AI$ is a diameter of $k_3$. Since, as already established, $D$ lies on $k_3$, angle $\angle \, ADI = 90^{\circ}$.

| cite | improve this answer | |

The problem in the OP can be stated equivalently in the form: The circles $\odot(ABC)$, $\odot(AEF)$, and the circle with diameter $AI$ have a common chord. (Which is $AD$ in the OP.) The first and the third in the list are "simpler" (for my taste, they depend on "simpler points"), so let $D'$ be their intersection, let us try to show that the third circle, $\odot(AEF)$ also passes through $D'$. (So $D=D'$ in the final.)

As it often happens in problems involving essentially the centers $O,I$ (of the circumcircle and the incircle) of a triangle, the following constellation of points is useful:

wiki page on Euler's formula

The following solution is based on the point $L$ from the above link (and the solution by inversion of Futurologist), and on the projection $Z$ of $I$ on the $A$-height. The idea of the following solution is to show that $L,E,Z,D'$ are collinear.

In the following picture, let $AH$ be the height in $A$, $H\in BC$, let $X,Y,Z$ be the projections of $I$ on $AB$, $BC$, $AH$. (So $EYX$ is the incircle.)

Let $S$ be the mid point of $AI$. Let $\odot(S)$ be the circle centered in $S$ with diameter $AI$.

Let $D'\ne A$ be the second intersection of the circles $\odot(ABC)=\odot(O)$ and $\odot(AXZIY)=\odot(S)$.

math stackexchange problem 3693462

We have: $$ \begin{aligned} \widehat{LD'A} &= \widehat{LBA} = \widehat{LBC} + \widehat{CBA} = \frac 12 \hat A+\hat B\ , \\ \widehat{ZIA} &= \widehat{BFA} = \frac 12\overset{\frown}{AB} + \frac 12\overset{\frown}{LC} =\hat C+\frac 12 A\ , \\ \widehat{ZD'A} &= 180^\circ -\widehat{ZIA} =180^\circ -\left(\hat C+\frac 12 A\right) =\frac 12 \hat A+\hat B =\widehat{LD'A}\ . \end{aligned} $$ So $L,Z,D'$ are on the same line. Let us show now that $E$ is also on this line. For this, we compute two proportions, this seems to be the quick+dirty path: $$ \begin{aligned} \frac{IE}{AZ} &= \frac{ZH}{AZ} = \frac{FI}{IA} = \frac{BF}{BA} = \frac{ac/(b+c)}c = \frac a{b+c} \ ,\\[2mm] \frac{LI}{IA} &= \frac{LB}{LA} = \frac{\sin\widehat{BAL}}{\sin\widehat{ABL}} = \frac{\sin(\hat A/2)}{\sin(\hat A/2+\hat B)} = \frac{2\sin(\hat A/2)\cos(\hat A/2)}{2\sin(\hat A/2+\hat B)\cos(\hat A/2)} \\ &= \frac{\sin\hat A}{\sin\hat B+\sin\hat C} =\frac a{b+c}\ . \end{aligned} $$ So $\Delta LIE\sim\Delta LAZ$ (since the above proportions are equal, and the angles in $I$ and $A$ are correspondent), so they have the same angle in $L$, so $L,E,Z$ colinear.

We can now conclude. The quadrilaterals $ZIAD'$ and $EFAD'$ have the same angles, so the second one is also cyclic, as the first one. This means that $AD'$ is also a chord in $\odot(AD'XZIY)$. (So $D=D'$.)


Bonus: With the notations from the picture, $S$ is on $MN$. The circle $\odot(S)$, and the lines $CI$, $EX$ intersect in a point. The circle $\odot(S)$, and the lines $BI$, $EY$ intersect in a point.

| cite | improve this answer | |

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.