Iran TST 2009-Day3-P3

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

239 posts

#1 h May 16, 2009, 5:22 PM • 15 Y

Y by DrMath, anantmudgal09, Catoptrics, itslumi, 554183, pog, MehmetBurak, Newmaths, Lcz, Adventure10, Mango247, Rounak_iitr, Funcshun840, X.Luser, and 1 other user

In triangle $ABC$ , $D$ , $E$ and $F$ are the points of tangency of incircle with the center of $I$ to $BC$ , $CA$ and $AB$ respectively. Let $M$ be the foot of the perpendicular from $D$ to $EF$ . $P$ is on $DM$ such that $DP = MP$ . If $H$ is the orthocenter of $BIC$ , prove that $PH$ bisects $EF$ .

This post has been edited 1 time. Last edited by khashi70, May 16, 2009, 6:07 PM

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

mathVNpro

469 posts

mathVNpro

#2 h May 16, 2009, 5:42 PM • 1 Y

Y by Adventure10

khashi70 wrote:

In triangle $ABC$ , $D$ , $E$ and $F$ are the points of tangency of incircle with the center of $I$ to $BC$ , $CA$ and $AB$ respectively . Let $M$ be the feet of perpendicular from $D$ to $EF$ and $P$ is on $DM$ such that : $DP = MP$ . If $H$ be the orthocenter of $ABC$ , prove that $PH$ bisects $EF$ .

Maybe you have made a typo, check it again.

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

khashi70

239 posts

khashi70

#3 h May 16, 2009, 6:08 PM • 2 Y

Y by Adventure10, Mango247

Sorry you're right ... it's OK now .

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

Ahiles

374 posts

Ahiles

#4 h May 16, 2009, 6:52 PM • 3 Y

Y by K.N, Adventure10, and 1 other user

Let $CI$ cut $BH$ at $G$ and $BI$ cut $EF$ ate $Q$ . Then

$\angle GIB = 180^\circ - \dfrac{\angle B + \angle C}{2} = 90^\circ - \dfrac{A}{2} = \angle EFA$ .

Thus, the quadrilateral $BFGI$ is cyclic and

$\angle{BGI} = \angle BFI = 90^\circ$ .

So, points $B,G$ and $H$ are collinear. Denote $K \in HI \cap EF$ . Let $J$ be midpoint of $EF$ . By Menelaos theorem for $\triangle DMK$ and points $P,J,H$ it's enough to prove that

$\dfrac{PM}{PN}\cdot \dfrac{HD}{HK}\cdot \dfrac{JK}{JM} = 1 \Longleftrightarrow \dfrac{HD}{HK} = \dfrac{JM}{JK} \ (*)$ .

From $AF = AE$ we get $IJ \perp EF$ , so $JI \parallel DM$ . Therefore

$\dfrac{JM}{JK} = \dfrac{ID}{IK}$ .

We rewrite the relation $(*)$ as

$\dfrac{HD}{HK} = \dfrac{ID}{IK}$

This is equivalent to prove that $(D,I,K,H)$ is harmonical.
$QGD$ is orthic triangle of $HBC$ . So $GI$ is bisector of $DGK$ . From $\angle IGH = 90^\circ$ we get that $GH$ is exterior bisector of $DGK$ and the result follows.

Image not found

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

Umut Varolgunes

279 posts

Umut Varolgunes

#5 h May 16, 2009, 7:25 PM • 2 Y

Y by Adventure10, Mango247

Sketch of my solution
1. The intersection points of the altitudes from B and C with the corresponding sides in triangle HBC are on the line EF (well known lemma)
2. H, I and D are collinear. (easy with the use of harmonical division in both triangles ABC and HBC)
3. HBC and DEF are similar. (angle chasing)
4. Let the orthocenter of DEF be H' and the midpoint of EF be N. Now from the well known property, H'D is equal to 2 times IN. Using 3, we find that AI/AD=IN/PD. Then the result follows from the fact that IN//PD.

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

mathVNpro

469 posts

mathVNpro

#6 h May 17, 2009, 5:13 AM • 2 Y

Y by Adventure10, Mango247

Here is another aproach for this problem:
Let $\{K\}\equiv CI\cap FE$ , $\{G\}\equiv BI\cap EF$ . It is very well- know that $BK\bot CK$ and $BG\bot CG$ (this result can be proved easily by harmonic division). Hence $\{H\}\equiv BK\cap CG$ . Let $J$ be the midpoint of $EF$ . Denote $P$ the intersection by $HJ$ and $DM$ . We will prove that $P$ is the midpoint of $DM$ .
Indeed, let $S$ be the projection of $H$ onto $EF$ , in order to prove $P$ is the midpoint of $DM$ , it is enough to prove $H(MJYS) = - 1$ , where $\{Y\}\equiv HD\cap EF$ , which is equivalent to proving $\frac {\overline {JM}}{\overline {JY}} = - \frac {\overline {SM}}{\overline {SY}}$ .

Note that $J$ is the midpoint of $EF\Longrightarrow IJ\bot EF\Longrightarrow IJ//DM$ , which implies $\frac {\overline {JY}}{\overline {JM}} = \frac {\overline {IY}}{\overline {ID}}$ .

But $(HIYD) = - 1\Longrightarrow \frac {\overline {IY}}{\overline {ID}} = - \frac {\overline {HY}}{\overline {HD}} = - \frac {\overline {HY}}{\overline {HY} + \overline {YD}} = - \frac {\overline {SY}}{\overline {SY} + \overline {YM}} = - \frac {\overline {SY}}{\overline {SM}}$
$\Longrightarrow \frac {\overline {SY}}{\overline {SM}} = - \frac {\overline {JY}}{\overline {JM}}$ .
Which implies that $(MJYS) = - 1\Longrightarrow H(MDPS) = - 1$ , but $DM//HS\Longrightarrow P$ is the midpoint of $DM$ , which leads to the result of the problem.
Our proof is completed then.

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

No Reason

114 posts

No Reason

#7 h May 18, 2009, 2:47 PM • 2 Y

Y by Adventure10, Mango247

Sorry for my bad English...

Here is the proof without calculation.
Lemma:In a triangle ABC,let D be the tangent of incircle $(I)$ to BC, $(I_a)$ be the excircle of A wrt triangle ABC and M be midpoint of the altitude from A wrt triangle ABC,then $D,I_a,M$ are collinear.
Proof:
Let $(I_a)$ meet BC at N, $NI_a$ intersect $(I_a)$ at N and P
The homothety center A take $(I)$ to $(I_a)$ also take D to P,then A,D,P are collinear.
Then the homothety center D take the altitude from A wrt triangle ABC to NP,it also take M to $I_a$
We conclude that $D,I_a,M$ are collinear.

Return to the problem,apply the lemma to the orthic triangle of triangle BHC then P,H,M' are collinear (M' is the midpoint of EF).
Q.E.D

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

reza1370

59 posts

reza1370

#8 h May 20, 2009, 4:55 PM • 2 Y

Y by Arefe, Adventure10

let $K$ midpoint of $EF$ and $r$ radius of incircle of $\triangle ABC$ .
let $PH$ cut $IK$ at $K'$ so $\frac{IK'}{DP}=\frac{HI}{HD}$ ( $i$ )(because of $IK || PD$ )
it is easy to measure this lenght:
$IK=r \sin\frac{A}{2}$ (1)
$DM=BD\cos\frac{B}{2}=2r\cos\frac{c}{2}\cos\frac{B}{2}$ so $DP=r\cos\frac{B}{2}\cos\frac{C}{2}$ (2)
with (1),(2): $\frac{IK}{DP}=\frac{\sin\frac{A}{2}}{\cos\frac{B}{2}\cos\frac{C}{2}}$ (3).

and $ID=BD \tan\frac{B}{2}$ , $DH= \cot\frac{C}{2}$ so
$HI=BD(\cot\frac{C}{2}-\tan\frac{B}{2})$ so
$\frac{HI}{DH}=1-\frac{\tan\frac{B}{2}}{\cot\frac{C}{2}}=1-\tan\frac{B}{2}\frac{C}{2}=\frac{\sin\frac{A}{2}}{\cos\frac{B}{2}\cos\frac{C}{2}}$ (4).

with (3),(4): $\frac{IK}{DP}=\frac{HI}{HD}$ ( $ii$ ).

with ( $i$ ),( $ii$ ): $PH$ passer from $K$ .
problem solved.

Image not found

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

goodar2006

1347 posts

goodar2006

#9 h Dec 4, 2011, 8:24 AM • 3 Y

Y by Adventure10 and 2 other users

let $T$ be the intersection point of $PH$ and $EF$ , $S$ the intersection point of $IH$ and $EF$ , and $R,T'$ be the feet of the perpendicular from $H$ and $I$ to the line $EF$ respectively. since $HR||DM$ and $P$ is the midpoint of $DM$ , we have $H(RPDM)=-1$ , so $(RTSM)=-1$ . on the other side, $EF$ is the line joining feet of the altitudes of vertices $B$ and $C$ of triangle $HBC$ respectively, so $(HISD)=-1$ and therefore $(RT'SM)=-1$ . we cocnlude that $T=T'$ , hence $T$ is the midpoint of $EF$ , and the proof is complete.

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

v_Enhance

6877 posts

v_Enhance

#10 h Aug 6, 2012, 1:30 PM • 32 Y

Y by infiniteturtle, anantmudgal09, rkm0959, JasperL, huricane, zacchro, mathadventurer, AlastorMoody, myh2910, HolyMath, hellomath010118, Smita, SHREYAS333, nguyendangkhoa17112003, Gaussian_cyber, Modesti, amar_04, MrOreoJuice, HamstPan38825, PRMOisTheHardestExam, rayfish, Adventure10, Mango247, Assassino9931, IAmTheHazard, Rounak_iitr, ohiorizzler1434, Funcshun840, and 4 other users

$[asy]import graph; size(12.73cm); real lsf=0.5; pathpen=linewidth(0.7); pointpen=black; pen fp=fontsize(10); pointfontpen=fp; real xmin=-7.08,xmax=5.65,ymin=-3.59,ymax=3.72; pen qqfftt=rgb(0,1,0.2), ttwwqq=rgb(0.2,0.4,0), zzttqq=rgb(0.6,0.2,0), ccqqqq=rgb(0.8,0,0); pair A=(-3,2), B=(-4,-3), C=(3,-3), I=(-1.86,-1.24), D=(-1.86,-3), F=(-3.58,-0.9), H=(-1.86,2.92), B_1=(0.19,0.43), C_1=(-3.19,-0.76), M=(-2.71,-0.59), P=(-2.28,-1.79), X_C=(-4.46,1.37), X_B=(0.75,1.37); D(A--B--C--cycle,linewidth(1.2)+dotted+qqfftt); D(D--(-0.73,0.11)--F--cycle,linewidth(1.2)+dotted+ttwwqq); D(B--C--H--cycle,zzttqq); D(B_1--C_1--D--cycle,blue); D(A--B,linewidth(1.2)+dotted+qqfftt); D(B--C,linewidth(1.2)+dotted+qqfftt); D(C--A,linewidth(1.2)+dotted+qqfftt); D(CR(I,1.76),linetype("2 2")+ttwwqq); D(D--(-0.73,0.11),linewidth(1.2)+dotted+ttwwqq); D((-0.73,0.11)--F,linewidth(1.2)+dotted+ttwwqq); D(F--D,linewidth(1.2)+dotted+ttwwqq); D(B--C,zzttqq); D(C--H,zzttqq); D(H--B,zzttqq); D(B_1--C_1,blue); D(C_1--D,blue); D(D--B_1,blue); D(D--M); D(CR(I,0.9),linewidth(1.6)+linetype("2 2")+blue); D(P--H,linewidth(1.6)+linetype("2 2")+ccqqqq); D(C_1--X_C,linetype("4 4")+blue); D(B_1--X_B,linetype("4 4")+blue); D(CR(H,3.03),linetype("4 4")+blue); D(A); MP("A",(-2.96,2.07),NE*lsf); D(B); MP("B",(-3.96,-2.93),NE*lsf); D(C); MP("C",(3.04,-2.93),NE*lsf); D(I); MP("I",(-1.82,-1.18),NE*lsf); D(D); MP("D",(-1.82,-2.93),NE*lsf); D((-0.73,0.11)); MP("E",(-0.68,0.17),NE*lsf); D(F); MP("F",(-3.54,-0.84),NE*lsf); D(H); MP("H",(-1.82,2.99),NE*lsf); D(B_1); MP("B_1",(0.23,0.49),NE*lsf); D(C_1); MP("C_1",(-3.15,-0.7),NE*lsf); D((-2.15,-0.39)); MP("N",(-2.11,-0.33),NE*lsf); D(M); MP("M",(-2.67,-0.52),NE*lsf); D(P); MP("P",(-2.24,-1.73),NE*lsf); D(X_C); D(X_B); clip((xmin,ymin)--(xmin,ymax)--(xmax,ymax)--(xmax,ymin)--cycle); [/asy]$

Let $N$ be the midpoint of $EF$ , and let $B_1 = EF \cap HC$ , $C_1 = EF \cap HB$ . It is not hard to see, via angle chasing, that $BFC_1I$ is cyclic, from which it follows that $IC_1 \perp C_1B \implies CC_1 \perp HB$ . Similarly $BB_1 \perp HC$ , so that $DB_1C_1$ is the orthic triangle of $HBC$ .

Consider $\triangle DB_1C_1$ . $N$ is the tangency point of its incircle with $B_1C_1$ and $H$ is the excenter opposite $D$ . It is well-known that $P$ , $N$ , and $H$ are collinear, implying the conclusion.

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

subham1729

1479 posts

subham1729

#11 h Apr 12, 2013, 9:45 PM • 2 Y

Y by Adventure10, Mango247

(Writing the Idea only, calculation is easy).Take $F=(1,0),E=(-1,0),I=(0,r),M=(a,0),D=(a,b)$ now so equation of circle $\odot(DEF)$ be $x^2+(y-r)^2=r^2+1$ . Now note tangent at $(x_1,y_1)$ point be $xx_1+yy_1-r(y+y_1)=1$ , using this we get $B,C$ in terms of $a,b,r$ . Now easily we’ll get $H$ in terms of $a,b,r$ . We’ve midpoint of $MD$ is $P=(a,\frac {b}{2})$ , now basically we need to show now slope of the line is $tan^{-1}(\frac {b}{2})$ and that’s quite easy to show. Note here we’re saying , solution is easy by CG because no stuffs involving any power is coming, and that makes our calculation easier. (I got a trig solution too, but there calculation is damn hard to do.)

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

AnonymousBunny

339 posts

AnonymousBunny

#12 h Jul 15, 2014, 7:39 PM • 2 Y

Y by Adventure10, Mango247

$[asy] /* Geogebra to Asymptote conversion, documentation at artofproblemsolving.com/Wiki, go to User:Azjps/geogebra */ import graph; size(10cm); real labelscalefactor = 0.5; /* changes label-to-point distance */ pen dps = linewidth(0.7) + fontsize(10); defaultpen(dps); /* default pen style */ pen dotstyle = black; /* point style */ real xmin = -1.535709806738513, xmax = 7.699401969463958, ymin = 1.537544229253762, ymax = 7.025735038013770; /* image dimensions */ pen zzttqq = rgb(0.6000000000000006,0.2000000000000002,0.000000000000000); draw((4.120000000000004,6.740000000000007)--(0.6000000000000006,2.120000000000002)--(4.046266301600477,2.770413386903411)--cycle, zzttqq); /* draw figures */ draw((4.120000000000004,6.740000000000007)--(0.6000000000000006,2.120000000000002), zzttqq); draw((0.6000000000000006,2.120000000000002)--(4.046266301600477,2.770413386903411), zzttqq); draw((4.046266301600477,2.770413386903411)--(4.120000000000004,6.740000000000007), zzttqq); draw((2.219650741964543,4.245791598828465)--(4.061766128857575,3.604874537240598)); draw((3.226139448965370,2.615630953617537)--(3.623387848646997,3.757397078019321)); draw((0.6000000000000006,2.120000000000002)--(2.823005108246210,4.751669850437923)); draw((3.424763648806184,3.186514015818429)--(2.823005108246210,4.751669850437923)); draw((3.623387848646997,3.757397078019321)--(4.046266301600477,2.770413386903411)); draw((2.823005108246210,4.751669850437923)--(4.046266301600477,2.770413386903411)); draw((3.623387848646997,3.757397078019321)--(4.046266301600477,2.770413386903411)); draw((4.046266301600477,2.770413386903411)--(2.355699169206497,4.198457018789681)); draw((0.6000000000000006,2.120000000000002)--(3.385861897857764,3.840038173964708)); draw((3.035843169874220,3.623930701373472)--(2.823005108246210,4.751669850437923)); draw((3.226139448965370,2.615630953617537)--(3.035843169874220,3.623930701373472)); /* dots and labels */ dot((4.120000000000004,6.740000000000007),dotstyle); label("$A$", (4.153046114633986,6.797820137317756), NE * labelscalefactor); dot((0.6000000000000006,2.120000000000002),dotstyle); label("$B$", (0.6340400478875363,2.175705951202600), NE * labelscalefactor); dot((4.046266301600477,2.770413386903411),dotstyle); label("$C$", (4.080113346411262,2.822984269179279), NE * labelscalefactor); dot((3.035843169874220,3.623930701373472),dotstyle); label("$I$", (3.068171187320962,3.679944295796290), NE * labelscalefactor); dot((2.219650741964543,4.245791598828465),dotstyle); label("$F$", (2.256794140843153,4.299872825689447), NE * labelscalefactor); dot((3.226139448965370,2.615630953617537),dotstyle); label("$D$", (3.259619703905613,2.668002136705990), NE * labelscalefactor); dot((4.061766128857575,3.604874537240598),dotstyle); label("$E$", (4.098346538466943,3.661711103740609), NE * labelscalefactor); dot((3.623387848646997,3.757397078019321),dotstyle); label("$M$", (3.660749929130597,3.807576640186058), NE * labelscalefactor); dot((3.424763648806184,3.186514015818429),dotstyle); label("$P$", (3.460184816518106,3.242347686459944), NE * labelscalefactor); dot((2.823005108246210,4.751669850437923),dotstyle); label("$H$", (2.858489478680629,4.810402203248517), NE * labelscalefactor); dot((3.140708435411058,3.925333068034533),dotstyle); label("$N$", (3.177570339655048,3.980791964715028), NE * labelscalefactor); dot((2.355699169206497,4.198457018789681),dotstyle); label("$J_c$", (2.393543081260761,4.254289845550244), NE * labelscalefactor); dot((3.385861897857764,3.840038173964708),dotstyle); label("$J_b$", (3.423718432406743,3.898742600464463), NE * labelscalefactor); dot((2.967592072231618,3.985564490881980),dotstyle); label("$X$", (3.004355015126078,4.044608136909912), NE * labelscalefactor); clip((xmin,ymin)--(xmin,ymax)--(xmax,ymax)--(xmax,ymin)--cycle); /* end of picture */ [/asy]$

Let $J_b$ and $J_c$ be the feet of perpendiculars from $B,C$ on $CH,BH$ respectively. Since $CI \perp BH$ and $BI \perp CH,$ $B,I,J_b$ and $C,I,J_c$ are collinear. Furthermore, since $\angle BJ_cC = \angle BJ_bB = 90^{\circ},$ $(BI_cI_bC)$ is cyclic and since $ID \perp BD, IF \perp BF, IJ_c \perp BJ_c,$ $(BDIJ_cF)$ is also cyclic. Now, note that $\angle FJ_cB = \angle FDB = 90^{\circ} - \dfrac{\angle ABC}{2}$ and $\angle J_bJ_cC = \angle J_bBC = \dfrac{\angle ABC}{2},$ so $\angle \angle FJ_cB + \angle BJ_cC + \angle J_bJ_cC = 180^{\circ},$ implying $F,J_c, J_b$ are collinear. Similarly, $E, J_b, J_c$ are collinear. Combining them, we see that $J_b$ and $J_c$ both lie on line $EF.$

Since $ID \perp BC,$ $H,I,D$ are collinear. Let this line intersect $EF$ at $X.$ Using cyclic quadrilaterals $BDJ_bH$ and $CDJ_cH,$ we deduce that
$\angle J_cJ_bI = 90^{\circ} - \angle HJ_bJ_c = 90^{\circ} - \angle CJ_bE = \angle IJ_bC,$
so $J_bI$ internally bisects $\angle J_cJ_bC.$ Since $IJ_b \perp J_bH,$ $J_bH$ externally bisects $\angle J_bJ_bC,$ so by the angle bisector theorem we have $\dfrac{HD}{ID} = \dfrac{HX}{IX}.$

Let $N$ be the midpoint of $EF.$ Note that $A$ and $I$ both lie on the perpendicular bisector of $EF,$ so $IN \perp EF,$ implying $(H,I;X,D)$ is harmonic, or $\dfrac{KX}{IX} = \dfrac{DH}{DI}.$

Plugging these ratios, a simple calculation shows that
$\dfrac{PM}{PN} \cdot \dfrac{DH}{DX} \cdot \dfrac{XN}{MN} = 1,$
so the desired result follows by the converse of Menelaus theorem. $\blacksquare$

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

sayantanchakraborty

505 posts

sayantanchakraborty

#13 h Oct 3, 2014, 7:15 PM • 4 Y

Y by rashah76, Entrepreneur, Adventure10, Mango247

I will provide a more straightforward and bashy solution.(but a synthetic solution like you have done is obviously more appreciated)

Applying the sine rule in $\triangle{BFD}$ we get
$\frac{FD}{sinB}=\frac{s-b}{cos\frac{B}{2}} \implies FD=2sin\frac{B}{2}(s-b)$

So $MD=FDcos\frac{C}{2}=2sin\frac{B}{2}cos\frac{C}{2}(s-b) \implies DP=sin\frac{B}{2}cos\frac{C}{2}(s-b)$

Let $K$ be the midpoint of $EF$ .Then $IK=IEsin\frac{A}{2}=rsin\frac{A}{2}=(s-b)tan\frac{B}{2}sin\frac{A}{2}$

Now easy angle chasing gives $\angle{ICH}=\frac{A}{2}$ and $\angle{HIC}=90+\frac{C}{2}$

So applying sine rule in $\triangle{HIC}$ we get
$HI=\frac{HCsin\frac{A}{2}}{cos\frac{C}{2}}$

Also $HD=HCcos\frac{B}{2}$ .

Now note that $\frac{IK}{PD}=\frac{(s-b)tan\frac{B}{2}sin\frac{A}{2}}{sin\frac{B}{2}cos\frac{C}{2}(s-b)}=\frac{sin\frac{A}{2}}{cos\frac{B}{2}cos\frac{C}{2}}$

Again $\frac{HI}{HD}=\frac{\frac{HCsin\frac{A}{2}}{cos\frac{C}{2}}}{HCcos\frac{B}{2}}=\frac{sin\frac{A}{2}}{cos\frac{B}{2}cos\frac{C}{2}}$

Thus $\frac{HI}{HD}=\frac{IK}{PD}$ .Also note that $IK \parallel PD$ ,both being perpendicular to $EF$ .

So $H,K,P$ are collinear as desired.

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

andria

824 posts

andria

#14 h Jun 18, 2015, 5:03 PM • 4 Y

Y by buratinogigle, rashah76, Adventure10, Mango247

My solution:
Let $BI\cap EF=T$ note that $\angle BTF=\angle AFE-\angle ABI=\frac{C}{2},\angle ECI=\frac{C}{2}\longrightarrow IETC$ is cyclic so $\angle IEC=\angle ITC=90$ thus $C,T,H$ are collinear similarly $B,S,H$ are collinear now let the perpendicular lines from $D$ to $DE,DF$ intersect $EF$ at $K,L$ note that $BT||DL$ (because both of them are perpendicular to $HC$ ) and $\angle STB=\angle BTD=\frac{C}{2}$ thus $TL=TD$ so $TH$ is perpendicular bisector of $DL$ similarly $HS$ is perpendicular bisector of $DK$ so $H$ is circumcenter of $\triangle DKL$ now using this problem we get that $P,M,H$ are collinear.
DONE

Attachments:

This post has been edited 1 time. Last edited by andria, Jun 18, 2015, 5:05 PM

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

MouN

26 posts

MouN

#15 h Aug 25, 2015, 5:42 PM • 1 Y

Y by Adventure10

Let $Q=BI \cap EF$ , $N=AI \cap EF$ , and $K=BI \cap DF$ .
Since $\angle CIQ=\angle CBI + \angle ICB = 90^{\circ} - \angle IAE = \angle AEF = \angle CEQ$ and $90^{\circ}=\angle CDI = \angle IEC$ the pentagon $CDIEQ$ is cyclic with $CI$ as diameter, so $\angle IQC = 90^{\circ}$ and hence $Q \in CH$ .
Denote by $h$ the homothety centered at $H$ sending $I$ to $D$ , and let $R=h(Q)$ . We claim that $h(N)=P$ from which the assertion follows since $N$ is also the midpoint of $EF$ .
But obviously $h(IN)=DP$ since $IN \parallel DP$ , so we just have to show that $h(QN)=RP$ . Note first that $h(IQ)=DR$ so $IQ \parallel DR$ which means the pentagon $DKMQR$ is cyclic since $90^{\circ} = \angle QRD = \angle DKQ = \angle DMQ$ . Note also that $KP$ is the perpendicular bisector of $DM$ since it is the midline of the right angled triangle $DFM$ , so $\angle DKR = 90^{\circ} - \angle KRD = 90^{\circ} - \angle KMD = \angle DKP$ which yields $R \in KP$ , but $KP\parallel FM$ so we are done.

This post has been edited 1 time. Last edited by MouN, Aug 25, 2015, 7:06 PM
Reason: wrong names for points

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

john0512

4190 posts

john0512

#62 h Sep 8, 2023, 2:05 AM

Y by

Let $T_B$ and $T_C$ denote the feet from $B$ to $CH$ and the feet from $C$ to $BH$ respectively. Note that $BICH$ is an orthocentric system, so $B,I,T_B$ are collinear as well as $C,I,T_C.$ By that egmo lemma, $BI$ and $EF$ intersect at some point $K$ for which $KB\perp CK.$ Thus, $K=T_B$ , since the intersections of $BI$ and $(BC)$ are $B$ and $T_B$ .

Now, we have an orthic triangle configuration with $D,T_B,T_C$ in $\triangle HBC$ . Hence, $I$ is the incenter of $\triangle DT_BT_C.$ Since $T_B$ and $T_C$ lie on $EF$ , the foot from $I$ to $T_BT_C$ is just the midpoint of $EF$ .

Hence, in $\triangle DT_BT_C$ , the midpoint of the D-altitude, $P$ , the intouchpoint to $T_BT_C$ , which is the midpoint of $EF$ , and the $D$ -excenter, which is $H$ , are collinear, hence $PH$ bisects $EF$ as desired.

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

cursed_tangent1434

635 posts

cursed_tangent1434

#63 h Oct 26, 2023, 12:09 AM • 2 Y

Y by Shreyasharma, ihategeo_1969

My sol is crazy long bruh. Just noticed that this actually connects up with 2011 ISL G4.

One first considers the following claim.

Claim (Partial Iran Lemma) : The feet of the perpendiculars from $B$ to $CI$ and $C$ to $BI$ respectively lie on $\overline{EF}$ .

Let $C_1=\overline{EF} \cap \overline{CI}$ . Then, note that
$\begin{align*} 2 \measuredangle EC_1I &= 2 \measuredangle ECI +2 \measuredangle C_1EC\\ &= \measuredangle ACB + 2 \measuredangle FEA\\ &= \measuredangle ACB + \measuredangle FAE \\ &= \measuredangle ACB + \measuredangle BAC\\ &= \measuredangle ABC\\ &= 2\measuredangle FBI \end{align*}$ Thus, $\measuredangle EC_1I= \measuredangle FBI$ . This means that $FBIC_1$ is cyclic and thus $\measuredangle IC_1B = 90^\circ$ which means that indeed $C_1$ is the foot of the perpendicular from $B$ to $CI$ . Similarly, the foot of the perpendicular from $C$ to $BI$ also lies on $EF$ and we have our claim.

Now, we have the following very important claim.
Claim (Extended 11SLG4) :
Consider a triangle $ABC$ with $D,E,F$ the feet of the perpendiculars from $A,B,C$ to the opposite sides. Let $H$ be its orthocenter and $Y_A$ be the $A-$ Why Point of $\triangle ABC$ . Let $P$ be the foot of the altitude from $D$ to $EF$ Then, the following points are collinear,
1. $A$
2. Foot of the perpendicular from $H$ to $EF$ ( $Z$ )
3. Midpoint of the altitude from $D$ to $EF$ ( $M$ )
4. $Y_A$

First, consider the inversion centred at $A$ with radius $\sqrt{AD \cdot AH}$ . This clearly maps $D$ to $H$ , $B$ to $F$ and $C$ to $E$ (and vice versa). Thus, $\overline{EF}$ maps to $(ABC)$ and $\overline{BC}$ maps to $(AEF)$ and thus the $A-$ Ex Point $X_A$ maps to the $A-$ Queue Point $Q_A$ under this inversion. Further, since $D$ and $H$ map to each other, this in fact means that the circle $(X_AH)$ remains fixed under this inversion ( $Q_A$ clearly lies on this circle as $\measuredangle HQ_AX_A=90^\circ$ ). Further, $Z$ will map to the intersection of $(X_AH)$ and $(ABC)$ . The claim is that this intersection is $Y_A$ .

Let $A'$ be the reflection of $A$ across the perpendicular bisector of $BC$ . Let $G$ be the centroid of $\triangle ABC$ and $W$ the intersection of $\overline{AM}$ and $(ABC)$ . It is well known (from 11SLG4) that $Y_A-D-G-A'$ (with $A'G=2DG$ ). Simply note that
$GM \cdot GW = \frac{AG \cdot GW}{2} = \frac{A'G\cdot GY_A}{GD \cdot GY_A} = GD \cdot GY_A$ Thus, $DMWY_A$ is cyclic. Now, let $X_A' = \overline{Y_AW} \cap \overline{BC}$ . Then,
$X_A'D\cdot X_A'M = X_A'Y_A\cdot X_A'W = X_A'B \cdot X_A'C$ Thus, $X_A' \equiv X_A$ and we have that $X_A-Y_A-W$ . Let $H_A$ be the $A-$ Humpty Point. It is well known that $X_A-H-H_A$ and $HH_AMD$ is cyclic. Now, note that from this it follows that,
$\measuredangle X_AY_AD = \measuredangle WY_AD = \measuredangle AMD=\measuredangle H_AMD = \measuredangle H_AHA = \measuredangle X_AHD$ and thus, $X_AHDY_A$ is cyclic and $Y_A$ lies on $(X_AH)$ as desired.

Thus, $Z$ maps to $Y_A$ in the previously described inversion and we have that $A-Z-Y_A$ .

Now, we deal with the other part of the collinearity. Let $N$ be the midpoint of $X_AD$ . We already know that $X_AY_ADZ$ is cyclic. Further, since $B,C,H_A$ and $Y_A$ are cyclic and $(X_AD;BC)=-1$ and we also have,
$\measuredangle H_AY_AD = \measuredangle H_AA' =90^\circ$ $Y_A$ must be the $H_A-$ Humpty Point of $\triangle X_ADH_A$ . Further, this implies that $H_AY_A$ is the $H_A-$ median of this triangle which implies $H_A-Y_A-N$ . Thus, $\measuredangle NY_AD = 90^\circ$ . Further, $MN \parallel X_AP$ by Midpoint Theorem, which implies that $NY_ADM$ must indeed be cyclic.

Now,
$\measuredangle DY_AM = \measuredangle DNM = \measuredangle DX_AZ = \measuredangle DY_AZ$ This clearly implies that $Y_A-M-Z$ and we put these two together and conclude that indeed $A-Z-M-Y_A$ are collinear as claimed.

By the first claim, $EF$ is the line joining the feet of the altitudes from $B$ and $C$ to the opposite sides. One now notices that by applying the above-proved claim on $\triangle BHC$ , we must have $H, M$ and the foot of the perpendicular from $I$ to $EF$ collinear. But, clearly $AI$ is the perpendicular bisector of $EF$ meaning that, the perpendicular from $I$ to $EF$ passes through the midpoint of $EF$ .

Thus, we have that $\overline{PH}$ bisects $\overline{EF}$ as was required.

This post has been edited 4 times. Last edited by cursed_tangent1434, Oct 26, 2023, 12:12 AM
Reason: latex error

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

Cusofay

85 posts

Cusofay

#64 h Feb 6, 2024, 6:23 PM

Y by

Let $N$ be the midpoint of $[EF]$ , $B'=(BI)\cap (EF)$ and $C'= (CI)\cap(EF)$ . By the Iran lemma (provable by angle chase), the triangle $\triangle DB'C'$ with incenter $I$ is the orthic triangle of $\triangle HBC$ . Now since $P$ is the midpoint of the $D-$ altitude in $\triangle DB'C'$ and $H$ the $D-$ excenter point, the collinearity is well known(4.3 midpoints of Altitudes EGMO).

$\mathbb{Q.E.D.}$

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

shendrew7

796 posts

shendrew7

#65 h Feb 13, 2024, 6:53 PM

Y by

Let $K = BI \cap EF$ and $L = CI \cap EF$ . Iran Lemma tells us $\triangle DKL$ is the orthic triangle of $\triangle HBC$ . Hence Midpoint of the Altitude Lemma this triangle with altitude midpoint $P$ , touch point $M$ (midpoint of $EF$ ), and excenter $K$ finishes. $\blacksquare$

This post has been edited 1 time. Last edited by shendrew7, Feb 13, 2024, 6:54 PM

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

HamstPan38825

8867 posts

HamstPan38825

#66 h Mar 5, 2024, 2:10 PM • 1 Y

Y by panche

alright here's the stock solution I guess

Let $X$ and $Y$ be the feet from $C$ to $\overline{BI}$ and from $B$ to $\overline{CI}$ , which both lie on $\overline{EF}$ . It follows that $DXY$ is the orthic triangle of $HBC$ , i.e. $H$ is the $D$ -excenter of triangle $DXY$ and $I$ is the incenter. Midpoint of altitudes lemma implies $K = \overline{PH} \cap \overline{EF}$ satisfies $\angle IKE = 90^\circ$ , or $K$ is the midpoint of $\overline{EF}$ .

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

EpicBird08

1754 posts

EpicBird08

#68 h May 28, 2024, 2:31 AM

Y by

WOAH. Took a "hint" because I heard from somewhere that this is the problem after which the "Iran Lemma" is named.

Let $X$ and $Y$ be the feet of the altitudes from $B,C$ to $CI,BI,$ respectively. By the Iran Lemma, we see that $X,Y,E,F$ are collinear. The condition that $PH$ bisects $EF$ is equivalent to $PH$ passing through the foot of the altitude from $I$ to $XY.$ Since $DYX$ is the orthic triangle of $\triangle BHC,$ we see that $I$ is the incenter of $\triangle XYD,$ so the foot from $I$ to $XY$ is the $D$ -intouch point of $\triangle DXY.$ Additionally, it is clear that $H$ is the $D$ -excenter of $\triangle DXY.$ Thus we must show that the midpoint of the $D$ -altitude, the $D$ -excenter, and the $D$ -intouch point are collinear, which is well-known.

This post has been edited 1 time. Last edited by EpicBird08, May 28, 2024, 2:50 AM

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

ezpotd

1277 posts

ezpotd

#69 h Jun 11, 2024, 10:10 PM • 1 Y

Y by L13832

We use complex bash with the unit circle as the incircle. First, we calculate the circumcenter of $(BIC)$ . We know that $(AB)(AC) = (AI)(AI_A)$ , and by considering angles we can see that $\frac{(a - b)(a - c)}{(a - i)} = (a - i_a)$ . So we have $a = \frac{2ef}{e + f}$ trivially and cyclic variants. Now we write $\frac{(a - b)(a - c)}{(a - i)} = \frac{4ef(\frac{f}{e + f} - \frac{d}{d + e}) (\frac{e}{e + f} - \frac{d }{d + f})}{\frac{2ef}{e + f}} = \frac{2(ef - de)(ef - df)}{(e + f)} = \frac{2ef(d - f)(d - e)}{(e + f)(d+e)(d+f)}$ . Now we find $i_a = \frac{2ef}{e + f} (1 - \frac{(d - e)(d-f)}{(d+e)(d+f)}) = \frac{2ef}{e + f} \frac{2de + 2df}{(d + e)(d+f)} = \frac{4def}{(d+e)(d+f)}$ . The circumcenter of $(BIC)$ is the midpoint of $II_a$ , which is just $\frac{2def}{(d + e)(d+f)}$ . Now the orthocenter is just given by $b + i + c - 2o = \frac{2de}{d + e} + \frac{2df}{d + f} - \frac{4def}{(d + e)(d + f)} = 2d(\frac{e}{d + e} + \frac{f}{d + f} - \frac{2ef}{(d + e)(d+f)}) = 2d(\frac{de + ef + df + ef - 2ef}{(d + e)(d + f)}) = \frac{2d^2(e+f)}{(d+e)(d+f)}$ .

Now, our strategy will be to show that the midpoint of $EF$ is collinear with $P,H$ . Find the midpoint as $\frac{e + f}{2}$ , now find $M$ as $\frac 12 (d + e + f - \frac{ef}{d})$ , and then $P$ is given as $\frac{3d + e + f - \frac{ef}{d}}{4}$ . Then we desire $\frac{2e + 2f - 4p}{e + f - 2h}$ is real, which is equivalent to showing $\frac{e + f - 3d + \frac{ef}{d}}{e + f - 4d^2\frac{(e+f)}{(d+e)(d+f)}} = \frac{(d+e)(d+f)(e + f - 3d + \frac{ef}{d})}{((d+e)(d+f)(e+f)-4d^2(e+f))} = \frac{(d+e)(d+f)(e + f -3d + \frac{ef}{d})}{e^2f+f^2e+e^2d+f^2d + 2def - 3d^2e-3d^2f} = \frac{(d+e)(d+f)}{(de+df)}$ which is obviously self conjugating, so we are done.

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

Eka01

204 posts

Eka01

#70 h Aug 17, 2024, 7:19 AM • 1 Y

Y by AaruPhyMath

Notice that in $\Delta BHC$ , $I$ is the orthocenter due to Iran lemma and it is also the incenter of orthic triangle of $\Delta BHC$ . The midpoint of $EF$ is the foot of altitude from $I$ to $EF$ and $P$ is the midpoint of the $D$ altitude of the orthic triangle of $BHC$ and $H$ is the $D$ excenter of this orthic triangle so the desired result follows by midpoint of altitude lemma.

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

cj13609517288

1922 posts

cj13609517288

#71 h Oct 28, 2024, 5:46 PM • 1 Y

Y by ehuseyinyigit

Let $Q$ be the midpoint of $EF$ , let $S$ be the foot of the perpendicular from $H$ to $EF$ , and let $K$ be the "Iran lemma point" (the concurrency point of $EF$ , $CH$ , $BI$ ). Then
$-1=(EF\cap BC,D;C,B)\stackrel{K}{=}(RD;HI)\stackrel{\infty_{\perp EF}}{=}(RM;SQ)\stackrel{H}{=}(D,M;\infty_{DM},HQ\cap DM),$ so $HQ$ bisects $MD$ , as desired.

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

shendrew7

796 posts

shendrew7

#72 h Nov 4, 2024, 6:34 PM

Y by

Let $N$ be the midpoint of $EF$ , and note
$(MD;P\infty) \overset{N}{=} (MN \cap HD, D; HI) = -1. \quad \blacksquare$

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

Mathandski

766 posts

Mathandski

#73 h Nov 28, 2024, 2:04 AM

Y by

Synthetic sol

Subjective Rating (MOHs) $\text{[math]}$

Attachments:

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

awesomeming327.

1721 posts

awesomeming327.

#74 h Dec 27, 2024, 10:00 PM

Y by

Let $H_B$ and $H_C$ be the feet of the altitudes from $B$ and $C$ to $CI$ and $BI$ , respectively. It's easy to see that $FH_BIDB$ and $EH_CIDC$ are cyclic.

Claim: $H_B$ and $H_C$ lie on $EF$ .

It's clear by
$\measuredangle H_BFI=\measuredangle H_BBI=90^\circ-\measuredangle BIH_B=90^\circ-\measuredangle BIC=-\tfrac{\measuredangle BAC}{2}=\measuredangle EFI$ with $H_C$ following analogously.

Let $N$ be the intersection of $HP$ with $EF$ . Let $H_A$ be the foot of the altitude from $H$ to $EF$ . Let $T$ be the intersection of $DH$ and $EF$ . We will use LinPOP on $(BID)$ and $(CID)$ . For any point $X$ let $f(X)$ be $\text{Pow}_{(BID)}(X)-\text{Pow}_{(CID)}(X)$ . LinPOP states that $f$ is linear. It suffices to show that $f(N)=\tfrac{f(F)+f(E)}{2}$ .

Since $T$ is on the radical axis of the two circles, $f(T)=0$ . By Menelaus on $PNH$ traversing $\triangle MTD$ ,
$\frac{MN}{TN}\cdot \frac{TH}{DH}\cdot \frac{DP}{MP}=1\implies \frac{MN}{TN}=\frac{DH}{TH}\implies \frac{NT}{MT}=\frac{TH}{DH+TH}$ Therefore,
$f(N) = \frac{TH}{DH+TH}f(M)$ Since $\angle DFE=\angle CBH$ and $\angle DEF=\angle BCH$ , $\triangle DEF\sim \triangle HCB\sim \triangle HH_BH_C$ . Therefore,
$\frac{TH}{DH+TH}=\frac{d(H,H_BH_C)}{2d(H,H_BH_C)+d(D,EF)}=\frac{H_BH_C}{2H_BH_C+EF}$ Let $MF=u$ , $ME=v$ , $MH_B=x$ , $MH_C=y$ . Furthermore, set $MD=1$ . We have
$\begin{align*} 2f(N) &= \frac{2x+2y}{2x+2y+u+v} \cdot (xu-vy) \\ f(F)+f(E) &= (u+v)(v+x-u-y) \end{align*}$ Now, we relate $x$ with $v$ and $y$ with $u$ .

Claim: $2x=\tfrac{1}{v}-v$ ; $2y=\tfrac{1}{u}-u$ .

It is simple. We have by orthocenter-incenter duality that $\triangle DH_BH_C\sim \triangle ABC$ implying $\angle DH_CM=\angle C$ while $\angle MDF=90^\circ-\angle EFD=\tfrac{\angle C}{2}$ so
$y=\cot{\angle C}=\frac{1-\tan^2\left(\frac{\angle C}{2}\right)}{2\tan\left(\frac{\angle C}{2}\right)} = \frac{1-u^2}{2u}$ and the result follows.

Therefore, we have
$\begin{align*} 2f(N) &= \left(\frac{\frac{1}{v}-u+\frac{1}{u}-v}{\frac{1}{u}+\frac{1}{v}}\right) \left(xu-vy\right) \\ &= \frac{1}{2}\left(\frac{u+v-uv(u+v)}{u+v}\right) \left(\frac{u}{v}-\frac{v}{u}\right) \\ &= \frac{1}{2}\left(\frac{u+v-uv(u+v)}{u+v}\right) \left(\frac{(u-v)(u+v)}{uv}\right) \\ &= \frac{1}{2}\left(\frac{(u+v-uv(u+v))(u-v)}{uv}\right) \\ &= (u+v)\left(\frac{(1-uv)(u-v)}{2uv}\right) \\ &= (u+v)\left(\frac{u+uv^2}{2uv}-\frac{v+u^2v}{2uv}\right) \\ &= (u+v)\left(\frac{1+v^2}{2v}-\frac{1+u^2}{2u}\right) \\ &= (u+v)\left(\frac{1-v^2}{2v}+v-\frac{1-u^2}{2u}-u\right) \\ &= (u+v)\left(x+v-y-u\right) \end{align*}$ and we are done.

Remark

This post has been edited 1 time. Last edited by awesomeming327., Dec 27, 2024, 10:04 PM

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

ihategeo_1969

236 posts

ihategeo_1969

#75 h Mar 30, 2025, 12:50 AM • 1 Y

Y by cursed_tangent1434

We will define some new points.
$\bullet$ Let $\triangle DXY$ be the orthic triangle of $\triangle IBC$ .
$\bullet$ Let $\triangle D'E'F'$ be the medial triangle of $\triangle DEF$ .
$\bullet$ Let $T=\overline{ID'} \cap \overline{E'F'}$ .

By Iran Lemma we have $X=\overline{BF'I} \cap \overline{EF}$ and $Y=\overline{CE'I} \cap \overline{EF}$ .

Now consider the homothety at $I$ that sends $X \to F'$ and $Y \to E'$ ; see this also maps $(IXHY) \to (IF'DE')$ and so $H \to D$ and also $D' \to T$ . This also means $\overline{D'P} \to \overline{T\infty_{D'P}}$ .

So all we need to prove is $\overline{TD} \parallel \overline{D'P}$ which is trivial just because $D'F'DE'$ is a parallelogram and gliding principle or whatever.

Remark: I was doing this while watching ``Grave of the Fireflies" which is probably the saddest film ever. Idk why I said it, I just thought I should. Bye.

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

Ilikeminecraft

656 posts

Ilikeminecraft

#76 h May 4, 2025, 6:53 AM

Y by

Define $K$ the foot from $C$ to $BH.$ Define $L$ the intersection of $DI$ with $EF.$ Iran lemma implies $\angle EKD$ is bisected by $CK.$ By apollonian circles configuration, it follows $(HI;DL) = -1.$
Define $N$ the midpoint of $EF.$ We clearly have $(MD;P\infty) \stackrel N= (L, D; PN\cap ID, I)$ but uniqueness of harmonic bundles implies the result.

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

Markas

150 posts

Markas

#77 h May 11, 2025, 5:46 PM

Y by

Let $BL \perp IC$ and $L \in IC$ , $CK \perp BI$ and $K \in BI$ . Now from the Iran lemma it follows that $L \in EF$ and $K \in EF$ . Now from the conditions we have that $\triangle DKL$ is the orthic triangle for $\triangle BCH$ and I is the incenter of $\triangle DLK$ . We have that H is the $I_d$ excenter for $\triangle DLK$ . Now let N be the midpoint of EF $\Rightarrow$ if we show that $N \in PH$ we will be ready. We know that IE = IF and N is the midpoint of EF $\Rightarrow$ $IN \perp EF$ $\Rightarrow$ N is the intouch point for the incircle of $\triangle DKL$ and EF. Now we have that DM is an altitude in $\triangle DLK$ , P is midpoint of DM, N is an intouch point and H is the $I_d$ excenter $\Rightarrow$ from the Midpoint of an altitude configuration it follows immediately that P, N, H lie on one line $\Rightarrow$ PH bisects EF and we are ready.

Z K Y