8

$D$ is a point in $\triangle{ABC}$ so that $\angle{ABD}=10^{\circ}$, $\angle{DBC}=20^{\circ}$, $\angle{BCD}=40^{\circ}$, $\angle{DAC}=50^{\circ}$.

Find $\angle{BAD}$.

This problem is easily done with trigonometric Ceva theorem as:

$$ \begin{aligned} &\dfrac{\sin x}{\sin50^{\circ}} \cdot \dfrac{\sin(60^\circ-x)}{\sin40^{\circ}} \cdot \dfrac{\sin20^\circ}{\sin10^\circ}=1 \\ \implies & \sin x \cdot \sin(60^\circ-x) = \dfrac {\sin50^\circ\cdot \sin40^\circ\cdot \sin10^\circ} {\sin20^\circ} \\ &\qquad = \dfrac { \cos40^\circ\cdot \sin40^\circ \cdot \sin10^\circ} {2\sin10^\circ\cdot \cos10^\circ} = \dfrac{\sin80^\circ}{4\cos10^\circ} =\dfrac{1}{4} \\ \implies & -\dfrac{\cos60^\circ - \cos(2x-60^\circ)}2 = \dfrac{1}{4} \\ \implies & \cos(2x-60^\circ)=1 \\ \implies &2x-60^\circ=0 \\ \implies &x=\boxed{30^\circ}\ . \end{aligned} $$

Any idea on how to solve this problem in pure geometric approach? Thanks.

Problem Image

r ne
  • 794
  • 3
  • 9
  • The result is based on property $\tan 20° \cdot \tan 30° = \tan 10° \cdot \tan 50°$, which can be easily shown by trigonometric manipulation but is hard to show by pure geometry. – Ivan Kaznacheyeu Apr 11 '22 at 14:37
  • This is a nice problem, everybody seems to have her or his own fun with it, it is a good problem for contests (below the level of having Ceva and trigonometry at hand). – dan_fulea Apr 17 '22 at 00:36

4 Answers4

3

Solution Pic

I just got this solution (no phantom point and few verbal words):

Let $E$ be the circumcenter of $\triangle{ACD}$, so $AE=CE=DE, \angle{DEC}=2\angle{DAC}=100^{\circ}$ $\implies \angle{EDC}=\angle{ECD}=40^{\circ}=\angle{BCD} \implies DE \parallel BC$

Extend $ED$ and intersects with $AB$ at $F$, so $\angle{AFE}=\angle{ABC}=30^{\circ}$. Make $G$ on $BC$ so that $\angle{EFG}=100^{\circ}=\angle{FEC} \implies CEFG$ is an isosceles trapezoid $\implies FG=CE$

Let $H$ be the circumcenter of $\triangle{BFG}$, so $BH=GH=FH, \angle{FHG}=2\angle{FBG}=60^{\circ} \\ \implies \triangle{FGH} \text{ is an equilateral triangle}\\ \implies FH=GH=BH=FG=CE=DE=AE, \angle{HFG}=60^{\circ} \\ \implies \angle{HBF}=\angle{HFB}=\angle{HFG}-\angle{BFG}=\angle{HFG}-(180^{\circ}-\angle{FBG}-\angle{FGB}) \\ =\angle{HFG}-(180^{\circ}-\angle{FBG}-\angle{GFE}) =60^{\circ}-(180^{\circ}-100^{\circ}-30^{\circ})=10^{\circ} \\ \implies FH \parallel BD, \angle{HBD}=20^{\circ}=\angle{DBC}=\angle{FDB} \\ \implies BDFH \text{ is an isosceles trapezoid} \\ \implies BH=DF \implies DF=DE=AE$

Check $\triangle{AEF}, \angle{AFE}=30^{\circ}, AE=DE=DF \implies EF=2AE$.

Let $O$ be the circumcenter of $\triangle{AEF}$, so $\angle{AOE}=2 \angle{AFE}=60^{\circ}, AO=FO=EO$ $\implies \triangle{AOE} \text{ is equilateral triangle}\\ \implies AO=OE=AE=DE=DF=FO=EO \\ \implies AD=AO=FD \\ \implies \angle{BAD}=\angle{AFE}=\boxed{30^{\circ}}$

r ne
  • 794
  • 3
  • 9
  • Good solution. Need of coinciding point is excluded by using more advanced theory (sine rule). This problem can be solved before learning this theory. – Ivan Kaznacheyeu Apr 13 '22 at 10:51
  • @MathLover It's very natural right, if you could give an alternative approach for the last step, it's also good. – r ne Apr 14 '22 at 04:28
  • 1
    $\triangle EAF$ is a right triangle given $EF = 2 AE$ and $\angle AFE = 30^0$ – Math Lover Apr 14 '22 at 04:28
  • If $O$ is the circumcenter of the triangle, $AO = OE$ and $\angle AOE = 60^0$. So $FO = OE = AE$. As $EF = 2 AE$, $O$ must be on $AE$ and its midpoint. – Math Lover Apr 14 '22 at 04:33
  • @MathLover I made new edit and remove Law of Sines. – r ne Apr 14 '22 at 04:38
  • @IvanKaznacheyeu Yes this part can be saved. Right now it's more elementary. – r ne Apr 14 '22 at 04:39
  • What if $E$ is on side $AC$, why is that not a possibility? – Vasili Apr 14 '22 at 23:36
  • @Vasili It could be, right? The process does not specify that $E$ is NOT on $AC$. It's the result that shows $E$ is not on $AC$ but it does not matter, right? – r ne Apr 15 '22 at 08:46
  • @rne: It it was, the angle $C$ would be $80$ degrees which leads to unknown angle being $20$, a contradiction. – Vasili Apr 15 '22 at 17:57
  • @Vasili: then you have answered your question above: it is not possible... another proof is: if $E$ is on $AC$, $\angle{DAC}=50^{\circ}=\angle{DAE}=60^{\circ}$. – r ne Apr 15 '22 at 19:35
  • @Vasili: It's like: I solve the equation 2x=1: divide by 2 on each side gives x=0.5. Then you say: what if x=1, why is that not a possibility? I said it does not matter, you just divide by 2 on each side... You said: it matters, if x=1, then 2x=2. You are right. So you have proved x cannot be 1... – r ne Apr 15 '22 at 19:42
  • @rne: This is different, you based your proof on the assumption that point $E$ is located outside of the triangle. You need to show that this is the only possible position. Don't take me wrong, I like the proof but as a mathematicians, we have to doubt and apply critical thinking, – Vasili Apr 15 '22 at 20:01
  • @Vasili You need show that the proof process is based on the assumption of location of point $E$ is outside of the triangle. How did you come to this statement. To apply critical thinking, you need provide reasoning for your statement first. – r ne Apr 17 '22 at 19:01
2

I'll give a first quick solution, one can stop after it if the problem implicitly gives the information that the configuration to be realized exists and is unique. Some discussion follows to cover also this point and have a complete solution, that cannot lose points e.g. in the conditions of an olympiad.


It is natural to try to realize the given picture using vertices $A,B,C$ among those of a regular poligon. It turns out, we need a regular $18$-gon, in the picture the vertices have the labels from $\Bbb Z$ modulo $18$, and for easy using as a point there is a label $k'$ instead of $-k$ for $-k$ from $-8=10$ to $-1=17$. We place the letters $A,B,C$ at positions $0,7,3'$, and show that the point $D$ defined as $3'3\cap 71'$ (so that $\Delta DBC$ has angles of $20^\circ$, $40^\circ$ in $B,C$) lies on the chord from $A=0$ making the $50^\circ$ angle with $AC$, which is $08'$.

math stackexchange 4423750 regular 18-gon

In other words, we have to show:

Proposition $(P)$: In the given regular $18$-gon the lines $08'$ and $71'$ and $3'3$ intersect in a point $D$.

Proof: Let $I$ be the incenter of $\Delta 173'=\Delta 1BC$. So $I=17'\cap 71'\cap 3'4$, and since $11'7'7$ is an isosceles trapezium, it is also its symmetry axis $09$. Now observe that $D$ is the orthocenter of $\Delta IAC=\Delta I03'$, the heights are the three lines to be shown concurrent: $$ D= \underbrace{\ 7I1'\ }_{\perp\text{ on } 03'}\cap \underbrace{\ 3'3\ }_{\perp\text{ on } 9I0}\cap \underbrace{\ 08\ }_{\perp\text{ on } 3'I4}\ . $$

$\square$

The angle $\widehat {BAD}$ "sees an arc covering three sides", so it is $\color{blue}{\boxed{30^\circ}}$.



We are done, but let's turn the above argument into a constructive solution not assuming that the given constellation exists and is unique. We make the Proposition $(P)$ work in this context.

So we know by it that $08'\cap 71'\cap 3'3$ is a point, call it $D$.

We restart with what we certainly have, this is the triangle $\Delta BCD$, it is fully known (modulo similarity), we are drawing it, so that $B,C$ are placed in $7,3'$.

Based on it, we have to construct a point $A$ such that $\widehat{DAC}=50^\circ$, a point located in the half-plane w.r.t. $CD$ not containing $B$. The locus of such a point is an arc of a circle, call this circle $\omega$, its center $\Omega$. We know that the vertex $0$ is on this circle, so $\omega=(D0C)$. I need the picture (without any point $A$ on it):

math stackexchange 4423750 second picture regular polygon 18-gon

Now consider a moving ray, in the radar with origin in the point $B$. It intersects $\omega$ in either two points, or in one point, or in no point. We fix this ray to make an angle of $10^\circ$ with $BD$, as given. This ray intersects $(\omega)$ in at least one point, $0$, as guaranteed by $(P)$. Does $B0=70$ intersect $\omega$ also in an other point, or it is tangent to it?! We show:

Lemma: In the above situation, $B0$ is tangent to $\omega$ in $0$.

(So $A=0$ is the unique solution to our problem.)

Proof of the Lemma: $\Delta D0C$ has angles of $30^\circ$, $50^\circ$ in $C$, resp. $0$. So $\widehat {0\Omega C}=2(50^\circ+30^\circ)=160^\circ$, which results in angles of $10^\circ$ in $0,C$ in $\Delta 0\Omega C$. This gives: $$ \widehat{B0\Omega}= \widehat{B0D}+ \widehat{D0C}+ \widehat{C0\Omega} = 30^\circ + 50^\circ + 10^\circ = 90^\circ \ . $$ $\square$

dan_fulea
  • 32,856
1

enter image description here

We use this fact that the mirrors of circumcircle about the sides of triangle meet at it's orthocenter.

Step 1:Draw circumcircle(1) of ABC center is O.Draw altitude AK.Mirror AB, AK, circle (1) about base BC, circle (2) is mirror of circle (1). Altitude AK intersects the circle 2 at E. E is orthicenter of triangle ABC. Connect D to E and extend it to touch AC at H. We show that DH is collinear with BD and in this way BH is altitude of triangle,Clearly $\angle BA'A=60^o$

In circle (2):

$\overset{\large\frown}{BE}=120^o$

$\overset{\large\frown}{EC}=40^o$

$\Rightarrow \overset{\large\frown}{BA'C}=360-(120+40)=200^o$

$\overset{\large\frown}{A'C}=60^o$

$\Rightarrow \overset{\large\frown}{BA'}=140^o$

That is if we connect B to E we have:

$\angle BEA'=\frac {140}2=70^o$

This ls competent with situation in right angled triangle BEK where $\angle BEK =180-(90+20)=70^o$. This means that BD and BE are collinear and BH is an altitude of triangle.

Step 2:Draw circumcircle (3) of triangle BCH, BC will be it's diameter. Extend CD to meet the circle (3) at point F.We have:

$\overset{\large\frown}{CH}=40^o$

$\overset{\large\frown}{BF}=80^o$

$\Rightarrow$$\overset{\large\frown}{FH}=60^o$

$\Rightarrow \angle FCH=30^o$

which finally gives $\angle BAD=30^o$

Update: All triangles has this property that the mirror of their circumcircle about it's sides meet at orthocenters of the them. See following figure:

enter image description here

sirous
  • 10,751
  • Why do we have $$O'E=O'A'\ ?$$ The point $E$ was constructed as $BD\cap AK$. That proposition used as first argument in the answer, "We show that circle (2) passes E (...) . Since circle (2) also passes B and C, ..." Where is shown that $E\in(2)$?! What does it mean "also passes", the points $B,C$ are so far the only points (easily) known to be on that circle $(2)$. Also, the used point $H$ was not defined, it seems that this very start already assumes $BD\perp AC$, which makes all the rest of the story obsolete, knowing this perpendicularity makes $\angle DCA$ known, so everything. – dan_fulea Apr 17 '22 at 00:52
1

Let's make following construction. $E$ is point of $BC$ such that $BE=BD$, $F$ is point symmetric to $E$ around $CD$. $G$ is point symmetric to $D$ around $AB$. Angular measures which can be restored from problem statement are given in picture.

$\angle DFC=100$°$=2\angle DAC$, then $F$ is center of circumcircle of triangle $ACD$. Then $FA=FD=FC=DE=CE$. Triangle $BDG$ is congruent to triangle $BDE$, then $DG=DE$.

$AF=FD$, $AD=AG$. Then $A$ is common point of perpendicular bisector of $DG$ and circle with center $F$ and radius $FD$.

Let point $H$ is symmetric to $D$ around $FG$, then $FDGH$ is rhombus. Because of $\angle FDG=120$° in this rhombus $DH=GH$ and $H$ is on perpendicular bisector of $DG$ which is $BA$. Then $\angle BHD=30$°, $\angle BHF=90°$ and $FH=FD$. Then perpendicular bisector of $DG$ is tangent to circle with center $F$ and radius $FD$. As tangent has only one common point with circle, then $A=H$ and $\angle BAD=\angle BHD=30$°.

  • Thanks for the reply. First of all, why $FDGH$ is rhombus? All we know is $FD=DG, DH=GH, \angle{FDG}=120^{\circ}$. Second, why $\angle{BHD}=30^{\circ}, \angle{BHF}=90^{\circ}, FH=FD$? There should be some reasoning here. I am not convinced by this brief statement. – r ne Apr 11 '22 at 22:44
  • If $FD=DG$ and $\angle FDG=120$°, then we always can construct point $H$ such, that $FDGH$ is rhombus. If rhombus has angle 120°, then it consists of two equilateral triangles, so triangles $DGH$ and $FDH$ are equilateral. 3. Then perpendicular bisector of $DG$ is angle bisector of $DGH$. Then $\angle BHD=30$°. $\angle BHF=\angle BHD+\angle DHF=90$°. – Ivan Kaznacheyeu Apr 12 '22 at 09:46
  • Yes, we can always construct point $H$ such, that $FDGH$ is rhombus, but you still need prove $H$ point lies on the perpendicular bisector of $DG$ which is $BA$, right? What if $H$ is not on $BA$? Why it has to be on $BA$? – r ne Apr 12 '22 at 10:08
  • To my experience, for pure geometry approach of calculation/proof, surely you need to say something like "extend line AB to D so that..." or "Apply the XXX theorem so that...", this is hard to be replaced with pure mathematical expression, but the deduction part should be done without any verbal words. The more words used in deduction part, usually it means there are weakness inside it. – r ne Apr 12 '22 at 10:21
  • 1
    According to definition of $G$, $BA$ is perpendicular bisector of $DG$. In the answer: "$H$ is on perpendicular bisector of $DG$ which is $BA$". In my comment I gave explanation, why $H$ is on perpendicular bisector of $DG$ (triangle $DGH$ is equilateral). I believe my answer is reasonably short, one can ask me for unclear passages. – Ivan Kaznacheyeu Apr 12 '22 at 10:44
  • Thanks. This phantom point approach may work. I have just got a clearer approach as following. – r ne Apr 13 '22 at 03:20
  • @IvanKaznacheyeu: This works if $F$ is to the right of $AC$ but we can't assume that. It can be on side $BC$, for example or above $AD$. – Vasili Apr 15 '22 at 22:06