Stay ahead of learning milestones! Enroll in a class over the summer!

2014 AIME I Math Jam

Go back to the Math Jam Archive

AoPS instructors discuss all 15 problems of the 2014 AIME I.

Copyright © 2024 AoPS Incorporated. This page is copyrighted material. You can view and print this page for your own use, but you cannot share the contents of this file with others.

Facilitator: Dave Patrick

DPatrick 2014-03-15 19:00:19
Welcome to the 2014 AIME I Math Jam!
DPatrick 2014-03-15 19:00:28
I'm Dave Patrick, and I'll be leading our discussion tonight.
DPatrick 2014-03-15 19:00:36
Before we get started I would like to take a moment to explain our virtual classroom to those who have not previously participated in a Math Jam or one of our online classes.
DPatrick 2014-03-15 19:00:52
This classroom is moderated, meaning that you can type into the classroom, but these comments will not go directly into the room. These comments go to the moderators, who may choose to share your comments with the room.
DPatrick 2014-03-15 19:01:02
This helps keep the session organized and on track. This also means that only well-written comments will be dropped into the classroom, so please take time writing responses that are complete and easy to read.
DPatrick 2014-03-15 19:01:20
There are a lot of students here! As I said, only a relatively small fraction of the well-written comments will be passed to the entire group. Please do not take it personally if your comments do not get posted, and please do not complain about it. I expect this Math Jam to be much larger than our typical class, so please be patient with me---there are quite a few of you here tonight!!
DPatrick 2014-03-15 19:01:40
Also, we won't be going through all the math quite as thoroughly as we do in our classes -- I can't teach all the necessary material for every problem as we go. Another difference between tonight and our regular online classes is that it is very unlikely that we'll be able to answer every single question you ask. We always to try do so in our regular online classes, but we have a large number of students tonight! So, please go ahead and ask questions, but also please understand if we aren't able to answer them all!
DPatrick 2014-03-15 19:02:15
We do have two teaching assistants with us tonight to help answer your questions: Elena (Anna Smith) and Alyssa (baozhale). (At least I expect to have two -- one is apparently running late.)
DPatrick 2014-03-15 19:02:30
They can answer questions by whispering to you or by opening a window with you to chat 1-on-1. However, due to the large size of the session tonight, they may not be able to get to you right away (or at all). Repeating your question over and over is more likely to annoy us than to get it answered faster, so please, just ask your question once and be patient, and please understand that we may not be able to answer all the questions tonight.
DPatrick 2014-03-15 19:03:06
Please also remember that the purpose of this Math Jam is to work through the solutions to AIME problems, and not to merely present the answers. "Working through the solutions" includes discussing problem-solving tactics. Also on occasion we may stop to prove things that you wouldn't necessary need to prove while doing the contest. So please, when a question is posted, do not simply respond with the final answer. That's not why we're here. We're going to work through the problems step-by-step, so please don't write comments that skip key steps or jump ahead in the problem.
DPatrick 2014-03-15 19:03:34
Let's get started! We're going to work through all 15 problems, in order.
DPatrick 2014-03-15 19:03:42
1. The 8 eyelets for the lace of a sneaker all lie on a rectangle, four equally spaced on each of the longer sides. The rectangle has a width of 50 mm and a length of 80 mm. There is one eyelet at each vertex of the rectangle. The lace itself must pass between the vertex eyelets along a width side of the rectangle and then crisscross between successive eyelets until it reaches the two eyelets at the other width side of the rectangle as shown. After passing through these final eyelets, each of the ends of the lace must extend at least 200 mm farther to allow a knot to be tied. Find the minimum length of the lace in millimeters.
DPatrick 2014-03-15 19:03:48
DPatrick 2014-03-15 19:03:59
(Note that I will always place the current problem under discussion at the top of the classroom window. You can resize the top area of the classroom by dragging the horizontal bar separating it from the rest of the classroom.)
DPatrick 2014-03-15 19:04:17
What is the total lace length that we have to count up?
DPatrick 2014-03-15 19:04:25
How would you describe it, in words?
poweroftwo 2014-03-15 19:04:58
ends, bottom, and diagonals
minimario 2014-03-15 19:04:58
The 2 ends + the 6 crosses + the bottom line
borntobeweild 2014-03-15 19:04:58
The 6 digs, the bottom, and the two trailing ends
checkmate1021 2014-03-15 19:04:58
the six diagonal things, the bottom width, and the two ends
cellobix 2014-03-15 19:04:58
2 ends, 6 diagonals, and 1 horizontal piece
DuoCapital 2014-03-15 19:04:58
The bottom, plus the complementary diagonals, plus the two ends
Quadratic64 2014-03-15 19:04:58
6 * diagonals + 2 * 200mm extensions + straight part on the bottom
DPatrick 2014-03-15 19:05:07
Right. Simply put, we have:
2 * (the length of laces that extends to either side) +
6 * (the length of the diagonal segments) +
(the length of the bottom segment)
DPatrick 2014-03-15 19:05:23
The first of these is easy: the laces extend for 200 mm on either side.
joshualee2000 2014-03-15 19:05:42
the bottom is just 50mm
DPatrick 2014-03-15 19:05:51
The last (the bottom piece) is also easy: it's just the width of 50 mm.
DPatrick 2014-03-15 19:05:56
How about the diagonal lengths?
mathtastic 2014-03-15 19:06:23
You can use a well known theorem, known as the Pythagorean Theorem, to find the lengths of the diagonals.
aimingforimo 2014-03-15 19:06:23
draw right triangles
ingridzhang97 2014-03-15 19:06:23
the diagonals are easy too. just use pythagorean
blizzard10 2014-03-15 19:06:23
pythagoream theorem
vinayak-kumar 2014-03-15 19:06:23
Pythagorean Theorem
hamup1 2014-03-15 19:06:23
each eyelet section is 80/3, and the width is 150/3 (pythag triples)
DPatrick 2014-03-15 19:06:34
Right. We can zoom in on one of the diagonal segments.
DPatrick 2014-03-15 19:06:40
DPatrick 2014-03-15 19:06:59
As you can see, it's a right triangle. The top side has length 50 mm, and the right side has length (80/3) mm.
jeff10 2014-03-15 19:07:24
hypotenuse of the right triangle with legs 50 and 80/3, which is 150/3 and 80/3, so each of the 6 crosses has a length of 170/3
LightningX48 2014-03-15 19:07:24
so it's 170/3 (8, 15, 17)
Tommy2000 2014-03-15 19:07:24
use pythagorean theorem for 50, 80/3 to get the diagonal as 170/3
sat113 2014-03-15 19:07:24
8-15-17
dli00105 2014-03-15 19:07:24
170/3
summitwei 2014-03-15 19:07:24
8-15-17
Rectangle_Square 2014-03-15 19:07:24
one side 80/3 other is 150/3 knowing 8-15-17 triple the diagonal is 170/3
DPatrick 2014-03-15 19:07:38
Exactly. You could use the Pythagorean Theorem, but we know that the final answer is an integer, so we suspect that this triangle must be a multiple of a Pythagorean triple.
DPatrick 2014-03-15 19:07:48
Indeed, the triangle is similar to an 8-15-17 triangle (with scale factor 10/3: 80/3 = (10/3)*8 and 50 = (10/3)*15).
DPatrick 2014-03-15 19:08:08
So the hypotenuse has length 170/3 mm.
ninjataco 2014-03-15 19:08:37
So the total is 2*200 + 50 + 6*170/3 = 790
1915933 2014-03-15 19:08:37
2(200)+6(170/3)+1(50)=790
Eudokia 2014-03-15 19:08:37
multiply by 6 to get 340 for the diagonals
tRIG 2014-03-15 19:08:37
multiply that by 6?
bunniesrcute 2014-03-15 19:08:37
so our final answer is 50 + 6*170/3 + 200*2 = 790
DPatrick 2014-03-15 19:08:49
Right, summing up all these lengths, we get a total length of $2(200) + 6(170/3) + 50 = 400 + 340 + 50 = \boxed{790}.$
DPatrick 2014-03-15 19:09:23
That was the traditional AIME warmup question -- pretty straightforward but as always you have to be careful with the computation.
DPatrick 2014-03-15 19:09:30
On to #2:
DPatrick 2014-03-15 19:09:34
2. An urn contains 4 green balls and 6 blue balls. A second urn contains 16 green balls and $N$ blue balls. A single ball is drawn at random from each urn. The probability that both balls are of the same color is $0.58.$ Find $N.$
DPatrick 2014-03-15 19:09:51
How can we approach this?
VietaFan 2014-03-15 19:10:28
Cases: Blue and blue or green and green
bob2 2014-03-15 19:10:28
casework for 2 green balls and 2 blue balls
aimingforimo 2014-03-15 19:10:28
find probability of green, then add to blue
genesis2 2014-03-15 19:10:28
probability of green balls from both boxes + same for blue
navygoat123 2014-03-15 19:10:28
find the probability for two green balls, and find the probability for two blue balls and add them up
DPatrick 2014-03-15 19:10:33
Right. There are two cases for us to "succeed" in drawing two balls of the same color: we can draw two green balls, or we can draw two blue balls.
DPatrick 2014-03-15 19:10:43
These cases are exclusive (they don't overlap), so we can compute the probability separately of each case, and then add those probabilities to get the overall probability of success.
DPatrick 2014-03-15 19:10:53
What's the probability of drawing 2 green balls?
bellyflop 2014-03-15 19:11:21
both green=> (4/10)(16/16+N)
brian22 2014-03-15 19:11:21
Green balls: (4/10)*(16/16+N)
forthegreatergood 2014-03-15 19:11:21
Zer0waltz 2014-03-15 19:11:21
2/5 times 16/16+n
raptorw 2014-03-15 19:11:21
4/10*16/(N+16)
DPatrick 2014-03-15 19:11:29
There's a $\dfrac25$ chance of drawing a green ball from the first urn.
DPatrick 2014-03-15 19:11:36
There's a $\dfrac{16}{N+16}$ chance of drawing a green ball from the second urn.
DPatrick 2014-03-15 19:11:44
So there's a $\dfrac{32}{5(N+16)}$ of drawing green balls from both urns.
DPatrick 2014-03-15 19:11:51
What's the probability of drawing 2 blue balls?
sparkles257 2014-03-15 19:12:14
3/5 * N/16+N
maxwellfl 2014-03-15 19:12:14
(6/10)*(N/16+N) for blues
checkmate1021 2014-03-15 19:12:14
$\frac{6}{10}\times \frac{N}{16 + N}$
chezbgone2 2014-03-15 19:12:14
$\dfrac{6}{10}\cdot\dfrac{n}{n+16}$
bgu 2014-03-15 19:12:14
both red => (3/5)(N/16+N)
borntobeweild 2014-03-15 19:12:14
3/5*N/(N+16)
samuelczhao 2014-03-15 19:12:14
3/5 * N/N+16
1023ong 2014-03-15 19:12:14
6/10 * N/(N+16)
DPatrick 2014-03-15 19:12:18
There's a $\dfrac35$ chance of drawing a blue ball from the first urn.
DPatrick 2014-03-15 19:12:24
There's a $\dfrac{N}{N+16}$ chance of drawing a blue ball from the second urn.
DPatrick 2014-03-15 19:12:29
So there's a $\dfrac{3N}{5(N+16)}$ of drawing blue balls from both urns.
DPatrick 2014-03-15 19:12:40
So what is the probability of drawing balls of the same color?
niraekjs 2014-03-15 19:13:12
This is the equation we can write $$ \frac{4}{10}\cdot\frac{16}{16+N}+\frac{6}{10}\cdot\frac{N}{16+N}=\frac{29}{50}. $$
mathtastic 2014-03-15 19:13:12
sum of them=58/100
shreshrej 2014-03-15 19:13:12
Add them up!
bluephoenix 2014-03-15 19:13:12
29/100
popcorn1 2014-03-15 19:13:12
0.58
Lightningwing 2014-03-15 19:13:12
(32+3n)/5(16+n)
Tarin57 2014-03-15 19:13:12
32 + 3N / 5n + 80
DPatrick 2014-03-15 19:13:26
Right, since these are exclusive cases. we add the above probabilities: $$\dfrac{32}{5(N+16)}+\dfrac{3N}{5(N+16)} = \dfrac{32+3N}{5(N+16)}.$$
DPatrick 2014-03-15 19:13:33
Thus we have the equation $\dfrac{32+3N}{5(N+16)} = 0.58 = \dfrac{29}{50}.$
jigglypuff 2014-03-15 19:13:57
cross multiply
6stars 2014-03-15 19:13:57
Cross mulitply, get N = 144
ViolinNinja256 2014-03-15 19:13:57
so now we can cross multiply
cellobix 2014-03-15 19:13:57
Cross-multiply.
picuberoot 2014-03-15 19:13:57
Solve for N!
Rectangle_Square 2014-03-15 19:13:57
cross multiply and solve
DPatrick 2014-03-15 19:14:02
Right, it's straightforward algebra from here.
DPatrick 2014-03-15 19:14:07
Multiply both sides by 5 to get $\dfrac{32+3N}{N+16} = \dfrac{29}{10}.$
DPatrick 2014-03-15 19:14:14
Cross-multiply to get $320 + 30N = 29N + 464.$
MathStudent2002 2014-03-15 19:14:24
dli00105 2014-03-15 19:14:24
N=144
LightningX48 2014-03-15 19:14:24
n = 144!
Tommy2000 2014-03-15 19:14:24
answer = 144
vinayak-kumar 2014-03-15 19:14:24
$N=\boxed{144}$
DPatrick 2014-03-15 19:14:28
And yes, $N = 464 - 320 = \boxed{144}.$
DPatrick 2014-03-15 19:15:11
Again, a reminder: there are nearly 200 students here! So please don't feel bad if your comments don't get posted -- I can only post a small portion of what you guys type in. But please keep participating!
DPatrick 2014-03-15 19:15:20
On to #3:
DPatrick 2014-03-15 19:15:25
3. Find the number of rational numbers $r$, $0 < r < 1$, such that when $r$ is written as a fraction in lowest terms, the numerator and denominator have a sum of 1000.
DPatrick 2014-03-15 19:15:37
What do these fractions look like?
basketball8 2014-03-15 19:16:01
x/1000-x
bob2 2014-03-15 19:16:01
r/(1000-r)
chezbgone2 2014-03-15 19:16:01
$\dfrac{n}{1000-n}$
tRIG 2014-03-15 19:16:01
a/(1000-a) where a<500
Eudokia 2014-03-15 19:16:01
n/(1000-n)
hwl0304 2014-03-15 19:16:01
n/(1000-n)
TheEconomist 2014-03-15 19:16:01
n/1000-n
allaoc 2014-03-15 19:16:01
n/(1000-n)
DPatrick 2014-03-15 19:16:08
Right: they're of the form $\dfrac{n}{1000-n}$ where $0 < n < 500.$ (If $n \ge 500$, then the fraction is greater than 1.)
DPatrick 2014-03-15 19:16:25
What does it mean that it is in lowest terms?
alex31415 2014-03-15 19:16:49
gcd(n,1000-n)=1
ashgabat 2014-03-15 19:16:49
They have no common divisors
fz0718 2014-03-15 19:16:49
GCD(N,1000-N)=1
sat113 2014-03-15 19:16:49
n and 1000-n are relatively prime
mathman523 2014-03-15 19:16:49
gcf of n and 1000-n is 1
Bluedevil98 2014-03-15 19:16:49
relatively prime
VCheep 2014-03-15 19:16:49
n and 1000-n are relatively prime
DPatrick 2014-03-15 19:16:55
Exactly. It means that $n$ and $1000-n$ have no common factors.
DPatrick 2014-03-15 19:17:03
But what's equivalent to this, and simpler?
El_Ectric 2014-03-15 19:17:24
gcd(n,1000-n) => gcd(n,1000)=1 by euclidean algorihtm
mentalgenius 2014-03-15 19:17:24
n and 1000 have no common factors
Quadratic64 2014-03-15 19:17:24
n and 1000 are relatively prime
martian179 2014-03-15 19:17:24
n coprime with 1000
DrMath 2014-03-15 19:17:24
Euclidean algorithm -----------> gcd(a,1000)=1
sjain 2014-03-15 19:17:24
no common factors between 1000 and n
ingridzhang97 2014-03-15 19:17:24
n and 1000 are relatively prime
DPatrick 2014-03-15 19:18:05
Right. Note that $\gcd(n,1000-n) = \gcd(n,1000)$, so $n$ and 1000 cannot have any common factors. (That's one step of the Euclidean Algorithm for finding greatest common divisor. Or, you can note that if $n$ and $1000-n$ are both multiples of the some number, then so is their sum of 1000.)
claudeaops 2014-03-15 19:18:20
n is not divisible by 2 or 5
Verjok 2014-03-15 19:18:20
n is not divisible by 2 or 5
bengals 2014-03-15 19:18:20
2 or 5 does not divide n
Tan 2014-03-15 19:18:20
n doesnt not have any prime factors of 2 and 5
DPatrick 2014-03-15 19:18:37
Right, going one step further, $n$ is relatively prime to 1000 if and only if $n$ has no prime factors of 2 or 5.
DPatrick 2014-03-15 19:18:47
So we need to count how many numbers in $0 < n < 500$ have no factors of 2 or 5.
DPatrick 2014-03-15 19:19:25
A lot of you are saying fancy things like "totient" or "Euler's Formula" or "inclusion-exclusion", but it's much simpler than that.
DPatrick 2014-03-15 19:19:31
How many odd numbers are between 0 and 500?
ninjataco 2014-03-15 19:19:43
250
summitwei 2014-03-15 19:19:43
250
poweroftwo 2014-03-15 19:19:43
250
minimario 2014-03-15 19:19:43
250
navygoat123 2014-03-15 19:19:43
250
popcorn1 2014-03-15 19:19:43
250
CornSaltButter 2014-03-15 19:19:43
250.
DPatrick 2014-03-15 19:19:47
There are 250 odd integers between 0 and 500.
DPatrick 2014-03-15 19:20:03
And how many of them do we have to throw away because they are (odd) multiples of 5?
bellyflop 2014-03-15 19:20:21
1/5 aof them are div by 5
ViolinNinja256 2014-03-15 19:20:21
1/5th are divisible by 5
forthegreatergood 2014-03-15 19:20:21
a fifth of these are multiples of 5 so 250-50=200 (ANSWER)
aimingforimo 2014-03-15 19:20:21
50
picuberoot 2014-03-15 19:20:21
50
kenneth102099 2014-03-15 19:20:21
50
darthvader1521 2014-03-15 19:20:21
50
joshualee2000 2014-03-15 19:20:21
50
samuelczhao 2014-03-15 19:20:21
50
raptorw 2014-03-15 19:20:21
50
DPatrick 2014-03-15 19:20:26
Exactly 50: 1*5, 3*5, 5*5, ..., up to 99*5.
DPatrick 2014-03-15 19:20:31
So that leaves $250 - 50 = 200$ values of $n$, and thus $\boxed{200}$ such rational numbers.
DPatrick 2014-03-15 19:20:54
OK, on to #4:
DPatrick 2014-03-15 19:20:58
4. Jon and Steve ride their bicycles along a path that parallels two side-by-side train tracks running in the east/west direction. Jon rides east at 20 miles per hour, and Steve rides west at 20 miles per hour. Two trains of equal lengths, traveling in opposite directions at constant but different speeds, each pass the two riders. Each train takes exactly 1 minute to go past Jon. The westbound train takes 10 times as long as the eastbound train to go past Steve. The length of each train is $\frac{m}{n}$ miles, where $m$ and $n$ are relatively prime positive integers. Find $m+n.$
DPatrick 2014-03-15 19:21:11
Yikes, that's a lot of words. With a wordy problem like this, be sure to read it carefully.
Bluedevil98 2014-03-15 19:21:24
d=rt
sat113 2014-03-15 19:21:24
d=rt
cellobix 2014-03-15 19:21:24
d=rt
DPatrick 2014-03-15 19:21:40
Indeed, the equation (distance) = (rate)(time), or $d=rt$ for short, is probably going to be our main tool in this problem.
DPatrick 2014-03-15 19:22:24
Some of you might find that drawing a picture is helpful to keep track of which train is which and which cyclist is which. If you want to do that, that's great. I didn't in fact draw one myself.
ninjataco 2014-03-15 19:22:29
I would like to define several variables.
Bluedevil98 2014-03-15 19:22:29
assign variables
DPatrick 2014-03-15 19:22:39
Sure, let's start by assigning variables to the various unknown quantities.
DPatrick 2014-03-15 19:22:43
Let:
$L$ be the length of each train in miles
$v_E$ be the speed of the eastbound train in miles per hour
$v_W$ be the speed of the westbound train in miles per hour
$t$ be the time that it takes the eastbound train to pass Steve (in hours)
(The time that the westbound train takes to pass Steve is then $10t$.)
DPatrick 2014-03-15 19:23:03
This is everything that's unknown in the problem as far as I can tell. And remember that we're trying to find $L.$
DPatrick 2014-03-15 19:23:18
OK, now how can we use the data given in the problem?
DPatrick 2014-03-15 19:23:56
I should mention that there are lots of different ways you could assign variables and set up equations...I'm just going to present the method that seemed most natural to me. (And it turned out to work nice in the end.)
DPatrick 2014-03-15 19:24:12
What equation(s) can we write for "Each train takes exactly 1 minute to go past Jon"?
DPatrick 2014-03-15 19:24:51
Let's focus on the eastbound train first.
LightningX48 2014-03-15 19:25:22
1 minute * v_e = 1/3 mile + L
DPatrick 2014-03-15 19:25:43
That's right, but let's try to keep all our units in miles and hours.
DPatrick 2014-03-15 19:25:50
To pass Jon in 1 minute, how much distance does the eastbound train have to cover?
bellyflop 2014-03-15 19:26:15
1/3+L
mdlu 2014-03-15 19:26:15
L+1/3 mile
danusv 2014-03-15 19:26:15
1/3+L
hamup1 2014-03-15 19:26:15
1/3+L miles
Tommy2000 2014-03-15 19:26:15
L+1/3
DPatrick 2014-03-15 19:26:39
Right. Jon bikes $\frac13$ mile in 1 minute, going in the same direction as the train. So the back end of the train to be even with Jon 1 minute after the front of the train is even with Jon. Thus it has to go $L + \frac13$ miles in that 1 minute.
claudeaops 2014-03-15 19:27:14
L+1/3=v_e*1/60
DPatrick 2014-03-15 19:27:31
Exactly, but let's get rid of the fraction. In an hour, the eastbound train travels $60L + 20$ miles. Thus its velocity is $$v_E = 60L + 20.$$
shreshrej 2014-03-15 19:27:57
Now for the westbound train!
DPatrick 2014-03-15 19:28:19
Right. You can probably guess the answer by symmetry, but what's the velocity of the westbound train, given that it also takes 1 minute to pass Jon?
hwl0304 2014-03-15 19:28:45
60L-20?
ninjataco 2014-03-15 19:28:45
$L - 1/3 = \frac{1}{60}v_W \implies 60L - 20 = v_W$
othuum0149 2014-03-15 19:28:45
60L-20
vinayak-kumar 2014-03-15 19:28:45
$60L-20$
allaoc 2014-03-15 19:28:45
60L - 20
CornSaltButter 2014-03-15 19:28:45
60L-20?
mathwizard888 2014-03-15 19:28:45
60L-20
dantx5 2014-03-15 19:28:45
60L-20
Eudokia 2014-03-15 19:28:45
v_W=60L-20
picuberoot 2014-03-15 19:28:45
v_w = 60L - 20
DPatrick 2014-03-15 19:28:59
Right. Since Jon has moved $\frac13$ mile in the opposite direction, the westbound train only has to go $L - \frac13$ miles in 1 minute.
DPatrick 2014-03-15 19:29:12
So in an hour, it travels $60L - 20$ miles. Its velocity is thus $$v_W = 60L - 20.$$
DPatrick 2014-03-15 19:29:24
Let's recap. Now we have:
$L$ is the length of each train in miles
The eastbound train is moving at a speed of $60L + 20$ miles per hour, and passes Steve in time $t$.
The westbound train is moving at a speed of $60L - 20$ miles per hour, and passes Steve in time $10t$.
DPatrick 2014-03-15 19:29:50
So let's try to write the same equations for passing Steve for each train.
DPatrick 2014-03-15 19:30:12
How far does the eastbound train have to travel to pass Steve?
DPatrick 2014-03-15 19:30:52
Remember, it has to go a distance of $L$ minus however long Steve has traveled in time $t$. (Steve is moving opposite the train, shortening the necessary distance.)
othuum0149 2014-03-15 19:31:16
L-20t
bellyflop 2014-03-15 19:31:20
L-20t
hwl0304 2014-03-15 19:31:20
L-20t?
DPatrick 2014-03-15 19:31:26
Right. Steve has moved $20t$ miles in time $t$.
DPatrick 2014-03-15 19:31:36
Thus the train has to travel $L - 20t$ miles in time $t$.
DPatrick 2014-03-15 19:31:42
What equation does that allow us to write?
mdlu 2014-03-15 19:32:22
L-20t=(60L+20)t
Eudokia 2014-03-15 19:32:22
L-20t=(60L+20)t
Tarin57 2014-03-15 19:32:22
L - 20t = (60L + 20)(t)
navygoat123 2014-03-15 19:32:22
L-20t/t = 60L-20
DPatrick 2014-03-15 19:32:30
We can use the usual $d = rt$ equation to write
$$ L - 20t = (60L + 20)(t).$$
DPatrick 2014-03-15 19:32:42
How about the westbound train?
LightningX48 2014-03-15 19:33:03
And for the westbound train, it has to go L + 200t miles
DPatrick 2014-03-15 19:33:24
Right: for the westbound train, the signs are reversed, and $t$ gets replaced by $10t$.
cellobix 2014-03-15 19:33:30
The train travels L+200t, so L+200t=(60t-20)(10t)
DPatrick 2014-03-15 19:34:01
Again, though, in words:
DPatrick 2014-03-15 19:34:15
The westbound train takes $10t$ hours to pass Steve.
DPatrick 2014-03-15 19:34:35
It has to travel its length $L$ plus the distance Steve has traveled. (Steve is moving in the same direction as the train, thus lengthening the distance necessary.)
DPatrick 2014-03-15 19:34:41
But Steve has moved $200t$ miles in time $10t$.
DPatrick 2014-03-15 19:34:49
Thus the train has to travel $L + 200t$ miles in time $10t$.
DPatrick 2014-03-15 19:34:54
Again using $d=rt$, we have $$L + 200t = (60L - 20)(10t).$$
LightningX48 2014-03-15 19:35:01
yeay two equations two variables; let's solve!
popcorn1 2014-03-15 19:35:01
combine the two equations
DPatrick 2014-03-15 19:35:09
Indeed, now we have a system of two equations in two variables:
\begin{align*}
L - 20t &= (60L + 20)t, \\
L + 200t &= (60L - 20)(10t).
\end{align*}
DPatrick 2014-03-15 19:35:27
Remember: our goal is to solve for $L.$ How do we continue?
VietaFan 2014-03-15 19:35:47
isolate L on one side
DPatrick 2014-03-15 19:36:14
As usual, there are tons of ways to solve "two equations in two variables", but I think the most efficient is to isolate $L$ in each of them, because something nice happens.
DPatrick 2014-03-15 19:36:22
We can rearrange slightly to get
\begin{align*}
L &= t(60L+40), \\
L &= t(600L - 400).
\end{align*}
swamih 2014-03-15 19:36:44
the t's cancel out
Quadratic64 2014-03-15 19:36:44
equate them and the t's cancel
mathtastic 2014-03-15 19:36:44
set equal divide by t
room456 2014-03-15 19:36:47
60L+40=600L-400
ninjataco 2014-03-15 19:36:47
$600L - 400 = 60L + 40$
DPatrick 2014-03-15 19:37:09
Right: equating them and dividing by $t$ just leaves $60L + 40 = 600L - 400.$
shreshrej 2014-03-15 19:37:21
Simplify for L = 22/27, so 22+27= 49
raptorw 2014-03-15 19:37:21
L=22/27
bellyflop 2014-03-15 19:37:21
22/27=L so our answer is 49.
joshxiong 2014-03-15 19:37:28
$540L = 440 \implies L=\frac{22}{27}$
dli00105 2014-03-15 19:37:28
22/27 so 49
DPatrick 2014-03-15 19:37:34
Yes. This simplifies to $440 = 540L$, so $L = \dfrac{44}{54} = \dfrac{22}{27}.$ Hence the answer is $22 + 27 = \boxed{049}.$
DPatrick 2014-03-15 19:38:08
It wasn't super-hard if you defined your variables carefully and just applied $d=rt$ consistently throughout.
DPatrick 2014-03-15 19:38:15
It was a bit tedious, though.
DPatrick 2014-03-15 19:38:23
On to #5:
DPatrick 2014-03-15 19:38:27
5. Let the set $S = \{P_1,P_2,\ldots,P_{12}\}$ consist of the twelve vertices of a regular 12-gon. A subset $Q$ of $S$ is called communal if there is a circle such that all points of $Q$ are inside the circle, and all the points of $S$ not in $Q$ are outside of the circle. How many communal subsets are there? (Note that the empty set is a communal subset.)
DPatrick 2014-03-15 19:38:46
Here a picture that might help clarify things:
DPatrick 2014-03-15 19:38:50
bellyflop 2014-03-15 19:39:11
only way to get a communal set is to have consecutive vertices
DrMath 2014-03-15 19:39:11
after testing some communal subsets we have that the points must be consecutive
mathtastic 2014-03-15 19:39:11
the word "communal" strongly implies that all verticies are next to each other
vinayak-kumar 2014-03-15 19:39:11
Aren't these just consecutive points
picuberoot 2014-03-15 19:39:11
For the circle to only contain points in Q, the points all have to be next to eachother. Casework!!
DPatrick 2014-03-15 19:39:32
How do we know that? Why is it that the communal sets are exactly the sets of consecutive points around the 12-gon?
shreshrej 2014-03-15 19:39:52
Hmm, start drawing some circles?
Wolstenholme 2014-03-15 19:39:52
circles intersect at at most two points
DPatrick 2014-03-15 19:40:09
Exactly: look at what happens if we draw a circle to carve out a communal subset:
DPatrick 2014-03-15 19:40:15
DPatrick 2014-03-15 19:40:30
The points that are inside the blue circle are consecutive -- it's a bunch of points from the 12-gon in a row, with none skipped.
DPatrick 2014-03-15 19:40:50
And we know that any intersecting circle (that's not tangent) intersects the circle of our 12-gon in exactly two points, and one of the two arcs between those two points will lie inside the new circle, and the other arc will lie outside the new circle.
DPatrick 2014-03-15 19:41:23
So we can always only get an arc-full of points in our communal set. And we can get any arc we want if we take a big enough circle.
DPatrick 2014-03-15 19:41:44
So we just need to count the subsets of consecutive points. How?
wangt 2014-03-15 19:42:20
so 12*11 (because subsets with 1-11 points all have 12) + 1 for all 12 + 1 for none?
bobthesmartypants 2014-03-15 19:42:20
casework on 0 points and 12 points, and 1->11 points
poweroftwo 2014-03-15 19:42:20
there's 12 for each number of elements 1-11
danusv 2014-03-15 19:42:20
12*11 + 1 + 1
allaoc 2014-03-15 19:42:20
Define a starting point for sets from 1 to 11 elements, and there is just one for 12
room456 2014-03-15 19:42:20
casework bash: 1 (empty) + 1 (full) + 11*12 (any other case)
summitwei 2014-03-15 19:42:20
for 1-11 points there are 12 ways
DPatrick 2014-03-15 19:42:27
Right!
checkmate1021 2014-03-15 19:42:36
there are 12 possible "starting" location for each length of consecutive points, and 11 possible lengths, so 11(12) = 132
DPatrick 2014-03-15 19:42:58
Exactly -- we can get a "nontrivial" subset (non-empty and not all 12 points) by picking the "first" point (say, counting counterclockwise) and then picking the number of additional number of points to add. This number of additional points can be from 0 to 10 inclusive.
DPatrick 2014-03-15 19:43:12
So that gives us $12 \cdot 11 = 132$ consecutive subsets with a "start" point plus 0 through 10 additional points.
mathwizard888 2014-03-15 19:43:30
adding the empty set and whole set gives 134 ans
ninjataco 2014-03-15 19:43:30
So it's just $1 + 1 + 12(11) = \boxed{134}$.
Tommy2000 2014-03-15 19:43:30
132+2 = 134
Satyaprakash2009rta 2014-03-15 19:43:30
132+2=134
ninjataco 2014-03-15 19:43:30
Then the empty set and full subset add two more
hjl00 2014-03-15 19:43:30
132+1+1=134
guilt 2014-03-15 19:43:30
and also the empty and full sets so 134
raptorw 2014-03-15 19:43:30
two more cases of empty and full set
claudeaops 2014-03-15 19:43:30
132+1(empty set) + 1 (full set)
DPatrick 2014-03-15 19:43:51
Right: this doesn't yet include the empty set (if the blue circle is entirely inside the 12-gon) and the set of all 12 points (if the blue circle contains the entire 12-gon) that are also both communal.
DPatrick 2014-03-15 19:43:57
Adding the two trivial subsets (the empty set and the set of all 12 points) gives us a total of $132 + 2 = \boxed{134}$ subsets.
DPatrick 2014-03-15 19:44:22
OK, we're 1/3 of the way home! On to #6:
DPatrick 2014-03-15 19:44:27
6. The graphs of $y = 3(x-h)^2 + j$ and $y = 2(x-h)^2 + k$ have $y$-intercepts of 2013 and 2014, respectively, and each graph has two positive integer $x$-intercepts. Find $h.$
bengals 2014-03-15 19:45:00
expand first
mathtastic 2014-03-15 19:45:00
expand them both
swamih 2014-03-15 19:45:00
expand the binomials
DPatrick 2014-03-15 19:45:12
Perhaps we can use the equations in their current form, but we're probably more familiar with $y = ax^2 + bx + c$, so let's convert them into that form.
DPatrick 2014-03-15 19:45:18
The first equation becomes $y = 3x^2 - 6hx + (3h^2+j).$
DPatrick 2014-03-15 19:45:24
But what else do we know?
tRIG 2014-03-15 19:45:56
3h^2 + j = 2013 and 2h^2+k = 2014
pedronr 2014-03-15 19:45:56
3h^2+j=2013, 2h^2+j=2014
samuelczhao 2014-03-15 19:45:56
u know the y intercept
sat113 2014-03-15 19:45:56
3h^2+j=2013
cellobix 2014-03-15 19:45:56
3h^2 + j = 2013
DrMath 2014-03-15 19:45:56
the y intercept is 2013
blizzard10 2014-03-15 19:45:56
when y=2013, x=0
othuum0149 2014-03-15 19:45:56
y intercept of 2013, so 3h^2+j=2013
sparkles257 2014-03-15 19:45:56
3h^2+j = 2013
Ericaops 2014-03-15 19:45:56
when x-0, y is 2013
DPatrick 2014-03-15 19:46:03
Right: when we plug in $x=0$ we need to get $y = 2013$.
DPatrick 2014-03-15 19:46:13
Thus the first equation is just $y = 3x^2 - 6hx + 2013.$
DPatrick 2014-03-15 19:46:26
What about the second equation?
CornSaltButter 2014-03-15 19:46:50
it goes through (0,2014) so plug those values in
raptorw 2014-03-15 19:46:50
likewise, 2h^2+k=2014
vinayak-kumar 2014-03-15 19:46:50
we have the constant term to be 2014
shreshrej 2014-03-15 19:46:50
$2x^2-4hx+2014$
DPatrick 2014-03-15 19:46:56
By the same logic, the second equation is $y = 2x^2 - 4hx + 2014.$
DPatrick 2014-03-15 19:47:07
So to quickly recap, our equations are
\begin{align*}
y &= 3x^2 - 6hx + 2013, \\
y &= 2x^2 - 4hx + 2014.
\end{align*}
And we also have the data that each has two positive integer $x$-intercepts.
DPatrick 2014-03-15 19:47:14
What's a less-fancy way of saying that?
vinayak-kumar 2014-03-15 19:47:42
they have positive integer roots
MSTang 2014-03-15 19:47:42
Roots are pos ints
VietaFan 2014-03-15 19:47:42
the roots are positive integer values of x
swamih 2014-03-15 19:47:42
the roots of the equation are integers
sat113 2014-03-15 19:47:42
it has two positive integer roots
DPatrick 2014-03-15 19:47:57
Yes: "$x$-intercept" is just a fancy word for "root".
DPatrick 2014-03-15 19:48:09
(The $x$-intercepts are the values of $x$ that make $y=0.$)
DPatrick 2014-03-15 19:48:15
So we are given that the solutions to the two quadratics
\begin{align*}
0 &= 3x^2 - 6hx + 2013, \\
0 &= 2x^2 - 4hx + 2014,
\end{align*}
are all positive integers.
ViolinNinja256 2014-03-15 19:48:28
simplify
mathtastic 2014-03-15 19:48:37
divide first by 3 and second by 2
DPatrick 2014-03-15 19:48:46
Sure, we can simplify a little and divide by the quadratic coefficients of each and not change the roots:
\begin{align*}
0 &= x^2 - 2hx + 671, \\
0 &= x^2 - 2hx + 1007.
\end{align*}
DPatrick 2014-03-15 19:48:52
Now what?
dli00105 2014-03-15 19:49:30
prime factor 671 and 1007
brian22 2014-03-15 19:49:30
Vieta: sums of the roots are the same
joshxiong 2014-03-15 19:49:30
find the factors of $671$ and $1007$
ninjataco 2014-03-15 19:49:30
The sums of the roots are the same, look for numbers that multiply to get 671 and 1007 and have same sum
Eudokia 2014-03-15 19:49:30
Factor 671 and 1007
amudol 2014-03-15 19:49:30
prime factor 671 and/or 1007 to see what positive integer roots you can get
DPatrick 2014-03-15 19:49:44
Right. To say all this another way, now we can use Vieta's Formulas to get information about the roots of the two equations.
DPatrick 2014-03-15 19:49:52
In the first equation, the roots have product 671 and sum $2h.$
In the second equation, the roots have product 1007 and sum $2h.$
DPatrick 2014-03-15 19:50:14
Happily, 671 and 1007 don't have a lot of positive integer factors, and we're looking for a pair for each that have the same sum.
mathmaster2012 2014-03-15 19:50:35
671=61*11, 1007=19*53
room456 2014-03-15 19:50:35
671=61*11, 1007=19*53
cellobix 2014-03-15 19:50:35
671=11*61, 1007=19*53
mathwizard888 2014-03-15 19:50:35
11 and 61, 19 and 53
DPatrick 2014-03-15 19:50:41
The prime factorization of 671 is $11 \cdot 61,$ and the prime factorization of 1007 is $19 \cdot 53.$
guilt 2014-03-15 19:50:54
11+61 = 19+53 = 72 = 2h ; so h = 36
adamz 2014-03-15 19:50:54
they both sum to 72
acegikmoqsuwy2000 2014-03-15 19:51:04
11+61=19+53
DPatrick 2014-03-15 19:51:13
What do you know, those pairs of factors have the same sum of 72!
LightningX48 2014-03-15 19:51:35
yeay 2h = 72 h = 36 | 036 is ans
mdlu 2014-03-15 19:51:35
they both add up to 72 so $h=\boxed{36}.$
navygoat123 2014-03-15 19:51:35
h = 36
kquittman 2014-03-15 19:51:35
so h=36.
willwin4sure 2014-03-15 19:51:35
DPatrick 2014-03-15 19:51:44
So the roots of the first quadratic are 11 and 61, and the roots of the second quadratic are 19 and 53. And thus $2h = 11+61 = 19+53 = 72$, and our answer is $h = \boxed{036}.$
DPatrick 2014-03-15 19:52:08
That's all for page 1 of the contest. On to #7:
DPatrick 2014-03-15 19:52:14
7. Let $w$ and $z$ be complex numbers such that $|w| = 1$ and $|z| = 10.$ Let $\theta = \arg\left(\frac{w-z}{z}\right).$ The maximum possible value of $\tan^2\theta$ can be written as $\frac{p}{q},$ where $p$ and $q$ are relatively prime positive integers. Find $p+q.$ (Note that $\arg(w),$ for $w \not= 0,$ denotes the measure of the angle that the ray from 0 to $w$ makes with the positive real axis in the complex plane.)
DPatrick 2014-03-15 19:52:35
The AIME always seems to have one of this sort of problem, that's fairly easy if you know how to use the geometry of the complex plane, and nearly impossible if you don't.
DPatrick 2014-03-15 19:52:52
How can we get at the expression $\arg\left(\dfrac{w-z}{z}\right)?$
Wickedestjr 2014-03-15 19:52:55
Simplify (w-z)/z ?
DPatrick 2014-03-15 19:53:09
Yeah, $\dfrac{w-z}{z}$ is a bit hard to work with.
bobthesmartypants 2014-03-15 19:53:17
minimario 2014-03-15 19:53:17
$\frac{w}{z}-1$
PlatinumFalcon 2014-03-15 19:53:17
Split up into w/z-1
mdlu 2014-03-15 19:53:22
$\frac{w}{z}-1?$
raptorw 2014-03-15 19:53:22
w/z-1
Ericaops 2014-03-15 19:53:22
w/z-1
hjl00 2014-03-15 19:53:22
w/z-1
DPatrick 2014-03-15 19:53:26
Let's rewrite it as $\dfrac{w-z}{z} = \dfrac{w}{z} - 1.$
DPatrick 2014-03-15 19:53:31
What do we know about $\dfrac{w}{z}$?
claudeaops 2014-03-15 19:53:55
w/z has absolute value 1/10
hi how are you doing toda 2014-03-15 19:53:55
magnitude is 1/10
Tommy2000 2014-03-15 19:53:55
magnitude is 1/10?
othuum0149 2014-03-15 19:53:55
magnitude of 1/10
bunniesrcute 2014-03-15 19:53:55
norm is 1/10
DrMath 2014-03-15 19:53:55
has magnitude 1/10
TheStrangeCharm 2014-03-15 19:53:55
it has magnitude $\frac{1}{10}$
DPatrick 2014-03-15 19:54:03
Right. It's a point in the complex plane with magnitude $\left|\dfrac{w}{z}\right| = \dfrac{|w|}{|z|} = \dfrac{1}{10}.$
DPatrick 2014-03-15 19:54:37
Where do these points live on the complex plane?
PlatinumFalcon 2014-03-15 19:54:58
So it's on the circle centered at the origin with radius 0.1.
Nitzuga 2014-03-15 19:54:58
That's a circle with radius $1/10$
MSTang 2014-03-15 19:54:58
It can be any such point
csmath 2014-03-15 19:54:58
A circle radius 1/10
joshualee2000 2014-03-15 19:54:58
a circle of radius 1/10
pedronr 2014-03-15 19:54:58
circle with r=1/10
tiger21 2014-03-15 19:54:58
a circle with center at origin and radius 1/10
DPatrick 2014-03-15 19:55:03
Exactly. These points live on a circle of radius $\frac{1}{10}$ around the origin.
DPatrick 2014-03-15 19:55:07
So where do the points $\dfrac{w}{z} - 1$ live on the complex plane?
dli00105 2014-03-15 19:55:38
1/10 away freom -1
mathtastic 2014-03-15 19:55:38
that circle shifted 1 unit left
joshxiong 2014-03-15 19:55:38
circle translated to the left by 1
allaoc 2014-03-15 19:55:38
A circle of radius 1/10 centered at -1
checkmate1021 2014-03-15 19:55:38
circle radius 1/10 centered at (-1, 0)
DPatrick 2014-03-15 19:55:47
They live on a circle of radius $\frac{1}{10}$ centered at the real number $-1$ (on the negative real axis).
DPatrick 2014-03-15 19:55:52
DPatrick 2014-03-15 19:56:02
(This picture is not to scale -- the actual circle should be much smaller. But a picture drawn to scale would be too hard to use!)
DPatrick 2014-03-15 19:56:31
OK, so we're trying to maximize $\tan^2\theta$ where $\theta$ is the argument of some point on that circle. So what ray from the origin to the blue circle gives us an angle $\theta$ with the maximum possible $\tan^2\theta$?
tanuagg13 2014-03-15 19:56:58
the tangent ray
sat113 2014-03-15 19:56:58
the tangent line
guilt 2014-03-15 19:56:58
The tangent!
mentalgenius 2014-03-15 19:56:58
tan(gent) lol
ingridzhang97 2014-03-15 19:56:58
the tangent line
vinayak-kumar 2014-03-15 19:56:58
A tangent!
DPatrick 2014-03-15 19:57:15
Right! Remember that tangent is slope. Thus, we want the ray to be as far as possible away from the real axis. How can we get it as far away as possible from the real axis, and yet still intersect the circle? We pick the tangent line.
DPatrick 2014-03-15 19:57:22
DPatrick 2014-03-15 19:57:42
What is the tangent of that angle $\theta$?
negativebplusorminus 2014-03-15 19:58:25
Well, its cosine is 1/10.
DPatrick 2014-03-15 19:59:03
Is that true? That angle looks pretty close to 180 degrees to me...
hi how are you doing toda 2014-03-15 19:59:21
the sin is 1/10.
csmath 2014-03-15 19:59:21
Wait so sine=1/10
ashgabat 2014-03-15 19:59:21
Its sin is 1/10 no?
DPatrick 2014-03-15 19:59:45
Right, got it backwards. It's the sine that's 1/10: the sine is the ratio of the radius of the little circle to the distance to that circle's center.
DPatrick 2014-03-15 19:59:53
Drawing this little right triangle might help:
DPatrick 2014-03-15 19:59:58
dantx5 2014-03-15 20:00:19
cos = sqrt(99)/10
csmath 2014-03-15 20:00:19
Therefore its cos=sqrt(99)/10
Tommy2000 2014-03-15 20:00:19
cosine is sqrt99/10
DPatrick 2014-03-15 20:00:56
Well, technically we have $\cos\theta = -\frac{\sqrt{99}}{10}$ (the angle is between 90 and 180 degrees so the cosine is negative), but it doesn't matter since we're squaring later anyway.
pedronr 2014-03-15 20:01:12
and tan=-1/sqrt(99)
DPatrick 2014-03-15 20:01:36
Right: $\tan\theta = \dfrac{\sin\theta}{\cos\theta} = -\frac{1}{\sqrt{99}}.$
acegikmoqsuwy2000 2014-03-15 20:01:53
tan square is 1/99 then
buzzyun 2014-03-15 20:01:53
tan^2=1/99, so 1+99=100
mdlu 2014-03-15 20:01:53
$\frac{1}{99}$ so the answer is $\boxed{100}$
claudeaops 2014-03-15 20:01:53
the answer is 1/99 -> 100
summitwei 2014-03-15 20:01:53
so its 1/99 so the answer is 100
Eudokia 2014-03-15 20:02:01
so tan^2=1/99 => ans. is 100
DPatrick 2014-03-15 20:02:03
Right. Hence $\tan^2\theta = \dfrac{1}{99},$ and our answer is $1 + 99 = \boxed{100}.$
DPatrick 2014-03-15 20:02:50
Again, if you didn't know the geometry of complex numbers, you probably had little chance with this problem.
DPatrick 2014-03-15 20:02:58
Anyway, on to #8:
DPatrick 2014-03-15 20:03:02
8. The positive integers $N$ and $N^2$ both end in the same sequence of four digits $abcd$ when written in base 10, where digit $a$ is not zero. Find the three-digit number $abc.$
DPatrick 2014-03-15 20:03:29
What does it mean that $N$ and $N^2$ end in the same four digits?
guilt 2014-03-15 20:04:16
we see that N^2 - N = 10000x
tRIG 2014-03-15 20:04:16
n^2-n = 0 mod 10000
joshxiong 2014-03-15 20:04:16
N^2-N must be divisible by 10000
dli00105 2014-03-15 20:04:16
n==n^2 mod 10000
vinayak-kumar 2014-03-15 20:04:16
$N^2\equiv N\pmod{10000}$
aimingforimo 2014-03-15 20:04:16
n=n^2(mod 10000)
Wickedestjr 2014-03-15 20:04:16
They are equivalent mod 10,000
cellobix 2014-03-15 20:04:16
They are the same modulo 10 000
adi12 2014-03-15 20:04:16
N(N-1)= some multiple of 10000
tiger21 2014-03-15 20:04:16
N^2-N=0 mod 10000
DPatrick 2014-03-15 20:04:25
Right. It means that $N^2 - N$ is a multiple of 10000.
DPatrick 2014-03-15 20:04:33
But this means that $N(N-1)$ is a multiple of 10000.
DPatrick 2014-03-15 20:04:38
How can this happen?
henrikjb 2014-03-15 20:05:20
so either N or N-1 is a multiple of 16, and similarly for 625
pedronr 2014-03-15 20:05:20
N and N-1 are consecutive multiples of 625 and 16
DrMath 2014-03-15 20:05:20
N and N-1 are relatively prime so one of them is a multiple of 2^4 other 5^4.
TheEconomist 2014-03-15 20:05:20
factors must have 5^4 and 2^4
mathtastic 2014-03-15 20:05:20
one has to be a multiple of 5^4, other has to be multiple of 2^4
hamup1 2014-03-15 20:05:20
either N or N-1 is a multiple of 625
TheStrangeCharm 2014-03-15 20:05:20
if one is divisible by $2^4$, and the other is divisible by $5^4$.
infiniteturtle 2014-03-15 20:05:20
one is a multiple of 16, the other of 625
DPatrick 2014-03-15 20:05:37
We're getting slightly ahead of ourselves, but yes. Let me fill in the missing steps.
DPatrick 2014-03-15 20:05:52
One possibility, of course, is that $N$ could be a multiple of 10000 itself. But then its thousands digit $a$ would be 0.
DPatrick 2014-03-15 20:06:01
$N$ could also be 1 more than a multiple of 10000, so that $N-1$ is a multiple of 10000. But then $N$ would end in 0001 -- still not allowed.
DPatrick 2014-03-15 20:06:38
So that leaves the possibility that $N$ is a multiple of some factors of 10000, and $N-1$ is a multiple of the remaining factors of 10000.
DPatrick 2014-03-15 20:07:02
But we know that $10000 = 10^4 = 2^4 \cdot 5^4 = 16 \cdot 625.$ And we also know that $N$ and $N-1$ don't have any factors in common.
DPatrick 2014-03-15 20:07:13
So we must have either:
(I) $N$ is a multiple of 16 and $N-1$ is a multiple of 625
or
(II) $N$ is a multiple of 625 and $N-1$ is a multiple of 16
DPatrick 2014-03-15 20:07:41
If you know the Chinese Remainder Theorem, then you know that there is a unique value of $N$ (modulo 10000) that satisfies either of these conditions. (But it's not necessarily to know this to finish the problem.)
DPatrick 2014-03-15 20:08:03
Let's deal with case (II) first because it's easier. What value of $N$ satisfies case (II)?
othuum0149 2014-03-15 20:08:36
625
lovejj 2014-03-15 20:08:36
N=625
sat113 2014-03-15 20:08:36
625
Tommy2000 2014-03-15 20:08:36
625
firemike 2014-03-15 20:08:36
625?
cheeptricks 2014-03-15 20:08:36
625
sat113 2014-03-15 20:08:36
625 lol
checkmate1021 2014-03-15 20:08:36
N = 625
DPatrick 2014-03-15 20:08:51
Right: $N = 625$ is already the solution for case (II)! (Because $N-1 = 624 = 16 \cdot 39.$)
cellobix 2014-03-15 20:09:05
But this is 0625, which is not allowed
shreshrej 2014-03-15 20:09:05
but a is 0
ninjataco 2014-03-15 20:09:05
this doesn't work because a = 0
bookhunter7 2014-03-15 20:09:05
0625 isn't allowed though
DPatrick 2014-03-15 20:09:14
Indeed, this means that all numbers that satisfy (II) end in the digits 0625 -- which is still not allowed!
DPatrick 2014-03-15 20:09:23
So we must be looking for an $N$ with condition (I): $N$ is a multiple of 16 and $N-1$ is a multiple of 625.
DPatrick 2014-03-15 20:09:30
How do we find it?
bellyflop 2014-03-15 20:09:52
try multioles of 625 for n-1
dantx5 2014-03-15 20:09:52
try each multiple of 625 less than 10000
DPatrick 2014-03-15 20:10:15
That's a good strategy. There are lots of multiples of 16 to hunt through...what's easier is to look for $M$ such that $M+1$ is a multiple of 16 and $M$ is a multiple of 625. (So $N = M+1$ will be our answer.)
CornSaltButter 2014-03-15 20:10:27
notice that 625 is congruent to 1 mod 16
DPatrick 2014-03-15 20:10:34
Indeed, we already know that 625 itself is 1 more than a multiple of 16. How does that help?
dli00105 2014-03-15 20:11:01
try 15
shreshrej 2014-03-15 20:11:01
so if you multiply by 15, you will get 15 mod 16
ashgabat 2014-03-15 20:11:01
We try 15*625
MSTang 2014-03-15 20:11:01
15 * 625 + 1 is then 15 + 1 = 16 = 0 mod 16
negativebplusorminus 2014-03-15 20:11:01
15 of them will be -1 mod 16.
joshxiong 2014-03-15 20:11:01
so 15 * 625 = 15 = -1 (mod 16)
martian179 2014-03-15 20:11:01
15 * 625 + 1 = 0 mod 16
DPatrick 2014-03-15 20:11:15
That's exactly right.
DPatrick 2014-03-15 20:11:33
Or, to say it with less-fancy language: we know that $15 \cdot 625$ is 15 more than a multiple of 16. So $15 \cdot 625 + 1$ is 16 more than a multiple of 16, which again is a multiple of 16.
csmath 2014-03-15 20:11:50
By doing that we get N=9376 so answer = 937
navygoat123 2014-03-15 20:11:50
N-1 is 9375 so N is 9376 so abc = 937
adi12 2014-03-15 20:11:50
so its 9376 and 9375!
urmilla 2014-03-15 20:11:50
N-1 = 9375, so N = 9376
acegikmoqsuwy2000 2014-03-15 20:11:50
9376
DPatrick 2014-03-15 20:11:59
Thus $M = 15 \cdot 625 = 9375$, and hence $N = M+1 = 9376$ is the number we're looking for.
DPatrick 2014-03-15 20:12:06
(You can check that $9376^2 = 87909376.$)
DPatrick 2014-03-15 20:12:19
So our answer is the first three digits of $N$, or $\boxed{937}$.
DPatrick 2014-03-15 20:12:35
On to #9:
DPatrick 2014-03-15 20:12:39
9. Let $x_1 < x_2 < x_3$ be the three real roots of the equation $$\sqrt{2014}x^3 - 4029x^2 + 2 = 0.$$Find $x_2(x_1+x_3).$
checkmate1021 2014-03-15 20:13:10
vieta
LightningX48 2014-03-15 20:13:10
Vieta's!
bellyflop 2014-03-15 20:13:10
vieta
minimario 2014-03-15 20:13:10
Vieta's Formulae
guilt 2014-03-15 20:13:10
Vieta's!
hwl0304 2014-03-15 20:13:10
Vietas??
Tommy2000 2014-03-15 20:13:10
vieta's formula?
DPatrick 2014-03-15 20:13:21
Good idea. Since we want an expression that involves the roots, Vieta's Formulas are probably useful for this problem.
DPatrick 2014-03-15 20:13:38
(If you don't know what Vieta's Formulas are, you can look them up after the Math Jam, or ask on the message board...or try to figure them out on your own!)
DPatrick 2014-03-15 20:13:45
Let's go ahead and list the Vieta's Formulas for this equation.
DPatrick 2014-03-15 20:13:50
\begin{align*}
x_1 + x_2 + x_3 &= \dfrac{4029}{\sqrt{2014}}, \\[2ex]
x_1x_2 + x_2x_3 + x_3x_1 &= 0, \\[2ex]
x_1x_2x_3 &= -\dfrac{2}{\sqrt{2014}}.
\end{align*}
DPatrick 2014-03-15 20:14:03
That quantity that we're looking for...which does it resemble?
mishai 2014-03-15 20:14:27
using the second we need to find -x3x1
raptorw 2014-03-15 20:14:27
so what we're looking for is -x_3*x_1
mentalgenius 2014-03-15 20:14:27
second
vinayak-kumar 2014-03-15 20:14:27
The second equations
pedronr 2014-03-15 20:14:27
the second one
ViolinNinja256 2014-03-15 20:14:27
the 2nd equation
danusv 2014-03-15 20:14:27
so x2(x1+x3)=-x1x3
kcui 2014-03-15 20:14:27
the second formula!
room456 2014-03-15 20:14:27
2nd one
DPatrick 2014-03-15 20:14:35
What we want to find resembles the middle one quite a bit.
DPatrick 2014-03-15 20:14:42
Indeed, $x_2(x_1 + x_3) = -x_1x_3$ by the middle formula.
DPatrick 2014-03-15 20:14:54
And what does that resemble?
Tuxianeer 2014-03-15 20:15:11
third one
mathtastic 2014-03-15 20:15:11
third one!!
tiger21 2014-03-15 20:15:11
the third
samuelczhao 2014-03-15 20:15:11
last one
mikechen 2014-03-15 20:15:11
the last equation
Nitzuga 2014-03-15 20:15:11
the third
firemike 2014-03-15 20:15:11
the third one!
Acstar00 2014-03-15 20:15:11
the third equation
niraekjs 2014-03-15 20:15:11
the third one!!!
DPatrick 2014-03-15 20:15:22
Right! Then $-x_1x_3 = \dfrac{2}{x_2\sqrt{2014}}$ by the third formula.
DPatrick 2014-03-15 20:15:29
So the quantity we're looking for as our final answer is just $\dfrac{2}{x_2\sqrt{2014}}.$
DPatrick 2014-03-15 20:15:43
Thus, if we can find the middle root of the cubic, we're done.
DPatrick 2014-03-15 20:16:08
OK, that's probably as far as Vieta's Formulas will get us for now. Let's go back and look at our cubic.
DPatrick 2014-03-15 20:16:26
What do you notice about it that looks notable?
raptorw 2014-03-15 20:16:44
4029 is really close to 2014*2
bookhunter7 2014-03-15 20:16:44
4029=2*2014+1
Tuxianeer 2014-03-15 20:16:44
4029=2*2014+1
raptorw 2014-03-15 20:16:44
4029=2014*2+1
guilt 2014-03-15 20:16:44
4029 = 2*2014 + 1
joshxiong 2014-03-15 20:16:44
4029 = 2014 * 2 + 1
MSTang 2014-03-15 20:16:44
A bunch of 2014s
cellobix 2014-03-15 20:16:44
4029 is almost 2*2014
Eudokia 2014-03-15 20:16:44
4029=2014*2+1
DPatrick 2014-03-15 20:16:58
Yeah, you should notice the weird coefficients.
DPatrick 2014-03-15 20:17:02
And in particular, $4029 = 2(2014) + 1.$
henrikjb 2014-03-15 20:17:18
Substitute in $\sqrt{2014}=a$
vinayak-kumar 2014-03-15 20:17:25
Make a substitution $y=\sqrt{2014}$
MSTang 2014-03-15 20:17:29
Let $n = \sqrt{2014}$
DPatrick 2014-03-15 20:17:36
That's a great idea. I don't like carrying these big numbers around: let's set $a = \sqrt{2014}$.
DPatrick 2014-03-15 20:17:40
Then our cubic is $$ax^3 - (2a^2 + 1)x^2 + 2 = 0.$$
DPatrick 2014-03-15 20:17:53
Cubics are hard to solve in general -- but is there an "obvious" root?
DPatrick 2014-03-15 20:18:12
In particular...what would make that "+2" cancel out?
nickduensing 2014-03-15 20:18:38
-2
navygoat123 2014-03-15 20:18:38
-2
cheeptricks 2014-03-15 20:18:38
-2
willwin4sure 2014-03-15 20:18:38
-2
ChenthuranA 2014-03-15 20:18:38
-2
DPatrick 2014-03-15 20:18:54
Indeed...and what value of $x$ would produce a "-2" term to cancel with the "+2"?
dli00105 2014-03-15 20:19:20
+-1/a
dantx5 2014-03-15 20:19:20
1/a
jimBriggs 2014-03-15 20:19:20
1/a
SuperSnivy 2014-03-15 20:19:20
x = 1/a
negativebplusorminus 2014-03-15 20:19:20
1/a
danzhi 2014-03-15 20:19:20
1/a
minimario 2014-03-15 20:19:20
$\frac{1}{a}$
DPatrick 2014-03-15 20:19:31
Right. If $x = \frac{1}{a}$, then the 2's would cancel...
Tuxianeer 2014-03-15 20:19:45
and so does everything
martian179 2014-03-15 20:19:49
and so does everything else
DPatrick 2014-03-15 20:19:54
...and the other terms would cancel too!
DPatrick 2014-03-15 20:20:03
Specifically,
$$ a\left(\frac{1}{a}\right)^3 - (2a^2+1)\left(\frac{1}{a}\right)^2 + 2 = \frac{1}{a^2} - 2 - \frac{1}{a^2} + 2 = 0.$$
samuelczhao 2014-03-15 20:20:19
so one solution is 1/ sqrt (2014)
kquittman 2014-03-15 20:20:19
so x=1/sqrt(2014)!
DPatrick 2014-03-15 20:20:27
Indeed, we've checked that $\dfrac{1}{a} = \dfrac{1}{\sqrt{2014}}$ is one of the three roots.
checkmate1021 2014-03-15 20:20:37
but how to prove that that is the middle root?
csmath 2014-03-15 20:20:37
So we can "guess" x2=1/sqrt(2014) however we don't know if this is actually the middle root but it does give us integer
DPatrick 2014-03-15 20:20:52
Good question! Since this cancels nicely with the radical in our final answer and leaves us with just 2, I'd be tempted to guess that this is the middle root. But the AIME writers are often tricky...
DPatrick 2014-03-15 20:21:10
Is this in fact the middle root? How can we be sure (or discover that it's not)?
Petaminx 2014-03-15 20:21:30
factor the original equation, then use quadratic formula to find the other 2 roots
ksun48 2014-03-15 20:21:30
well, we can find the other roots with quadratic formula.
ashgabat 2014-03-15 20:21:30
Well we could long divide >.<
vinayak-kumar 2014-03-15 20:21:30
Hmm, factor out the root?
Verjok 2014-03-15 20:21:30
use polynomial division
DPatrick 2014-03-15 20:21:45
Ick. This might work, but I sure don't want to do it.
hi how are you doing toda 2014-03-15 20:21:53
we know the smallest root has to be negative
DPatrick 2014-03-15 20:22:08
Aha...let's go back and talk to our friend Vieta again.
DPatrick 2014-03-15 20:22:38
The product of the three roots is negative, right? So since we've already found a positive root, we know that there's exactly one negative root and one other positive root.
DPatrick 2014-03-15 20:23:05
And can we tell if the root we've found is the larger or the smalller or the positive roots?
negativebplusorminus 2014-03-15 20:23:27
And the other two roots sum to 2Sqrt[2014] so the positive one has to be big.
Tuxianeer 2014-03-15 20:23:27
use sum
room456 2014-03-15 20:23:37
first equation of vietas
allaoc 2014-03-15 20:23:37
And the other positive root is larger because of the sum of roots
adi12 2014-03-15 20:23:37
yes because of the sum
genesis2 2014-03-15 20:23:37
yes, by looking at the sum, we know that the sum is 2 sqrt 2014
DPatrick 2014-03-15 20:23:56
Right! Using the first of Vieta's Formula, we see that the sum is about $2\sqrt{2014}$, which is a big number.
DPatrick 2014-03-15 20:24:25
Since we have one negative root, and the root we found is pretty tiny, we know that the third root must be relatively big (at least $2\sqrt{2014}$ or so).
mondi 2014-03-15 20:24:35
so it must be the middle root
csmath 2014-03-15 20:24:47
This means our answer is 002
dantx5 2014-03-15 20:24:47
ans is 002
samuelczhao 2014-03-15 20:24:47
so 1/ sqrt 2014 must be in the middle ,so u can just plug it in to the earlier equation
MSTang 2014-03-15 20:24:47
So the answer IS 2
adi12 2014-03-15 20:24:47
therefore the answer of 2 is proven
DPatrick 2014-03-15 20:24:57
Yes. We've proved that indeed $x_2 = \dfrac{1}{\sqrt{2014}},$ and our answer is $\dfrac{2}{\frac{1}{\sqrt{2014}}\cdot\sqrt{2014}} = \boxed{002}.$
DPatrick 2014-03-15 20:25:41
OK, on to the double-digit problems!
DPatrick 2014-03-15 20:25:47
10. A disk with radius 1 is externally tangent to a disk with radius 5. Let $A$ be the point where the disks are tangent, $C$ be the center of the smaller disk, and $E$ be the center of the larger disk. While the larger disk remains fixed, the smaller disk is allowed to roll along the outside of the larger disk until the smaller disk has turned through an angle of $360^\circ.$ That is, if the center of the smaller disk has moved to point $D$, and the point on the smaller disk that began at $A$ has now moved to point $B$, then $\overline{AC}$ is parallel to $\overline{BD}.$ Then $\sin^2(\angle BEA) = \frac{m}{n},$ where $m$ and $n$ are relatively prime positive integers. Find $m+n.$
mathtastic 2014-03-15 20:26:08
draw draw draw
bookhunter7 2014-03-15 20:26:08
DRAW IT
vinayak-kumar 2014-03-15 20:26:08
Draw a picture!!
bellyflop 2014-03-15 20:26:08
draw a picture
guilt 2014-03-15 20:26:11
Draw a diagram!
mishai 2014-03-15 20:26:11
diagram?
DPatrick 2014-03-15 20:26:17
Yeah, let's draw a picture with all this information:
DPatrick 2014-03-15 20:26:22
DPatrick 2014-03-15 20:26:40
What else would be helpful in this picture?
DrMath 2014-03-15 20:27:04
radii
Tommy2000 2014-03-15 20:27:04
draw EA and ED
Wickedestjr 2014-03-15 20:27:04
Lines connecting ED and AE
nickduensing 2014-03-15 20:27:04
Radius of larger circle
kquittman 2014-03-15 20:27:04
the radii of the circles.
pedronr 2014-03-15 20:27:04
connect CE, ED, and EB
minimario 2014-03-15 20:27:04
Drawing BE
2kev111 2014-03-15 20:27:04
de and ce
danzhi 2014-03-15 20:27:04
connect centers
DPatrick 2014-03-15 20:27:12
We'll almost certainly want to connect $E$ with $C$ and $D$. We'll also need $E$ connected with $B$ (since we eventually want to investigate $\angle BEA$:
DPatrick 2014-03-15 20:27:20
DPatrick 2014-03-15 20:27:34
What else do we know?
DPatrick 2014-03-15 20:28:02
We know a lot of lengths -- but I won't put them in the diagram just yet, they're easy to add later.
DPatrick 2014-03-15 20:28:09
What do we know about the angles in this diagram?
Petaminx 2014-03-15 20:28:16
bd parallel to ac
imak64 2014-03-15 20:28:16
ac parallel to bd
raptorw 2014-03-15 20:28:22
<BDE=<CED
Quadratic64 2014-03-15 20:28:22
BD // AC
joshxiong 2014-03-15 20:28:29
AC || BD so BDE = DEC
Deathranger999 2014-03-15 20:28:29
DEC and BDE are equal.
DPatrick 2014-03-15 20:28:53
Right. $\overline{AC}$ and $\overline{BD}$ are parallel, so the angles $BDD$ and $DEC$ are equal. Let's call them $\theta$:
DPatrick 2014-03-15 20:28:59
DPatrick 2014-03-15 20:29:21
Now what? How do we use the data about the rotation? Can we figure out $\theta$?
vinayak-kumar 2014-03-15 20:29:56
We can compare arc lengths
kenneth102099 2014-03-15 20:29:56
by circumference
sinani206 2014-03-15 20:29:56
from the arc length ratio
DPatrick 2014-03-15 20:30:28
It can be a little unclear what's going on here, and some evidence of that is the number of you telling me that $\theta$ is $72^\circ$. It's not, so let's investigate more closely.
DPatrick 2014-03-15 20:30:50
The blue arc of the big circle shown below is the amount of arc of the big circle that the little circle rolled along.
DPatrick 2014-03-15 20:30:56
DPatrick 2014-03-15 20:31:10
How does that correspond to an arc on the little circle?
DPatrick 2014-03-15 20:31:51
The little circle covers the arc shown in blue below: (Imagine the big circle is covered in blue paint, and the little circle picks up the paint as it rolls.)
DPatrick 2014-03-15 20:32:00
DPatrick 2014-03-15 20:32:26
So those two blue arcs have to have the same length.
vinayak-kumar 2014-03-15 20:33:17
Now we have $5\theta=2\pi-\theta$
csmath 2014-03-15 20:33:17
So 5/6 on the little circle = 1/6 of the big circle so we get theta=60
mathtastic 2014-03-15 20:33:17
it goes 2pi/12pi=1/6=60 degrees around so theta=60^o
DPatrick 2014-03-15 20:33:35
Right -- now you can perhaps "eyeball" $\theta$ from the picture, or we can set up an equation.
DPatrick 2014-03-15 20:33:42
The blue arc length on the big circle is $5\theta$ (if $\theta$ is in radians).
DPatrick 2014-03-15 20:33:58
The blue arc length on the little circle is $2\pi - \theta$. (It's all but $\theta$ of the little circle of radius 1.)
DPatrick 2014-03-15 20:34:04
So $5\theta = 2\pi - \theta.$
And thus $\theta = \frac{2}{6}\pi = \frac{\pi}{3}.$ (Or, if you prefer, 60 degrees.)
DPatrick 2014-03-15 20:34:39
So we know that angles $BDE$ and $DEA$ are $60^\circ$. How do we finish from here?
tRIG 2014-03-15 20:35:00
find BE with law of cosines
claudeaops 2014-03-15 20:35:00
law of cosines
DPatrick 2014-03-15 20:35:14
Certainly, we could finish with some fancy trigonometry, but we don't need it.
DPatrick 2014-03-15 20:35:30
(Not that fancy I suppose: Law of Cosines is not exactly uncommon.)
ninjataco 2014-03-15 20:35:56
find $\sin BEA$ by dropping an perpendicular from B
ViolinNinja256 2014-03-15 20:35:56
would drawing a 30-60-90 triangle work?
DPatrick 2014-03-15 20:36:10
Right. Just drop a perpendicular from $B$ down to $\overline{AE}$. Call the points it hits along the way $X$ and $Y$.
DPatrick 2014-03-15 20:36:17
DPatrick 2014-03-15 20:36:27
Now $BDX$ and $EXY$ are 30-60-90 triangles!
DPatrick 2014-03-15 20:36:54
And we can just chase the lengths we need to compute $\sin(\angle BEA).$
DPatrick 2014-03-15 20:37:16
What are the lengths in triangle $BDX$?
ChilledLemonade 2014-03-15 20:37:44
BD=1
googol.plex 2014-03-15 20:37:44
1, sqrt3, 2
cellobix 2014-03-15 20:37:44
1, 2, sqrt3
claudeaops 2014-03-15 20:37:44
XD=2, BX=√3
torquoiseworld 2014-03-15 20:37:44
bx= sqrt 3
tiger21 2014-03-15 20:37:44
1, sqrt3, 2
ptes77 2014-03-15 20:37:44
BD=1, DX=2, BX=sqrt3
Tommy2000 2014-03-15 20:37:44
1, sqrt3, 2
raptorw 2014-03-15 20:37:44
1, 2, \sqrt3
DrMath 2014-03-15 20:37:44
1, root3, 2
DPatrick 2014-03-15 20:37:48
$BD = 1$, so $BX = \sqrt3$ and $DX = 2$.
DPatrick 2014-03-15 20:37:58
Then what does $EXY$ tell us?
MSTang 2014-03-15 20:38:10
EX = 4
room456 2014-03-15 20:38:10
ex is 4
mishai 2014-03-15 20:38:20
2, 2sqrt3, 4
Tommy2000 2014-03-15 20:38:20
then 2, 2sqrt3, 4
mathtastic 2014-03-15 20:38:20
xe=6-2=4 so xey has sides 2 4 and 2sqrt 3 mainly ey=2
Deathranger999 2014-03-15 20:38:25
The ratio from big to small is 2:1.
DPatrick 2014-03-15 20:38:35
Right. Since $ED = 5+1 = 6$ and $DX = 2$, we have that $EX = 4$. Therefore $EY = 2$ and $XY = 2\sqrt3$.
DPatrick 2014-03-15 20:38:55
So $BE$ is the hypotenuse of right triangle $BYE$ with sides $BY = BX + XY = 3\sqrt3$ and $EY = 2.$
summitwei 2014-03-15 20:39:05
$BE=\sqrt{31}$
csmath 2014-03-15 20:39:05
This means we have BYE as 2, 3sqrt(3), sqrt(31)
gamjawon 2014-03-15 20:39:05
DPatrick 2014-03-15 20:39:15
Right: $BE = \sqrt{27 + 4} = \sqrt{31}.$
gamjawon 2014-03-15 20:39:41
So then $ \sin^2{\angle BEA}=\frac{27}{31} $?
joshxiong 2014-03-15 20:39:47
so $\sin^2(\angle BEA) = \frac{27}{31}$
Deathranger999 2014-03-15 20:39:47
Which means sine squared is 27/31.
room456 2014-03-15 20:39:47
so sin^2(angle BEA) = 27/31
DPatrick 2014-03-15 20:39:59
Exactly: we have all the sides of $BEY$, so $$\sin\angle BEA = \sin\angle BEY = \frac{BY}{BE} = \frac{3\sqrt3}{\sqrt{31}}.$$
DPatrick 2014-03-15 20:40:08
Thus $\sin^2\angle BEA = \dfrac{27}{31}$, and our answer is $27 + 31 = \boxed{058}$.
DPatrick 2014-03-15 20:40:39
On to the final third! Next is #11:
DPatrick 2014-03-15 20:40:45
11. A token starts at the point $(0,0)$ of an $xy$-coordinate grid and then makes a sequence of six moves. Each move is 1 unit in a direction parallel to one of the coordinate axes. Each move is selected randomly from the four possible directions and independently of the other moves. The probability the token ends at a point on the graph of $|y| = |x|$ is $\frac{m}{n},$ where $m$ and $n$ are relatively prime positive integers. Find $m+n.$
DPatrick 2014-03-15 20:41:09
There are a few different ways to approach this problem. But virtually all of them involve careful bookkeeping of the various paths.
DPatrick 2014-03-15 20:41:34
What does the graph of $|y| = |x|$ look like?
turkeybob777 2014-03-15 20:42:01
an X
aimingforimo 2014-03-15 20:42:01
y=x and y=-x
MSTang 2014-03-15 20:42:01
y=x and y=-x
ashgabat 2014-03-15 20:42:01
It looks like a big X
CornSaltButter 2014-03-15 20:42:01
The graphs of y=x and y=-x
mikechen 2014-03-15 20:42:01
the lines y=x and y=-x combined
kquittman 2014-03-15 20:42:01
y=x and y=-x.
Rectangle_Square 2014-03-15 20:42:01
v and upside down v
allaoc 2014-03-15 20:42:01
X
dasobson 2014-03-15 20:42:01
a big x
aprobs21 2014-03-15 20:42:01
X
DPatrick 2014-03-15 20:42:11
It's the union of the lines $y=x$ and $y=-x$. If you drew it, it'd be a big "X" through the origin.
DPatrick 2014-03-15 20:42:24
So one strategy is that we can separately look at the number of 6-move paths that end on either line.
DPatrick 2014-03-15 20:42:34
Of course, the origin $(0,0)$ is on both lines, so it'll be double-counted and we'll have to correct for it later.
DPatrick 2014-03-15 20:42:41
Let's start with paths landing on $y=x$. What do we know about such paths?
MSTang 2014-03-15 20:42:57
y-x = 0
mishai 2014-03-15 20:43:11
same number of up down and right left
gamjawon 2014-03-15 20:43:11
y-x=0!
sinani206 2014-03-15 20:43:11
number of moves up or left = number of moves down or right
DPatrick 2014-03-15 20:43:16
Exactly!
DPatrick 2014-03-15 20:43:28
We land on $y=x$ if and only if 3 of the 6 moves are ups or lefts, and 3 of the 6 moves are downs or rights.
DPatrick 2014-03-15 20:43:45
One way to see this is to keep track of the value of $y-x$ as the point moves. Moving up or left increases $y-x$ by 1, and moving down or right decreases $y-x$ by 1. To land on $y=x$, it has to finish at $y-x = 0.$
DPatrick 2014-03-15 20:44:24
Or more broadly, moving up or left moves "northwest" and moving down or right moves "southeast".
DPatrick 2014-03-15 20:44:32
So we need to count the number of 6-move paths in which 3 of the moves are U or L and 3 of the moves are D or R.
DPatrick 2014-03-15 20:44:43
How many are there? How can we count them?
MSTang 2014-03-15 20:45:19
$\dbinom{6}{3} \cdot 2^6$
torquoiseworld 2014-03-15 20:45:19
6choose 3 and each of the 6 moves have two options
tiger21 2014-03-15 20:45:19
6 choose 3 times 2^6
aprobs21 2014-03-15 20:45:19
6C3 2^6?
mentalgenius 2014-03-15 20:45:27
choose 3 from 6 for which are DR, rest are UL. Then you have 2^3 for each, to get 2^6 * (6 choose 3)
DPatrick 2014-03-15 20:45:31
Good!
DPatrick 2014-03-15 20:45:54
First, we can decide which 3 of the 6 moves are U or L. We can choose 3 of the 6 moves to be U or L in $\binom63 = 20$ ways. (The other 3 will then be D or R.)
DPatrick 2014-03-15 20:46:04
Now each move has two choices: if it's slotted as an "U or L" move, we must choose U or L, and if it's slotted as a "D or R" move, we must choose D or R.
samuelczhao 2014-03-15 20:46:18
1280
Deathranger999 2014-03-15 20:46:18
Which is 20 times 64=1280.
DPatrick 2014-03-15 20:46:35
Yes: calculating this out, there are $\binom63 \cdot 2^6 = 20 \cdot 64 = 1280$ paths ending on $y=x.$
sinani206 2014-03-15 20:46:52
and you get the same number for y = -x so now you just have to get rid of doubles
cellobix 2014-03-15 20:46:52
And same thing for the line y=-x
summitwei 2014-03-15 20:46:52
same for y=-x
mishai 2014-03-15 20:46:58
by symmetry, are there the same amount landing in y=-x?
DPatrick 2014-03-15 20:47:13
Right! For paths landing on $y=-x$, we could do the same computation, or just note that by the symmetry of the problem, there are also $1280$ such paths. (For example, take any path landing on $y=x$ and rotate it 90 degrees.)
DPatrick 2014-03-15 20:47:26
So that's $2(1280) = 2560$ paths that land on $y=x$ or $y=-x$...
mathtastic 2014-03-15 20:47:40
then subtract for overcount of orgin
pedronr 2014-03-15 20:47:40
but you're overcounting the origin
ninjataco 2014-03-15 20:47:40
the only overcount is (0,0)
SockFoot 2014-03-15 20:47:40
but then subtract the ones where it ends on the origin
mentalgenius 2014-03-15 20:47:40
but you doublecounted the origin
shreshrej 2014-03-15 20:47:40
But we cannot overcount 0,0! That can be bad
DPatrick 2014-03-15 20:47:49
...indeed, as mentioned earlier, we've double-counted paths that land on the origin. So we need to subtract these from our count.
DPatrick 2014-03-15 20:48:00
So let's count 6-move paths that end at $(0,0)$. What can we say about such paths?
sinani206 2014-03-15 20:48:30
up = down and left = right
speck 2014-03-15 20:48:30
R=L, U=D
shreshrej 2014-03-15 20:48:30
The number of up is equal to down, the number of left is equal to right
allaoc 2014-03-15 20:48:30
Up = down, left = right
sparkles257 2014-03-15 20:48:30
all cancel out L and R or D and U
ChilledLemonade 2014-03-15 20:48:30
Number of up= down left = right
claudeaops 2014-03-15 20:48:30
# of Ups=# of Downs, # of Lefts= # of Rights
MSTang 2014-03-15 20:48:30
numOfU = numOfD and numOfL = numOfR
LightningX48 2014-03-15 20:48:35
for every right, there is a left and for every up there is a down
VietaFan 2014-03-15 20:48:35
ups = downs, lefts = rights
DPatrick 2014-03-15 20:48:41
Yep. They must have an equal number of Us and Ds, and an equal number of Ls and Rs.
mentalgenius 2014-03-15 20:49:00
UUUDDD, UUDDRL, UDRRLL, or RRRLLL (permuatations of each of these)
shreshrej 2014-03-15 20:49:05
RRRLLL, UUUDDD, RRLLUD, RLUUDD
DPatrick 2014-03-15 20:49:31
Right, all such paths must be UUUDDD, UUDDLR, UDLLRR, or LLLRRR in some order. So there's some simple casework to count the number of ordering of each of these. (I don't know a slick way to do them all at once.)
DPatrick 2014-03-15 20:49:43
How many with UUUDDD in some order?
ViolinNinja256 2014-03-15 20:50:17
2 * 6C3 + 2*6C4*4C2
Deathranger999 2014-03-15 20:50:17
6C3.
sparkles257 2014-03-15 20:50:17
6C3
sophiazhi 2014-03-15 20:50:17
20
shreshrej 2014-03-15 20:50:17
6C3 or 20
guilt 2014-03-15 20:50:17
6C3 = 20
Descrip 2014-03-15 20:50:17
6 choose 3
ninjataco 2014-03-15 20:50:17
20
aimingforimo 2014-03-15 20:50:17
20
lovejj 2014-03-15 20:50:17
20
tiger21 2014-03-15 20:50:17
6C3=20
othuum0149 2014-03-15 20:50:17
20
room456 2014-03-15 20:50:17
6C3
ViolinNinja256 2014-03-15 20:50:17
6C3
TheEconomist 2014-03-15 20:50:17
6!/(3!3!)
hjl00 2014-03-15 20:50:17
20
DPatrick 2014-03-15 20:50:24
We just need to choose 3 of the 6 positions in which to place a U (and the other 3 then becomes Ds). So there are $\binom63 = 20$ such paths.
DPatrick 2014-03-15 20:50:48
Or you can think of it as $\dfrac{6!}{3!3!}$, where we're arranging 6 letters, with two letters each repeated 3 times. (It's the same number!)
cellobix 2014-03-15 20:51:01
Same thing for LLLRRR
kgator 2014-03-15 20:51:01
RRRLLL is also 20
mathtastic 2014-03-15 20:51:01
same for RRRLLL
mishai 2014-03-15 20:51:01
same for RRRLLL
ViolinNinja256 2014-03-15 20:51:01
same for RRRLLL
allaoc 2014-03-15 20:51:01
LLLRRR is the same
DPatrick 2014-03-15 20:51:08
Indeed, by symmetry, there are also 20 paths with LLLRRR in some order.
DPatrick 2014-03-15 20:51:13
How about paths with UUDDLR?
sparkles257 2014-03-15 20:51:42
6!/4
dli00105 2014-03-15 20:51:42
180
kquittman 2014-03-15 20:51:42
180
MSTang 2014-03-15 20:51:42
6 * 5 * (4 choose 2) = 180
sharonmath 2014-03-15 20:51:42
180
bakugon321 2014-03-15 20:51:42
6!/(2*2)
DrMath 2014-03-15 20:51:42
180
danusv 2014-03-15 20:51:42
6!/2!2!
firemike 2014-03-15 20:51:42
180
mathmaster2012 2014-03-15 20:51:42
6!/(2!2!)
joey8189681 2014-03-15 20:51:42
6!/2!2!
WalkerTesla 2014-03-15 20:51:42
180
DPatrick 2014-03-15 20:51:52
There are $6!$ ways to arrange 6 letters, but we divide by $2 \cdot 2$ because there are 2 Us and 2Ds. Thus there are $\dfrac{6!}{4} = 180$ paths with UUDDLR in some order.
DPatrick 2014-03-15 20:52:13
You could also count this as 6 ways to slot the L, then 5 ways to slot to R, then $\binom42$ ways to pick 2 of the remaining slots for U.
kgator 2014-03-15 20:52:29
same with RRLLUD
cellobix 2014-03-15 20:52:29
Same for UDLLRR
sinani206 2014-03-15 20:52:29
so 180 llrrud as well
sparkles257 2014-03-15 20:52:29
same for the other one by symmetry
firemike 2014-03-15 20:52:29
same for LLRRUD
kunsun 2014-03-15 20:52:29
also 180 for UDLLRR
claudeaops 2014-03-15 20:52:29
same for UDLLRR
dli00105 2014-03-15 20:52:29
same for RRLLUD
bakugon321 2014-03-15 20:52:29
Same for udrrll
DPatrick 2014-03-15 20:52:33
By the same reasoning there are also 180 paths with UDLLRR in some order.
DPatrick 2014-03-15 20:52:40
So that's a total of $20+180+180+20 = 400$ paths that end at the origin.
DPatrick 2014-03-15 20:52:48
Thus how many total paths that work for this problem?
sinani206 2014-03-15 20:53:06
2560-400 = 2160
sparkles257 2014-03-15 20:53:06
overcounted 400
sparkles257 2014-03-15 20:53:06
2160
sinani206 2014-03-15 20:53:06
2160
MSTang 2014-03-15 20:53:06
2160
DrMath 2014-03-15 20:53:06
2160
ashgabat 2014-03-15 20:53:06
2160
guilt 2014-03-15 20:53:06
2560-400 = 2160
WalkerTesla 2014-03-15 20:53:06
2160
ninjataco 2014-03-15 20:53:06
2160
DPatrick 2014-03-15 20:53:26
Right: we initially counted 2560 paths, but then discovered that we overcounted by 400. Thus there are $2560 - 400 = 2160$ paths that work.
DPatrick 2014-03-15 20:53:46
And to finish we need to compute the probability of such a path randomly occurring.
mathtastic 2014-03-15 20:54:06
2160/4^6
kgator 2014-03-15 20:54:06
in total there are 4096 cases
aprobs21 2014-03-15 20:54:06
Out of 4^6 total
cellobix 2014-03-15 20:54:06
4^6 total paths (4 options for each move)
sophiazhi 2014-03-15 20:54:06
2160/4^6
joey8189681 2014-03-15 20:54:06
2160/2^12
Quadratic64 2014-03-15 20:54:06
The total number of paths is 4^6
mishai 2014-03-15 20:54:06
divided by 4^6=4096
samuelczhao 2014-03-15 20:54:06
2160/4096 = 135 / 256
bob2 2014-03-15 20:54:06
4^6 TOTAL
Satyaprakash2009rta 2014-03-15 20:54:06
2160/4^6=135/256 so our answer is 135+256=391
DPatrick 2014-03-15 20:54:11
Right. Each of the 6 moves has 4 possibilities, so there are $4^6 = 4096$ total paths.
DPatrick 2014-03-15 20:54:21
Thus our probability is $\dfrac{2160}{4096}.$ This reduces to $\dfrac{135}{256},$ so our answer is $135 + 256 = \boxed{391}.$
DPatrick 2014-03-15 20:54:42
I'm going to take a 2-minute break to get a glass of water -- brb!
DPatrick 2014-03-15 20:56:28
OK, I'm back and refreshed!
DPatrick 2014-03-15 20:56:37
We're about at the 2-hour mark, and just 4 problems to go!
DPatrick 2014-03-15 20:56:42
12. Let $A = \{1,2,3,4\},$ and let $f$ and $g$ be randomly chosen (not necessarily distinct) functions from $A$ to $A.$ The probability that the range of $f$ and the range of $g$ are disjoint is $\frac{m}{n},$ where $m$ and $n$ are relatively prime positive integers. Find $m.$
DPatrick 2014-03-15 20:57:15
I would have preferred it if #11 and #12 were spaced a bit apart -- this is two bookwork-y counting problems in a row.
DPatrick 2014-03-15 20:57:36
But we have to deal with it. As with #11, the trickiest part is just keeping careful track of stuff.
DPatrick 2014-03-15 20:57:58
By the way, since many of you are asking: "disjoint" means that the ranges have no numbers in common.
firemike 2014-03-15 20:58:21
casework again!
bengals 2014-03-15 20:58:21
casework
claudeaops 2014-03-15 20:58:21
Split into cases based on the # of outputs f has
DPatrick 2014-03-15 20:58:32
That's what I did. We can do casework on the number of elements in the range of $f$.
joshxiong 2014-03-15 20:58:39
Note that |range f| cannot = 4
DPatrick 2014-03-15 20:58:57
Indeed, we can't use all 4 elements in the range for either function, because then there'd be none left for the other one.
DPatrick 2014-03-15 20:59:12
So there are 3 cases we have to work them. Let's just have at them.
DPatrick 2014-03-15 20:59:18
Case 1: the range of $f$ has 1 element.
DPatrick 2014-03-15 20:59:22
How many $f$'s have just 1 element in the range?
sparkles257 2014-03-15 20:59:40
4
checkmate1021 2014-03-15 20:59:40
4
Tommy2000 2014-03-15 20:59:40
4?
guilt 2014-03-15 20:59:40
4
kquittman 2014-03-15 20:59:40
4
SockFoot 2014-03-15 20:59:40
4
buzzyun 2014-03-15 20:59:40
4
mathmaster2012 2014-03-15 20:59:40
4
inductor 2014-03-15 20:59:40
4
MSTang 2014-03-15 20:59:40
4: f(x) = 1, 2, 3, 4
DPatrick 2014-03-15 20:59:44
Just 4: the four constant functions.
DPatrick 2014-03-15 20:59:54
For each of these, how many $g$'s are there with disjoint range?
claudeaops 2014-03-15 21:00:22
3^4
alex31415 2014-03-15 21:00:22
3^4=81
apple.singer 2014-03-15 21:00:22
3^4
dli00105 2014-03-15 21:00:22
81
aprobs21 2014-03-15 21:00:22
3^4
TheStrangeCharm 2014-03-15 21:00:22
3^4 = 81
SuperSnivy 2014-03-15 21:00:22
3^4
jimBriggs 2014-03-15 21:00:22
3^4
joshxiong 2014-03-15 21:00:22
3^4 = 81
DPatrick 2014-03-15 21:00:43
Right. Each element in $A$ can map via $g$ to one of the three numbers not in the range of $f$.
DPatrick 2014-03-15 21:01:00
So since we have 3 choice for each of the 4 elements in the domain of $g$, we have $3^4 = 81$ such $g$'s.
DPatrick 2014-03-15 21:01:06
So, how many pairs of functions in case 1?
sinani206 2014-03-15 21:01:28
324
aimingforimo 2014-03-15 21:01:28
324
aprobs21 2014-03-15 21:01:28
324
dli00105 2014-03-15 21:01:28
324
SockFoot 2014-03-15 21:01:28
81*4
kquittman 2014-03-15 21:01:28
324
mishai 2014-03-15 21:01:28
4*81=324
cellobix 2014-03-15 21:01:28
4*81=324
CornSaltButter 2014-03-15 21:01:28
81*4=324
summitwei 2014-03-15 21:01:28
4*81=324
ViolinNinja256 2014-03-15 21:01:28
4*81=324
DPatrick 2014-03-15 21:01:43
Right: there are 4 $f$'s and each $f$ has 81 possible $g$'s. Hence, there are $4 \cdot 81 = 324$ pairs of functions in case 1.
DPatrick 2014-03-15 21:01:53
Case 2: the range of $f$ has 2 elements.
DPatrick 2014-03-15 21:01:58
How many such $f$ are there?
Wolstenholme 2014-03-15 21:02:38
14*6 = 84
othuum0149 2014-03-15 21:02:38
84
DPatrick 2014-03-15 21:02:46
84 is correct -- let's see how we get it.
DPatrick 2014-03-15 21:03:07
An easier question is: how many choices for a 2-element range of $f$?
DPatrick 2014-03-15 21:03:22
(Without worrying yet about which elements map to which.)
VietaFan 2014-03-15 21:03:39
4C2= 6
firemike 2014-03-15 21:03:39
4 choose 2?
allaoc 2014-03-15 21:03:39
4C2 = 6
ChilledLemonade 2014-03-15 21:03:39
4C2=6
ninjataco 2014-03-15 21:03:39
6
bengals 2014-03-15 21:03:39
6
ashgabat 2014-03-15 21:03:39
6
Quadratic64 2014-03-15 21:03:39
4 choose 2 = 6
martian179 2014-03-15 21:03:39
4 choose 2 = 6
DPatrick 2014-03-15 21:03:57
Right: there are $\binom42 = 6$ choices for the range of $f$. (We just have to choose the 2 elements out of the 4 that we want in the range.)
DPatrick 2014-03-15 21:04:11
And then for each of these choices, how many functions are there that have this range?
Petaminx 2014-03-15 21:04:38
$2^4-2$
joshxiong 2014-03-15 21:04:38
2^4 but we have to subtract 2 since we do not want |range f| = 1 so 14 in total
tiger21 2014-03-15 21:04:38
4C1+4C2+4C3=4+6+4=14
raptorw 2014-03-15 21:04:38
2^4?
aimingforimo 2014-03-15 21:04:38
2^4-2=14
martian179 2014-03-15 21:04:38
2^4 - 2 (to remove the two constant functions)
DPatrick 2014-03-15 21:05:03
Right. There are $2^4 = 16$ functions total that map from $A$ to 2-elements, but we don't want to count the 2 constant functions -- we already dealt with those in Case 1!
DPatrick 2014-03-15 21:05:15
So there are $16 - 2 = 14$ functions $f$ with a given $2$-element range.
DPatrick 2014-03-15 21:05:56
Therefore, since there were 6 choices for a possible range, and then 14 functions with that particular range, we conclude that there are $6 \cdot 14 = 84$ possible $f$'s in Case 2.
DPatrick 2014-03-15 21:06:03
And for each $f$, how many functions $g$ are there with disjoint range?
claudeaops 2014-03-15 21:06:37
2^4=16
MSTang 2014-03-15 21:06:37
16
larry333 2014-03-15 21:06:37
16
Petaminx 2014-03-15 21:06:37
$2^4$
Tommy2000 2014-03-15 21:06:37
2^4?
DPatrick 2014-03-15 21:07:00
Right, each element of $A$ has 2 choices of where it can map to under $g$ (because it has to miss the range of $f$), so there are $2^4 = 16$ such $g$ (for each $f$).
DPatrick 2014-03-15 21:07:28
Therefore, how many pairs of functions in Case 2?
Satyaprakash2009rta 2014-03-15 21:07:57
16*84=1344
mishai 2014-03-15 21:07:57
1344
Tommy2000 2014-03-15 21:07:57
16*84 = 1344
ChilledLemonade 2014-03-15 21:07:57
84*16=1344
sparkles257 2014-03-15 21:07:57
84*16
hwl0304 2014-03-15 21:07:57
84*16
cellobix 2014-03-15 21:07:57
So 84*16 pairs!
kgator 2014-03-15 21:07:57
and $84 * 16 = 1334$
vinayak-kumar 2014-03-15 21:07:57
84*16
guilt 2014-03-15 21:07:57
16*84 = 1344
DPatrick 2014-03-15 21:08:17
Yep: 84 possible $f$'s and then 16 possible $g$'s for each $f$, so $84 \cdot 16 = 1344$ pairs total in case 2.
DPatrick 2014-03-15 21:08:33
Case 3: the range of $f$ has 3 elements.
DPatrick 2014-03-15 21:08:39
How many such $f$ are there?
raptorw 2014-03-15 21:09:36
144
claudeaops 2014-03-15 21:09:36
4*3*12=144
nsun48 2014-03-15 21:09:36
4*4C2*6=144
joshxiong 2014-03-15 21:09:36
(4 choose 3) * (3^4-3*2^4+3*1^4) = 144
DPatrick 2014-03-15 21:09:55
Indeed there are 144, and there are a few different ways to count them.
DPatrick 2014-03-15 21:10:17
There's are ugly casework ways, but there's a simpler way too.
DPatrick 2014-03-15 21:10:31
How many TOTAL functions from $A$ to $A$ are there, with no restrictions?
dli00105 2014-03-15 21:10:55
256
negativebplusorminus 2014-03-15 21:10:55
4^4
jilynne1992 2014-03-15 21:10:55
256
shreshrej 2014-03-15 21:10:55
4^4
othuum0149 2014-03-15 21:10:55
4^4=256
apple.singer 2014-03-15 21:10:55
4^4=256
VietaFan 2014-03-15 21:10:55
4^4 = 256
aprobs21 2014-03-15 21:10:55
4^4
DPatrick 2014-03-15 21:11:10
Right: each of the four elements of $A$ can map to any of 4 elements, so there are $4^4 = 256$ total functions.
DPatrick 2014-03-15 21:11:17
How many functions are there that have a 4-element range?
joshxiong 2014-03-15 21:11:53
4! = 24
othuum0149 2014-03-15 21:11:53
24
ashgabat 2014-03-15 21:11:53
4! = 24
mathtastic 2014-03-15 21:11:53
4!=24
Quadratic64 2014-03-15 21:11:53
4!=24
phil9047 2014-03-15 21:11:53
24
DPatrick 2014-03-15 21:12:11
Right: a map from $A$ to $A$ with a 4-element range is just a permutation of $A$, and we know there are $4! = 24$ of those.
ashgabat 2014-03-15 21:12:36
So we know that the number of functions with a 3-element range is 256 - 84 - 24 - 4 = 144.
DPatrick 2014-03-15 21:12:48
So since we've already counted the 1-element and 2-element range functions and now we've counted the 4-element range functions, what's leftover are the 3-element range functions!
DPatrick 2014-03-15 21:12:58
So there are $256 - 4 - 84 - 24 = 144$ of them.
DPatrick 2014-03-15 21:13:15
And how many functions $g$ for each such $f$?
kquittman 2014-03-15 21:13:41
1
firemike 2014-03-15 21:13:41
1
summitwei 2014-03-15 21:13:41
1
ruiqiu2000 2014-03-15 21:13:41
one
danzhi 2014-03-15 21:13:41
1
allaoc 2014-03-15 21:13:41
1
DPatrick 2014-03-15 21:14:02
Just 1: our $g$ must map all four elements of $A$ to the fourth element that's not in the range of $f$.
DPatrick 2014-03-15 21:14:11
So there are 144 pairs of functions in case 3.
DPatrick 2014-03-15 21:14:38
Phew! That's all of the cases, so we've counted all of the pairs.
DPatrick 2014-03-15 21:14:46
Adding the cases, there are $324 + 1344 + 144 = 1812$ pairs of functions total.
allaoc 2014-03-15 21:14:58
So (324+1244+144)/(256^2)
shreshrej 2014-03-15 21:15:05
We add all of them, and divide by $256^2$
firemike 2014-03-15 21:15:12
so 144+324+1344/256^2 = 453/2^14 and our answer is 453
DPatrick 2014-03-15 21:15:35
Right. To finish, note that there are 256 functions, so there are $256^2 = 2^{16}$ pairs of functions possible.
DPatrick 2014-03-15 21:15:44
Thus the probability of picking a pair with disjoint ranges is $\dfrac{1812}{2^{16}}.$
DPatrick 2014-03-15 21:15:55
This simplifies down to $\dfrac{453}{2^{14}},$ so the answer is $\boxed{453}.$
DPatrick 2014-03-15 21:16:30
On to the final page! #13:
DPatrick 2014-03-15 21:16:37
13. On square $ABCD,$ points $E$, $F$, $G$, and $H$ lie on sides $\overline{AB}$, $\overline{BC}$, $\overline{CD}$, and $\overline{DA}$, respectively, so that $\overline{EG} \perp \overline{FH}$ and $EG = FH = 34.$ Segments $\overline{EG}$ and $\overline{FH}$ intersect at a point $P,$ and the areas of quadrilaterals $AEPH$, $BFPE$, $CGPF$, and $DHPG$ are in the ratio $269:275:405:411.$ Find the area of square $ABCD.$
DPatrick 2014-03-15 21:16:53
DPatrick 2014-03-15 21:17:07
They were kind enough to provide the picture on the contest.
DPatrick 2014-03-15 21:17:24
But this diagram is a bit silly -- let's put the numbers from the ratio into the diagram. But remember -- they're just ratios! They're not the actual areas! So let's call them $269k$, $275k$, etc. where $k$ is some constant that we want to determine later.
DPatrick 2014-03-15 21:17:34
room456 2014-03-15 21:17:50
the picture is not to scale
DPatrick 2014-03-15 21:18:06
Indeed, it is so not to scale. I hope no one tried to measure from this picture.
othuum0149 2014-03-15 21:18:15
total area of 1360k
j2002 2014-03-15 21:18:26
isnt the area 1360k
DPatrick 2014-03-15 21:18:39
Indeed, the total area is $1360k$. So if we can find $k$, then we're done.
DPatrick 2014-03-15 21:18:57
Let's also let $s$ be the side length of the square just so we have it around as a variable. So $s^2 = 1360k.$ (This will probably come in handy later.)
DPatrick 2014-03-15 21:19:12
Those numbers are very strange. Any patterns among them?
vinayak-kumar 2014-03-15 21:19:28
$269k+411k=275k+405k$
guilt 2014-03-15 21:19:28
note that 269+ 411 = 275+405
shreshrej 2014-03-15 21:19:28
the first thing i saw was 269+411=275+405
csmath 2014-03-15 21:19:28
275+405=269+411 so EG cuts this square in half
VietaFan 2014-03-15 21:19:28
k(269+411) = k(275+405)
guilt 2014-03-15 21:19:28
269+411 = 275+405
sojourner1 2014-03-15 21:19:38
680k, 680k
DPatrick 2014-03-15 21:19:47
Aha -- the area of $ADGE$ is $269k+411k = 680k$ and the area of $BCGE$ is $275k+405k = 680k$ too! So $\overline{EG}$ splits the square in half!
raptorw 2014-03-15 21:20:02
EG goes through center of square?
sojourner1 2014-03-15 21:20:02
EG passes through the center of ABCD
DPatrick 2014-03-15 21:20:13
Yes: for one thing, this means that $\overline{EG}$ passes through the center of the square. Let's call it $O$ and add it to the diagram:
DPatrick 2014-03-15 21:20:23
DPatrick 2014-03-15 21:20:37
That was fruitful, so let's look at the other segment now. How does $\overline{FH}$ spilt the square?
mssmath 2014-03-15 21:21:08
3(269+275)=2(405+411)
guilt 2014-03-15 21:21:08
also note (411+405)/(269+275) = 3/2
martian179 2014-03-15 21:21:08
ratio of 2 : 3
sparkles257 2014-03-15 21:21:08
2/3!
LightningX48 2014-03-15 21:21:08
2:3?
raptorw 2014-03-15 21:21:08
2:3
negativebplusorminus 2014-03-15 21:21:08
2 : 3
minimario 2014-03-15 21:21:08
$2:3$
DPatrick 2014-03-15 21:21:20
The area of $ABFH$ is $269k + 275k = 544k$ and the area of $CDHF$ is $411k+405k = 816k.$
DPatrick 2014-03-15 21:21:28
And $\dfrac{544k}{816k} = \dfrac{544}{816} = \dfrac23.$
DPatrick 2014-03-15 21:21:35
So $\overline{FH}$ splits the square in a $2:3$ ratio.
DPatrick 2014-03-15 21:21:40
How do we use that information?
martian179 2014-03-15 21:22:22
draw line through O parallel to BC
joshxiong 2014-03-15 21:22:22
consider the midpoint of FH
guilt 2014-03-15 21:22:22
draw line through oh parallel to bc. fh splits it into 2:3 ratio
DPatrick 2014-03-15 21:22:51
Right. We now know that the midpoint of $\overline{FH}$ is a point that is $\frac25$ths of the way from $\overline{AB}$ to $\overline{CD}$.
(One way to think of this is that the median of the trapezoid $AHFB$ has length $\frac25$ the side of the square.)
DPatrick 2014-03-15 21:22:59
So let's add that point and call it $Q.$ We get $Q$ by drawing the segment from $O$ that's parallel to $\overline{AD}$ and $\overline{BC}$.
DPatrick 2014-03-15 21:23:08
DPatrick 2014-03-15 21:23:27
And what do we know about $OQ$?
guilt 2014-03-15 21:23:48
note oq = s/10
ashgabat 2014-03-15 21:23:48
QO=1/10*s
martian179 2014-03-15 21:23:48
OQ = 1/10 s
guilt 2014-03-15 21:23:48
oq = s/10
apple.singer 2014-03-15 21:23:48
s/10?
Tommy2000 2014-03-15 21:23:48
1/10s?
othuum0149 2014-03-15 21:23:48
s/10
DrMath 2014-03-15 21:23:48
1/10 side length
DPatrick 2014-03-15 21:23:59
Right. $Q$ is $\frac25$ of the way from the left to the right, and $O$ is exactly halfway. So $OQ = \dfrac{s}{2} - \dfrac{2s}{5} = \dfrac{s}{10}.$
DPatrick 2014-03-15 21:24:12
Hmmm...we've got a tiny right triangle $OPQ$ in the middle, and we now know that its hypotenuse is $\dfrac{s}{10}$.
DPatrick 2014-03-15 21:24:18
What else can we learn about it?
DPatrick 2014-03-15 21:24:53
What other segment might be useful in this picture? (At least one person suggested it earlier but I lost it.)
mathtastic 2014-03-15 21:25:17
line through o parallel to HF
summitwei 2014-03-15 21:25:17
the line through O parallel to HF
DPatrick 2014-03-15 21:25:29
Yes -- I'd want to draw $\overline{F'H'}$, which is $\overline{FH}$ slid over to pass through $O$:
DPatrick 2014-03-15 21:25:35
DPatrick 2014-03-15 21:25:50
Why is that segment really really useful?
ptes77 2014-03-15 21:26:16
Now AH'F'B is equal to half the square
summitwei 2014-03-15 21:26:16
cuts it into four equal parts
sinani206 2014-03-15 21:26:16
it splits square in half
shreshrej 2014-03-15 21:26:16
It splits ABCD In half as well?
Deathranger999 2014-03-15 21:26:16
The lines are perpendicular, it divides the square equally.
ashgabat 2014-03-15 21:26:16
Splits ABCD in half
tennis1729 2014-03-15 21:26:16
goes through the center o
CornSaltButter 2014-03-15 21:26:16
it means that the parallelogram is 1/10 the area of the square
DPatrick 2014-03-15 21:26:46
Yeah, that new line really helps us keep track of the areas, because now $\overline{EG}$ and $\overline{F'H'}$ split the square into quarters!
DPatrick 2014-03-15 21:27:04
The picture is deceptive now because, for example, $411k$ is the area of $HDGP$, not of $H'DGO$, and similarly $405k$ is deceptive. So let's adjust the area labels for our new regions.
sinani206 2014-03-15 21:27:17
340k
gamjawon 2014-03-15 21:28:06
340k!
firemike 2014-03-15 21:28:06
340k
kgator 2014-03-15 21:28:06
If you eliminate the original HF then the resulting sectors have equal area of 340
vinayak-kumar 2014-03-15 21:28:06
we have 1/4 of the area now
kquittman 2014-03-15 21:28:06
340k
kgator 2014-03-15 21:28:06
340k
ptes77 2014-03-15 21:28:11
So HH'OP=71k and OPFF'=65k
DPatrick 2014-03-15 21:28:13
Right, our big upper-right hunk $H'DGO$ and our lower-right hunk $OGCF'$ are each exactly one-quarter of the entire square, so they're each $\frac14(1360k) = 340k.$
DPatrick 2014-03-15 21:28:36
That makes $HH'OP$ the remaining $71k$ from the upper-right, and $FF'OP$ the remaining $65k$ from the lower-right:
DPatrick 2014-03-15 21:28:49
DPatrick 2014-03-15 21:29:02
Now what does this tell us about the area of the little triangle $OPQ$?
ashgabat 2014-03-15 21:29:22
Still kinda a little deceptive
DPatrick 2014-03-15 21:29:41
Right...just to clarify, it's $OP$ that divides the skinny parallelogram into $71k$ and $65k$.
tiger21 2014-03-15 21:30:04
it's 3k!
pickten 2014-03-15 21:30:04
area 3k
negativebplusorminus 2014-03-15 21:30:04
3k
Tommy2000 2014-03-15 21:30:04
it is area 3k.
room456 2014-03-15 21:30:04
3k is area
summitwei 2014-03-15 21:30:04
3k
joshxiong 2014-03-15 21:30:04
3k
ViolinNinja256 2014-03-15 21:30:04
its 3k
ptes77 2014-03-15 21:30:04
it's 3k
DPatrick 2014-03-15 21:30:45
Right! Because we know that $OQ$ divides that parallelogram in half exactly, with $68k$ on each side.
DPatrick 2014-03-15 21:31:07
So $OPQ$ has area $3k$: it's the "extra" $3k$ of area on the top side of $OP$ and the "missing" $3k$ of area from the bottom side of $OP$.
DPatrick 2014-03-15 21:31:18
Getting closer...we've got right triangle $OPQ$ with area $3k$ and hypotenuse $\dfrac{s}{10}.$ But we still need more data.
tennis1729 2014-03-15 21:31:35
FH is 34
guilt 2014-03-15 21:31:43
use the data about 34
mishai 2014-03-15 21:31:43
we haven't used 34 yet
DPatrick 2014-03-15 21:31:54
Indeed, now is a good time to ask: What haven't we used yet?
DPatrick 2014-03-15 21:31:59
We haven't used at all the fact that $EG = FH = 34.$ How does that help?
DPatrick 2014-03-15 21:32:10
Here's the pic again:
DPatrick 2014-03-15 21:32:23
cxiong 2014-03-15 21:32:38
area of FF'H'H
DPatrick 2014-03-15 21:32:58
Good! In particular, the area of parallelogram $FF'H'H$ is $136k,$ and $FH = 34$. What can you conclude?
CornSaltButter 2014-03-15 21:33:31
34(OP)=136k, so OP=4k!
LightningX48 2014-03-15 21:33:31
height is 4k of ff'h'h
sojourner1 2014-03-15 21:33:31
OP=4k
room456 2014-03-15 21:33:31
op is 4k
MSTang 2014-03-15 21:33:31
PO = 4k
Tommy2000 2014-03-15 21:33:31
op = 4k
Quadratic64 2014-03-15 21:33:31
OP = 136k/34
SuperSnivy 2014-03-15 21:33:31
OP=4k
sirknightingfail 2014-03-15 21:33:31
OP=4k
DPatrick 2014-03-15 21:33:39
$OP$ is the height of that parallelogram! So we have $OP = \frac{136k}{34} = 4k.$
DPatrick 2014-03-15 21:33:54
Aha...now we have right triangle $OPQ$ with area $3k,$ hypotenuse $\dfrac{s}{10},$ and one leg $4k.$
SockFoot 2014-03-15 21:34:27
other leg is 3/2
mentalgenius 2014-03-15 21:34:27
the other leg is 1.5
VietaFan 2014-03-15 21:34:27
other leg is 3/2
othuum0149 2014-03-15 21:34:27
other leg is then 3/2
DrMath 2014-03-15 21:34:27
so one leg is 3/2
MSTang 2014-03-15 21:34:27
The other leg is 3/2
pickten 2014-03-15 21:34:27
other leg has length 3/2
DPatrick 2014-03-15 21:34:39
If the area is $3k$ and one leg is $4k$, then the other leg must be $\frac32.$
DPatrick 2014-03-15 21:34:47
So now our right triangle has legs $\frac32$ and $4k$ and hypotenuse $\dfrac{s}{10}.$
firemike 2014-03-15 21:34:55
pythagorean theorem
Tommy2000 2014-03-15 21:34:55
pythagorean theorem
hwl0304 2014-03-15 21:34:55
pyth theorem
DPatrick 2014-03-15 21:35:02
Pythagorean Theorem time!
DPatrick 2014-03-15 21:35:09
We have the equation $$\left(\frac32\right)^2 + (4k)^2 = \left(\frac{s}{10}\right)^2.$$
DPatrick 2014-03-15 21:35:18
This simplifies to $\dfrac94 + 16k^2 = \dfrac{s^2}{100}.$
raptorw 2014-03-15 21:35:38
use s^2=1360k
sparkles257 2014-03-15 21:35:38
now s^2 = 1360 k comes in handy
dli00105 2014-03-15 21:35:38
s^2=1360k
guilt 2014-03-15 21:35:38
aslo s^2 = 1360k
Deathranger999 2014-03-15 21:35:38
s squared equals 1360k.
DPatrick 2014-03-15 21:35:45
I knew that would come in handy later!
DPatrick 2014-03-15 21:35:51
Remember that $s^2 = 1360k.$
DPatrick 2014-03-15 21:35:58
So we have $\dfrac94 + 16k^2 = \dfrac{136k}{10}.$ That's a quadratic in $k$ so we can solve it!
DPatrick 2014-03-15 21:36:14
Multiply through by 20 to clear denominators, and move all the terms to one side:
$$320k^2 - 272k + 45 = 0.$$
DPatrick 2014-03-15 21:36:44
You can use the quadratic formula to solve this, but 45 doesn't have a ton of factors (and we know we want a rational solution) so you might find factoring easier.
SuperSnivy 2014-03-15 21:36:55
(8k-5)(40k-9)=0
DPatrick 2014-03-15 21:37:04
It factors as $(40k-9)(8k-5) = 0.$
DPatrick 2014-03-15 21:37:11
So the solutions are $k=\dfrac{9}{40}$ and $k=\dfrac{5}{8}.$
ws5188 2014-03-15 21:37:32
how do we know which root to use?
raptorw 2014-03-15 21:37:32
oh there's another positive sol...
mathtastic 2014-03-15 21:37:32
oh no
DPatrick 2014-03-15 21:37:38
Since the area of the square is $1360k,$ this gives possible areas of $1360\left(\frac{9}{40}\right) = 34 \cdot 9 = 306$ and $1360\left(\frac58\right) = 170 \cdot 5 = 850.$
DPatrick 2014-03-15 21:37:48
I didn't come all this way just to guess which of these is correct. Which one do we want?
ViolinNinja256 2014-03-15 21:38:10
oh it has be bigger than 578
ashgabat 2014-03-15 21:38:10
We know the area is greater than (34)^2/2 = 578
negativebplusorminus 2014-03-15 21:38:10
Must be greater than half of 34^2.
DPatrick 2014-03-15 21:38:19
Right. There's one part of the area of the entire square that we definitely know, and that's the area of $EFGH$ -- it's a quadrilateral with perpendicular diagonals of length 34, so its area is $\frac12(34)^2 = 578.$
DPatrick 2014-03-15 21:38:33
Thus the area of $ABCD$ is at least 578.
DPatrick 2014-03-15 21:38:41
And happily, only one of our two solutions is bigger than 578 -- that's $\boxed{850}$.
mssmath 2014-03-15 21:39:06
Hardest problem?
DPatrick 2014-03-15 21:39:28
For us here at AoPS, it definitely was. We got a lot of messy solutions but it took a while for us to come up with this fairly nice solution.
DPatrick 2014-03-15 21:39:34
Anyway, only 2 more to go!
DPatrick 2014-03-15 21:39:40
14. Let $m$ be the largest real solution to the equation
$$ \frac{3}{x-3} + \frac{5}{x-5} + \frac{17}{x-17} + \frac{19}{x-19} = x^2 - 11x - 4. $$There are positive integers $a$, $b$, and $c$ such that $m = a + \sqrt{b+\sqrt{c}}.$ Find $a+b+c.$
qwertyu 2014-03-15 21:40:12
Add four to both sides
ws5188 2014-03-15 21:40:12
add 4 to both sides
liuh008 2014-03-15 21:40:12
Add 4 to both sides.
SockFoot 2014-03-15 21:40:12
+4 to all sides
mssmath 2014-03-15 21:40:12
Add 4 to both sides
theinfinite 2014-03-15 21:40:12
add 4 to both sides
dli00105 2014-03-15 21:40:15
4 terms on LHS and -4 on RHS
DPatrick 2014-03-15 21:40:32
Yeah, that -4 looks weird on the right side. And there happens to be 4 terms on the left side.
DPatrick 2014-03-15 21:40:38
What if we add 4 to both sides, by adding 1 to each of the fractions on the left?
DPatrick 2014-03-15 21:40:45
$$
\left(\frac{3}{x-3} + 1\right)
+\left(\frac{5}{x-5} + 1\right)
+\left(\frac{17}{x-17} + 1\right)
+\left(\frac{19}{x-19} + 1\right)
= x^2 - 11x.
$$
gamjawon 2014-03-15 21:41:06
then simplify!
vinayak-kumar 2014-03-15 21:41:06
the numerators turn into an $x$ !
aprobs21 2014-03-15 21:41:06
Well then all the numerators have x's!
brian22 2014-03-15 21:41:06
x/(x-3)+x/(x-5)+x/(x-17)+x/(x-19)=x(x-11)
cxiong 2014-03-15 21:41:06
that simplifies a lot
DPatrick 2014-03-15 21:41:19
Ooh, those parentheses simplify nicely!
DPatrick 2014-03-15 21:41:25
$$
\frac{x}{x-3} + \frac{x}{x-5} + \frac{x}{x-17} + \frac{x}{x-19} = x^2 - 11x.
$$
theinfinite 2014-03-15 21:41:44
divide out x
summitwei 2014-03-15 21:41:44
and then divide out an x
mishai 2014-03-15 21:41:44
divide by x
hjl00 2014-03-15 21:41:44
divide by x
VietaFan 2014-03-15 21:41:44
divide by x
othuum0149 2014-03-15 21:41:44
divide both sides by x
DPatrick 2014-03-15 21:41:49
The number we're looking for is clearly not $x=0$, so let's divide by $x$.
DPatrick 2014-03-15 21:41:58
$$
\frac{1}{x-3} + \frac{1}{x-5} + \frac{1}{x-17} + \frac{1}{x-19} = x - 11.
$$
DPatrick 2014-03-15 21:42:08
Now what?
vinayak-kumar 2014-03-15 21:42:20
hmm, lots of symmetry
aprobs21 2014-03-15 21:42:20
symmetry about 11
othuum0149 2014-03-15 21:42:36
substitute for x-11
MSTang 2014-03-15 21:42:36
$y = x-11$
aprobs21 2014-03-15 21:42:36
symmetry about 11
sparkles257 2014-03-15 21:42:36
let a = x-11
cellobix 2014-03-15 21:42:36
sub y=x-11?
DPatrick 2014-03-15 21:42:42
There's some symmetry here that we'd like to take advantage of. The $x-11$ on the right side is also a clue.
DPatrick 2014-03-15 21:42:48
Let's make the substitution $y = x-11$ to capture the symmetry.
DPatrick 2014-03-15 21:42:54
Now we have
$$
\frac{1}{y+8} + \frac{1}{y+6} + \frac{1}{y-6} + \frac{1}{y-8} = y.
$$
raptorw 2014-03-15 21:43:21
pair up the first and last terms in LHS
mathtastic 2014-03-15 21:43:21
DIFF OF SOME SQUARES YO
dli00105 2014-03-15 21:43:21
pair them off!
raptorw 2014-03-15 21:43:21
Diff of squares
SockFoot 2014-03-15 21:43:21
group pairs of fractions together
checkmate1021 2014-03-15 21:43:21
difference of squares?
CornSaltButter 2014-03-15 21:43:21
Difference of squares yes
buzzyun 2014-03-15 21:43:21
difference of squares
DPatrick 2014-03-15 21:43:28
It makes sense to try to combine the fractions on the left with $y \pm 6$ and $y \pm 8$ in the denominators.
DPatrick 2014-03-15 21:43:37
$$\frac{2y}{y^2 - 36} + \frac{2y}{y^2 - 64} = y.$$
joshualee2000 2014-03-15 21:44:03
dvide by y
Rectangle_Square 2014-03-15 21:44:03
divide by y
hwl0304 2014-03-15 21:44:03
divide by y
sojourner1 2014-03-15 21:44:03
take out the y
liuh008 2014-03-15 21:44:03
Divide out y - the answer is not 11.
LightningX48 2014-03-15 21:44:03
divide by y
DPatrick 2014-03-15 21:44:14
Now $y=0$ is also not the solution we want (it gives $x=11$, which is not the right format), so we can divide by $y$:
$$\frac{2}{y^2 - 36} + \frac{2}{y^2 - 64} = 1.$$
mentalgenius 2014-03-15 21:44:47
nice easy quadratic
aprobs21 2014-03-15 21:44:47
Quadratic in y^2
DPatrick 2014-03-15 21:45:06
We could certainly make another substitution now, but at this point it's "simple" enough that I'd probably just plow through to the end from here.
DPatrick 2014-03-15 21:45:17
We can expand this out by multiplying by $(y^2-36)(y^2-64)$:
$$
2(y^2-64) + 2(y^2-36) = (y^2-36)(y^2-64).
$$
This simplifies to $y^4 - 104y^2 + 2504 = 0.$
googol.plex 2014-03-15 21:45:43
quadratic formula
theinfinite 2014-03-15 21:45:43
quadratic
gamjawon 2014-03-15 21:45:43
Quadratic formula?
DPatrick 2014-03-15 21:45:46
This is a quadratic in $y^2$ so we can use the quadratic formula:
$$
y^2 = \frac{104 \pm \sqrt{104^2 - 4 \cdot 2504}}{2} = 52 \pm \sqrt{52^2 - 2504} = 52 \pm \sqrt{200}.
$$
DPatrick 2014-03-15 21:46:06
And now we just unwind our substitution.
DPatrick 2014-03-15 21:46:13
Recall that $x = y + 11$, and we want the largest root.
danusv 2014-03-15 21:46:22
so x= 11 + root( 52 + root(200)
googol.plex 2014-03-15 21:46:26
a=11, b=52, c=200 so 263
csmath 2014-03-15 21:46:39
We need the LARGEST sol so we get 11+sqrt(52+sqrt(200))
ashgabat 2014-03-15 21:46:39
DPatrick 2014-03-15 21:46:41
So $x = 11 + \sqrt{52 + \sqrt{200}}$ is the largest root. Our answer is $11 + 52 + 200 = \boxed{263}.$
DPatrick 2014-03-15 21:46:58
On to the grand finale...
DPatrick 2014-03-15 21:47:04
15. In $\triangle ABC,$ $AB = 3,$ $BC = 4,$ $CA = 5.$ Circle $\omega$ intersects $\overline{AB}$ and $E$ and $B,$ $\overline{BC}$ at $B$ and $D,$ and $\overline{AC}$ and $F$ and $G.$ Given that $EF = DF$ and $\frac{DG}{EG} = \frac34,$ length $DE = \frac{a\sqrt{b}}{c},$ where $a$ and $c$ are relatively prime positive integers, and $b$ is a positive integer not divisible by the square of any prime. Find $a+b+c.$
ptes77 2014-03-15 21:47:23
Draw
room456 2014-03-15 21:47:23
DIAGRAM!
pedronr 2014-03-15 21:47:23
Draw a picture!!
ChilledLemonade 2014-03-15 21:47:23
diagram!
DPatrick 2014-03-15 21:47:29
Let's try to sketch a picture. It'll likely be out-of-whack at first, but we can refine it as we learn more.
DPatrick 2014-03-15 21:47:36
DPatrick 2014-03-15 21:47:55
We want the length $DE,$ which I've drawn in blue. What do we know about the blue segment?
joshxiong 2014-03-15 21:48:21
DE is a diameter
ninjataco 2014-03-15 21:48:21
diameter of circle
MSTang 2014-03-15 21:48:21
A diameter of the circle
Rectangle_Square 2014-03-15 21:48:21
diameter of circle
Deathranger999 2014-03-15 21:48:21
It's the diamter.
ws5188 2014-03-15 21:48:21
it is the diameter of the circle
shreshrej 2014-03-15 21:48:21
diameter because of right angle B
Deathranger999 2014-03-15 21:48:21
It's the diameter.
DPatrick 2014-03-15 21:48:28
Since $\angle DBE$ is right, the blue segment $\overline{DE}$ is a diameter of the circle.
DPatrick 2014-03-15 21:48:46
OK, let's next look at $F.$ We have $DF = EF$ -- what does that tell us?
MSTang 2014-03-15 21:49:16
DEF is 45-45-90
summitwei 2014-03-15 21:49:16
DEF is isoceles
vinayak-kumar 2014-03-15 21:49:16
its a 45 45 90 triangle
CornSaltButter 2014-03-15 21:49:16
DFE is 45-45-90 triangle
cxiong 2014-03-15 21:49:16
DEF is isosceles right triangle
othuum0149 2014-03-15 21:49:16
45-45-90 triangle
mishai 2014-03-15 21:49:16
iscosceles 90 45 45 triangle
larry333 2014-03-15 21:49:16
DEF is 45-45-90
mdlu 2014-03-15 21:49:16
isosceles right triangle
DPatrick 2014-03-15 21:49:27
Well, we can tell that I drew a pretty lousy picture, for one thing. $F$ needs to be equidistant from $D$ and $E$.
DPatrick 2014-03-15 21:49:49
Indeed, we know that $\angle DFE$ is a right angle too, since it intercepts a semicircle. (And $\angle DGE$ is too -- let's file that away for later.) So $DFE$ is a 45-45-90 isosceles right triangle. And in particular $F$ is the midpoint of the arc from $D$ to $E.$
LightningX48 2014-03-15 21:50:06
NDP = New Diagram Plz
DPatrick 2014-03-15 21:50:17
Yeah, let's try to make the picture more accurate in light of what we've got so far:
DPatrick 2014-03-15 21:50:24
DPatrick 2014-03-15 21:50:41
Notice anything else about $F$?
DPatrick 2014-03-15 21:51:06
Remember, we just learned that arcs $DF$ and $FE$ are each one-quarter of the circle. What does that tell us?
SuperSnivy 2014-03-15 21:51:34
BF is angle bisector of <ABC
ABCDE 2014-03-15 21:51:34
BF is angle bisector
DPatrick 2014-03-15 21:51:59
Exactly. $BF$ bisects the arc $DFE$, so it bisects the inscribed angle $\angle DBE$ too.
DPatrick 2014-03-15 21:52:17
Or, you can note that $\angle DBF$ and $\angle EBF$ are each $45^\circ$, because they each intercept a $90^\circ$ arc.
DPatrick 2014-03-15 21:52:39
OK, on to $G$.
DPatrick 2014-03-15 21:52:48
We have $\dfrac{DG}{EG} = \dfrac34$. Hmmm...those numbers are familiar...and $\angle DGE$ is a right angle from earlier.
mishai 2014-03-15 21:53:06
DG:EG=3/4, so DEG is similar to ACB
vinayak-kumar 2014-03-15 21:53:06
Draw triangle DGE
LightningX48 2014-03-15 21:53:06
yeay 3,4,5
othuum0149 2014-03-15 21:53:06
3-4-5 triangle
CornSaltButter 2014-03-15 21:53:06
HHMMMMM... 3-4-5??
ashgabat 2014-03-15 21:53:06
∆DGE ~ ∆ABC
mentalgenius 2014-03-15 21:53:06
3-4-5 right triangle
DPatrick 2014-03-15 21:53:15
Aha: triangle $DGE$ is similar to a 3-4-5 triangle. Let me draw it in red and mark the equal angles:
DPatrick 2014-03-15 21:53:23
DPatrick 2014-03-15 21:54:01
Any other angles equal to those angles too? (Maybe I need to add another segment?)
joshxiong 2014-03-15 21:54:28
BG
MSTang 2014-03-15 21:54:35
EBG
lovejj 2014-03-15 21:54:35
<DBG
DPatrick 2014-03-15 21:54:52
Let's add $\overline{BG}$:
DPatrick 2014-03-15 21:55:00
mdlu 2014-03-15 21:55:08
angles that intercept the same arcs
ashgabat 2014-03-15 21:55:08
Same arc.
DPatrick 2014-03-15 21:55:25
Right:
$\angle DEG$ and $\angle DBG$ intercept the same arc $DG$, so they're the same angle.
$\angle EDG$ and $\angle EBG$ intercept the same arc $EG$, so they're the same angle.
DPatrick 2014-03-15 21:55:35
So what?
mathwizard888 2014-03-15 21:55:45
CG=BG=AG
ChilledLemonade 2014-03-15 21:55:45
Isoceles triangles!
joshualee2000 2014-03-15 21:55:45
we get two isoceles triangles
cxiong 2014-03-15 21:55:54
G is the midpoint of AC
dli00105 2014-03-15 21:55:54
G miodpoint of AC
checkmate1021 2014-03-15 21:55:54
BG = AG
MSTang 2014-03-15 21:55:54
GA = GB = GC
DPatrick 2014-03-15 21:56:00
This means that $CGB$ is isosceles (it has two green base angles), so $CG = BG.$
DPatrick 2014-03-15 21:56:10
Also that $AGB$ is isosceles (with two pink base angles), so $AG = BG.$
DPatrick 2014-03-15 21:56:19
This means that $AG = BG = CG.$ In other words, $G$ is the midpoint of $\overline{AC}$!
DPatrick 2014-03-15 21:56:35
So to summarize so far, we've concluded that:
(1) -- $\overline{DE}$ is the diameter of the circle
(2) -- $\overline{BF}$ is the angle bisector of right angle $\angle ABC$
(3) -- $G$ is the midpoint of $\overline{AC}$.
DPatrick 2014-03-15 21:56:50
Here's the "clean" picture again:
DPatrick 2014-03-15 21:56:56
DPatrick 2014-03-15 21:57:22
We have a lot of information now, and there are many ways to finish from here.
joshxiong 2014-03-15 21:57:43
Power of a point maybe
mathtastic 2014-03-15 21:57:43
ANGLE BISECTOR THEOREM
DPatrick 2014-03-15 21:58:33
I agree: I think that's probably the easiest way to finish. We know a lot of info about how $F$ and $G$ divide up the distances on $\overline{AC}$, so we should be able to use Power of a Point on points $A$ and $C$.
DPatrick 2014-03-15 21:58:51
Let's look at Power of a Point on $A$: $(AF)(AG) = (AE)(AB).$
DPatrick 2014-03-15 21:59:06
Which of these quantities do we know?
mathawesomeness777 2014-03-15 21:59:19
AB = 3
Petaminx 2014-03-15 21:59:19
AB=3
Tommy2000 2014-03-15 21:59:19
ag = 5/2
brian22 2014-03-15 21:59:26
AF=15/7, AG=5/2
MSTang 2014-03-15 21:59:26
all except AE
DPatrick 2014-03-15 21:59:37
Yes, we know three of them!
DPatrick 2014-03-15 21:59:45
$AB = 3$ (we were given that).
DPatrick 2014-03-15 21:59:55
$AG = \dfrac52$ because $G$ is the midpoint.
DPatrick 2014-03-15 22:00:21
And $AF = \dfrac{15}{7}$ by the Angle Bisector Theorem: we know that $AF:FC = 3:4$ and $AF + FC = 5$.
brian22 2014-03-15 22:00:38
AE=25/14
DPatrick 2014-03-15 22:01:04
Right: plugging these in we get $\dfrac{15}{7} \cdot \dfrac{5}{2} = AE \cdot 3.$
DPatrick 2014-03-15 22:01:18
So $AE = \dfrac{25}{14}$ when we solve this.
brian22 2014-03-15 22:01:31
->BE=17/14
cxiong 2014-03-15 22:01:31
BE = 17/14
summitwei 2014-03-15 22:01:31
So BE=17/14
DPatrick 2014-03-15 22:01:44
And more importantly, $BE = 3 - \dfrac{25}{14} = \dfrac{17}{14}.$
brian22 2014-03-15 22:01:54
Then do it again to C to get BD
cxiong 2014-03-15 22:02:07
Do the same for BD, then use Pythagorean Thereom
othuum0149 2014-03-15 22:02:07
proceed similarly to find BD
googol.plex 2014-03-15 22:02:07
power of a point on C, 4*CD=(5/2)(20/7)
DPatrick 2014-03-15 22:02:22
Exactly: we do the same computation of Power of a Point at point $C$.
DPatrick 2014-03-15 22:02:50
I'll skip the details since the computation is exactly the same: we know $CF = \frac{20}{7}$, $CG = \frac52$, and $CB = 4$, so $CD = \frac{25}{14}$ and hence $BD = \frac{31}{14}.$
DPatrick 2014-03-15 22:03:27
(Note that it came out that $AE = CD$ -- there are other ways to see that too. As I said, once you identify $F$ and $G$ there are tons of ways to finish.)
Galindon 2014-03-15 22:03:39
Pythagorean theorem done
joshxiong 2014-03-15 22:03:39
pythagorean theorem to finish
LightningX48 2014-03-15 22:03:39
pythag gives us
DPatrick 2014-03-15 22:04:09
Right, using the Pythagorean Theorem on triangle $BDE$, we get $$DE = \frac{1}{14}\sqrt{17^2 + 31^2} = \frac{\sqrt{1250}}{14} = \frac{25\sqrt2}{14}.$$
DPatrick 2014-03-15 22:04:20
So our answer is $25 + 2 + 14 = \boxed{041}.$
DPatrick 2014-03-15 22:04:35
That's it for the AIME I! Thanks for hanging out with me on a Saturday night.
DPatrick 2014-03-15 22:04:43
A full transcript of this Math Jam will be available on the website in the very near future. Go to the "Math Jams" link under the "School" menu.
DPatrick 2014-03-15 22:04:52
And please join us again for the AIME II Math Jam on Friday, March 28, when we'll do all this again for the AIME II.
DPatrick 2014-03-15 22:08:45
Bye!
ksun48 2014-03-15 22:09:01
Bye!
sojourner1 2014-03-15 22:09:12
Bye
joshxiong 2014-03-15 22:09:15
Bye!

Copyright © 2024 AoPS Incorporated. This page is copyrighted material. You can view and print this page for your own use, but you cannot share the contents of this file with others.