5

enter image description here

Point $D$ is in $\triangle{ABC}$ such that $\angle{ABD}=\angle{CBD}=6^{\circ}$,$\angle{BCD}=12^{\circ}$,$\angle{ACD}=18^{\circ}$, $\color{red}{\angle{BAD}=?}$

Can we find a pure elementary geometric approach to answer this question?

D.W.
  • 6,040
  • Comments have been moved to chat; please do not continue the discussion here. Before posting a comment below this one, please review the purposes of comments. Comments that do not request clarification or suggest improvements usually belong as an answer, on [meta], or in [chat]. Comments continuing discussion may be removed. – Xander Henderson May 24 '24 at 00:35

3 Answers3

3

The following approach should work generally to identify solutions in $6n^\circ$ adventitious triangles. Points $P,Q$ mark the triangle in question and the inner intersection in question. Within the larger context of $6n^\circ$ triangles, these are vertices of a rhombus marked with the other two pink dots (since the diagonals are perpendicular), hence the opposing angles are equal, and bisected by the diagonals, and we instantly know the relevant angle: $42^\circ$.

enter image description here

Proof: The image below shows the lattice again, this time marking all points that lie on $6n^\circ$ rays from all $3$ vertices. $P$ and $Q$ will lie on such point. Once this fact is established, the proof above follows.

enter image description here

I) By an argument first of mirror symmetry, then by $120$-degree radial symmetry, all filled gray points are on rays from each vertex measuring $6n^\circ$.

II) Points $Q$ and $R$ are also on $3$ rays, and again by mirror symmetry then radial symmetry these points populate the lattice as the hollow diamonds.

Diamond proof: That rays from each corner meet in the diamond vertices can be first guessed and then confirmed in the diagram. For $Q$ the image below shows how to construct on rays the two triangles that, when rotated (upper black clockwise $120^\circ$, pink counter-clockwise $120^\circ$), will perfectly match the triangle on the bottom to complete the equilateral (corners sum to $60^\circ$, equilateral sides, central angles sum to $360^\circ$). We can confirm that these triangles meet in a point by using the law of sines and Ceva's theorem, which yield $\frac{\sin{6}\sin{42}\sin{42}}{\sin{12}\sin{18}\sin{54}}=1$. This is the only possible pair of triangles that can meet at one point to complete the equilateral (at $Q$). The argument for $R$ is similar.

enter image description here

Alternate diamond proof: we can also follow synthetic proofs to the same conclusion, such as the one the OP added for $Q$ and this one for $R$: With initial triangle $ARB$ subtending $6$ and $18$ degrees, form the darts $AB_{12}RB$ and $AB_{12}SB$, note the formation of the $4$ isosceles triangles within the darts. We also have $CS$ at right angles to the base, so $\angle{CSR}=90-6=84$. But $AB=AB_{12}=AC$ so we have isosceles $\triangle{AB_{12}C}$ and $\angle{CB_{12}A}=66^\circ$, which, when combined with the angle in the large dart, also yields $84^\circ$. Thus we have the kite $CSRB_{12}$ and diagonal $CR$ bisecting $36^\circ$ at $C$, so $\angle{RCB_{12}}=18^\circ$ and within that $\angle{BCR}=12^\circ$.

enter image description here

The lattice above provides natural paths to solutions for virtually all $6n$ adventitious triangles. The solution added by the OP essentially follows the rhombus path, establishing the $3$-ray convergence for point $Q$ as as lemma, then using a kite instead of a rhombus, and the overall diagram makes it clear why this is a useful route toward a solution. Fun fact: Bottoma's triangle, a counterexample to "equal external angle bisectors $\Rightarrow$ isosceles" is hiding in the lattice as $\triangle{ABM}$ in the second diagram.

RobinSparrow
  • 3,984
  • I'm not convinced by the first "diamond proof" paragraph yet, which is based on the necessary condition that the pink $(54^\circ,42^\circ,84^\circ)$-triangle and the other $(48^\circ, 18^\circ, 114^\circ)$-triangle, once rotated, give matching angle sums ($42^\circ + 18^\circ = 60^\circ$; $84^\circ+114^\circ +162^\circ = 360^\circ$). But this is not sufficient to prove that these are indeed the correct pair of triangles to meet at $Q$. Why are they not, say, pink with $(54^\circ,40^\circ,86^\circ)$ and black with $(48^\circ, 20^\circ, 112^\circ)$? [TBC] – peterwhy May 21 '24 at 18:52
  • I can verify numerically by the sine rule that the angle numbers work for $Q$:

    $$QC = \frac{b\sin 84^\circ}{\sin54^\circ} = \frac{a\sin 114^\circ}{\sin48^\circ},$$

    (therefore not other angles), but the "diamond proof" paragraph does not prove it, and does not extend generally to show the angles $CR$ makes.

    – peterwhy May 21 '24 at 18:53
  • 1
    I am aware of the necessary conditions at corners $A$ and $B$ of equilateral $\triangle ABC$, so I gave an (incorrect) alternative hypothesis for $Q$, that pink $\triangle ACQ$ has angles $(A=54^\circ, C=40^\circ, Q=86^\circ)$ and black $\triangle BCQ$ has angles $(B=48^\circ, C=20^\circ, Q=112^\circ)$ after rotated. A proof should identify this as wrong, but the first "diamond proof" paragraph "(corners sum to $60^\circ$, equilateral sides [presumably $\triangle ABC$], central angles sum to $360^\circ$)" does not seem sufficient. – peterwhy May 21 '24 at 20:34
  • I don't have a clean fix. I agree that these adventitious angles geometry questions are hard and I can't solve this by my own probably. Thanks for your investigations though! – peterwhy May 21 '24 at 22:27
  • 6-12-18 are the only intersections of three lines inside an equilateral triangle for degrees being multiples of 6. There are two kinds of structures and six situations for only one unknown angle. They can all be induced using pure elementary geometric approach. I have got all these solved by now. – Adventitious Angles Qs Poster May 26 '24 at 16:30
1

Here is the solution in pure elementary geometric approach:

\begin{multline} \shoveleft \text{First we need solve a related problem: }\\ \shoveleft \text{Let }D \text{ be a point in equilateral }\triangle{ABC} \text{ such that }\angle{BCD}=12^{\circ}, \angle{CBD}=6^{\circ}, \color{red}{\angle{CAD}=?} \end{multline}

enter image description here

This can be solved:

enter image description here

\begin{multline} \shoveleft \text{Make }E \text{ be the reflect of }C \text{ along the line }BD \implies \\ \shoveleft CD=DE,BC=BE, \angle{BCD}=\angle{BED}=6^{\circ}, \angle{BCD}=\angle{BED}=12^{\circ}\\ \shoveleft \implies \angle{BAE}=\angle{BEA}=\dfrac{180^{\circ}-\angle{ABE}}{2}=\dfrac{180^{\circ}-48^{\circ}}{2}=66^{\circ}\implies \angle{CAE}=6^{\circ}\\ \shoveleft \text{Extend }CD \text{ and intersect }BE \text{ at }F \implies \angle{BCF}=\angle{CBF}=12^{\circ}\implies \\ \shoveleft \begin{cases} BF=CF\implies \triangle{ABF}\cong\triangle{ACF}\implies \angle{CAF}=30^{\circ}\implies \angle{EAF}=36^{\circ}\\ \angle{CFE}=24^{\circ} \implies \angle{CDE}=36^{\circ}=\angle{EAF}\\ \end{cases}\\ \shoveleft \implies AFDE \text{ is cyclic }\implies \angle{DAE}=\angle{DFE}=24^{\circ}\implies \angle{DAC}=\bbox[5px, border: 1px solid black]{18^{\circ}}\\ \shoveleft \\ \end{multline}

Now we can come back to original problem:

enter image description here

\begin{multline} \shoveleft \text{Make }\triangle{EBC} \text{ be an equilateral triangle}\\ \shoveleft \implies \angle{DEC}=18^{\circ} \implies \angle{BED}=42^{\circ}\\ \shoveleft \text{And we know }\angle{ABE}=48^{\circ} \implies BA\perp DE, \text{ suppose }BA\cap DE=G \\ \shoveleft \text{Extend }CA \text{ to }F\text{ on }BE, \angle{BAF}=30^{\circ} \implies CF \perp BE, BF=FE\\ \shoveleft \text{Make }H \text{ on }AF \text{ such that }\angle{ABH}=6^{\circ} \implies \angle{ABH}=42^{\circ}\\ \shoveleft \triangle{BFH}\cong\triangle{EFH}\implies BH=EH\implies \angle{FEH}=42^{\circ}=\angle{BED}\\ \shoveleft \implies EHGD \text{ are collinear} \implies BD=BH \implies \triangle{ABD}\cong\triangle{ABH}\\ \shoveleft \implies \angle{BAD}=\angle{BAH}=90^{\circ}-\angle{AHG}=90^{\circ}-\angle{EHF}=\angle{FEH}=\bbox[5px, border: 1px solid black]{42^{\circ}}\\ \end{multline}

0

Here is a trigonometric solution, it leads to a different equation that you mentioned, but it is still quite difficult, using law of sines in triangles $ABC$, $ABD$ and $BCD$ we have \begin{align*} &\frac{AB}{\sin{30}} = \frac{BC}{\sin{42}} \Longrightarrow AB = \frac{BC}{2\sin{42}}\\ &\frac{BD}{\sin{x}} = \frac{AB}{\sin{(174-x)}} \Longrightarrow BD = \frac{AB\sin{x}}{\sin{(x+6)}} = \frac{BC\sin{x}}{2\sin{42}\sin{(x+6)}} \\ &\frac{BC}{\sin{18}} = \frac{BD}{\sin{12}} \Longrightarrow BC = \frac{BC\sin{x}\sin{18}}{2\sin{12}\sin{42}\sin{(x+6)}} \end{align*} which leads us to $$ \frac{\sin{(x+6)}}{\sin{x}} = \frac{\sin{18}}{2\sin{12}\sin{42}} $$ i.e. \begin{equation} \cos{6}+\cot{x}\sin{6} = \frac{\sin{18}}{2\sin{12}\sin{42}} \tag{*} \end{equation} so we have \begin{align*} \cot{x}\sin{6} &= \frac{\sin{18}}{2\sin{12}\sin{42}} - \cos{6} = \frac{\sin{18}-2\sin{12}\cos{6}\sin{42}}{2\sin{12}\sin{42}} \\ &= \frac{\sin{18}-\left(\sin{18}+\sin{6}\right)\sin{42}}{2\sin{12}\sin{42}} = \frac{\sin{18}-\sin{18}\sin{42}-\sin{6}\sin{42}}{2\sin{12}\sin{42}} \\ &=\frac{\sin{18}-\frac{1}{2}\left(\cos{24}-\cos{60}\right)-\frac{1}{2}\left(\cos{36}-\cos{48}\right)}{2\sin{12}\sin{42}} \\ &=\frac{2\sin{18}+\cos{60}-\cos{36}-\cos{24}+\cos{48}}{4\sin{12}\sin{42}} \\ &=\frac{\frac{\sqrt{5}-1}{2}+\frac{1}{2}-\frac{(\sqrt{5}+1)}{4}-\cos{24}+\cos{48}}{4\sin{12}\sin{42}} = \frac{\frac{\sqrt{5}-1}{4}-\cos{24}+\cos{48}}{4\sin{12}\sin{42}} \\ &= \frac{\cos{72}+\cos{48}-\cos{24}}{4\sin{12}\sin{42}} = \frac{2\cos{60}\cos{12}-\cos{24}}{4\sin{12}\sin{42}} = \frac{\cos{12}-\cos{24}}{4\sin{12}\sin{42}} \\ &= \frac{2\sin{18}\sin{6}}{4\sin{12}\sin{42}} = \frac{\sin{18}}{2\sin{12}\sin{42}}\sin{6}, \end{align*} now, using (*) again we have \begin{align*} &\cot{x}\sin{6} = \left(\cos{6}+\cot{x}\sin{6}\right)\sin{6} \\ &\cot{x} = \cos{6}+\cot{x}\sin{6} = \sin{84}+\cot{x}\cos{84} \\ &(1-\cos{84})\cot{x} = \sin{84} \\ &2\sin^{2}{42}\cot{x} = 2\sin{42}\cos{42} \\ &\cot{x} = \cot{42}, \end{align*} hence $x=42$.