bad tommy >:(
by tastymath75025, Dec 11, 2017, 1:00 AM
Problem Set #3
by Tommy2000, Dec 23, 2016, 2:38 AM
This is overdue, but I have been rather busy recently... Let's start with some jokes to alleviate the mood.
-1) [minimario] Prove that
.
Solution
-2) [Classic] Prove that for
,
.
Solution
Ok now some real stuff:
1) USACO Dec 16 #1-2
2) [EGMO] Let
be a triangle, and
be such that
is a
triangle with obtuse angle at
and
. Furthermore, let
be the midpoint of
and
such that
. Prove that
.
Solution
I'll add more as I remember them (I kind of forgot to record stuff)
-1) [minimario] Prove that

Solution
By Wilson's theorem,
. Now, multiply both sides by
to obtain the result.


-2) [Classic] Prove that for

![$\sqrt[n]{2} \not\in \mathbb Q$](http://latex.artofproblemsolving.com/e/e/6/ee6f13c79d9b7b87a55c6afcb090f7c5cb950369.png)
Solution
Suppose for contradiction that it is
, then we get
, a contradiction by Fermat's Last Theorem.


Ok now some real stuff:
1) USACO Dec 16 #1-2
2) [EGMO] Let











Solution
Let
be the other trisection point on
. Then clearly
is cyclic, so it follows that
. Since
is the midpoint and on the angle bisector, the triangle is isosceles, and thus equilateral. But then
is clearly the midpoint of
, so
. Furthermore
, so
. Since
is a midline of
, the result follows.












I'll add more as I remember them (I kind of forgot to record stuff)

This post has been edited 1 time. Last edited by Tommy2000, Dec 23, 2016, 2:42 AM
Reason: Fun :D
Reason: Fun :D
Problem Set #2
by Tommy2000, Dec 7, 2016, 2:51 AM
1) USACO 2016 Jan Plat #1
Solution
2) CF 587C
Solution
3) Find all
such that
.
Solution
4) Let
be cyclic with
. Let
. Let
be the midpoints of
, respectively. The bisector of
intersects
at
and the bisector of
intersects
at
. Prove
.
Progress
Solution
As it turns out, an optimized
does in fact work.
However, the better
solution is as follows: Consider the starting and ending row for the rectangle. They key observation is given a valid rectangle, you would never move the right edge leftwards. Now, keep moving the right most boundary until you can't, then skip a block until you can do a
rectangle, and repeat (Line Sweep).

However, the better


2) CF 587C
Solution
This is rather nice, and the approach is essentially the same algorithm as LCA. For each vertex
, store the best
people up to its the
-th parent, so everything is
.




3) Find all


Solution
To show it's surjective, take
to get
.
To show it's injective, suppose
. Then we see that taking
and
, we get
To show
, it suffices to show
for some
. Since it's surjective, take
such that
, and
, so
.
Now, since it's injective, consider
, which implies
, where
. Stripping off the outer
, we get
, which is linear. It remains to check that all equations of this form do work, so the answer is
.


To show it's injective, suppose



![\[ 2b + f(f(y) - b) = f(f(b) + y) = f(f(a) + y) = 2a + f(f(y) - a) \]](http://latex.artofproblemsolving.com/2/f/b/2fb306913feaaa21ea00fcd992f1d3e83e38db9a.png)







Now, since it's injective, consider






4) Let












Progress
Lemma 1:
bisects 
Proof
Notice that the intersection of
and
is
in complex numbers. Since
and
lie on
, it follows that
concur.
Claim:
.


Proof
Let
. Observe that
so
is the midpoint as desired.

![\[ \frac{BP}{DP} = \frac{BC}{DC} \cdot \frac{\sin BCA}{\sin DCA} = \frac{BC}{DC} \cdot \frac{\sin BDA}{\sin DBA} = \frac{BC}{DC} \cdot \frac{BA}{DA} = 1\]](http://latex.artofproblemsolving.com/8/e/d/8ed205c78efaeaef664a750725d496c8f218b124.png)

Notice that the intersection of







Claim:

This post has been edited 1 time. Last edited by Tommy2000, Dec 7, 2016, 5:01 AM
Reason: Added #3
Reason: Added #3
Problem Set #1
by Tommy2000, Nov 29, 2016, 5:29 AM
Hmm, I guess I'll record some problems I recently did/had explained to me word for word '-' I'll post sols later if I have time...
1) Show that if
is prime for
, then
are primitive roots.
Solution
2) Prove that for
,
Solution
3) USAMO 2005/3
This problem isn't that fun...so I'm too lazy to post a solution.
4) WOOT 11/25 PoTD
Copied from Jeffery because too lazy to type
5) USACO Dec 15 Plat #1-3
Max Flow
High Card Low Card
Counting Haybales
6) 2016 PUMaC Finals #1-3
2016 PUMaC Finals 3
7) Find all
satisfying
and
.
Solution
8) (2016 Turkey TST #5) Find all functions
such that
and
.
Solution
It's also worth noting this nice result in fractals' solution. It was something I tried to prove, but ultimately could not.
1) Show that if



Solution
Suppose
is not a primitive root. Then
. By QR Law,
. However,
, so contradiction.
Suppose
is not a primitive root. Then
. By QR Law,
. However,
, so contradiction.




Suppose




2) Prove that for

![\[ \sum_{m = 1} ^ p \left( \sum_{n = 1} ^ h \left( \frac{m + n}{p} \right) \right) ^ 2 = h (p - h) \]](http://latex.artofproblemsolving.com/4/f/e/4fe29ff7873efd8744184b1018d3c742090f6c81.png)
Let
. We can write the sum as
In the first part, all are equal to
unless
, in which case it is
. Thus, the first part contributes
to the sum.
Claim:
for all
.
Proof
Since
, we can plug this in to obtain a sum of
as desired.

![\[ h \sum_{i = 1} ^ p \left( \frac{i}{p} \right) ^ 2 + 2 \sum_{k = 1} ^ h (h - k) S_k \]](http://latex.artofproblemsolving.com/c/4/4/c44f103ebd911d969aa45349cead8d1d1a482846.png)




Claim:


Proof
Note that
It suffices to find when
for some
. We can divide through by
(it's nonzero) to obtain an equation of the form
where
is nonzero. It's clear that the
in this are bijective with the
in the previous equation. We can write
as
and
for some
.
cannot be
or else
. Since
and
yield the same pair, we see there are
values of
which yield
. On the other hand, there are
values of
which yield a nonzero value. Hence, there are
values of
which yield
. Thus,
as desired.
![\[ S_k = \sum_{i = 1} ^ p \left( \frac{i ^ 2 + ik}{p} \right) = \sum_{i = 1} ^ p \left( \frac{(i + k/2) ^ 2 - (k / 2) ^ 2}{p} \right) \]](http://latex.artofproblemsolving.com/9/4/9/9492612a13b46fe382d0fe4ed1e11a83eb7f01d3.png)

























Since


3) USAMO 2005/3
This problem isn't that fun...so I'm too lazy to post a solution.
4) WOOT 11/25 PoTD
Copied from Jeffery because too lazy to type
Let
. Note that
and
so
is cyclic, so
lies on the perpendicular bisector of
. Since we want to show that the locus of
is a line perpendicular to
, which is parallel to
, we want
, which occurs when
. But, this is true as
. Thus, the result follows.












5) USACO Dec 15 Plat #1-3
Max Flow
For each flow, add 1 to each endpoint, subtract 1 from LCA and parent of LCA. Then to compute number of flows which go through a node, do a DFS.
High Card Low Card
Do it as if whole game were high card from beginning and whole were low card from end. Then this works by greedy.
Counting Haybales
Direct application of Segment Tree.
6) 2016 PUMaC Finals #1-3
2016 PUMaC Finals 3
We proceed by complex bashing. Let
be the unit circle. Then we have,

Now, since
is inside the quadrilateral, we have
iff
Hence, the two angles are equal.


Now, since



7) Find all



Solution
We claim
is odd. Suppose
, then
and
. Clearly these cannot be equal in magnitude unless
, in which case
is already odd.
is either
or
. However, if
,
and this contradicts
. Hence,
. Observe that
. Now, suppose
such that
. Then
. However,
is negative. Since
is odd and
, we conclude
, a contradiction. Thus, for
,
.
In particular,
. We claim
is increasing. Suppose
and
. Then we know that
. Since we know [that
is bounded between floor and ceil, this is a contradiction for sufficiently large
. Now notice that numbers of the form
for integers
are dense in reals, so we're done.






















![$f(c) \in [0, 1]$](http://latex.artofproblemsolving.com/d/6/9/d69c715e5e86ee1cd3f5364e1fd7ea53d1237f83.png)
In particular,
![$f(k) \in [\lfloor k \rfloor, \lceil k \rceil]$](http://latex.artofproblemsolving.com/0/3/2/032622cb8582feb4cfa8533d59ac1b49bf94b55e.png)








8) (2016 Turkey TST #5) Find all functions



Solution
Let (1) denote the first equation and (2) denote the second equation. Notice
by plugging
in (2). Furthermore, it's clear
for all
and
by (1).
[We claim that the answer is
for
. Since
is completely multiplicative, it suffices to find the values for
for prime
.
We first show that for each
,
is an odd power of
. First suppose
is not a power of
, then there is a prime
such that
. It's clear that there exists
such that
. Then note that
However, we know
, so this is a contradiction. Hence,
is a power of
. Now, observe that
for
, so it must be an odd power.
It remains to show that all of the primes have the same exponent. Take two arbitrary primes
with
and
. WLOG
. Observe that
. By (2) on
, we get
Applying the euclidean algorithm, we see
However,
is arbitrary, so this is a contradiction for sufficiently large
. Hence, the answer is indeed
for
.





[We claim that the answer is





We first show that for each









![\[ 1 + pm = qn \mid f(1) + f(pm) = 1 + f(pm) \]](http://latex.artofproblemsolving.com/5/8/9/589fe59975b6800a310e14f0e31623f7282fb017.png)





It remains to show that all of the primes have the same exponent. Take two arbitrary primes






![\[ p + q ^ x \mid p ^ a + \left( q ^ x \right) ^ b \]](http://latex.artofproblemsolving.com/f/3/f/f3feb186f8518dcdcb19ea05c264ac05aef11d31.png)
![\[ p + q ^ x \mid p ^ {a - b} - 1 \]](http://latex.artofproblemsolving.com/1/6/5/165b8622fabc0a1eb15f5f096ae407c306365631.png)




It's also worth noting this nice result in fractals' solution. It was something I tried to prove, but ultimately could not.
This post has been edited 15 times. Last edited by Tommy2000, Dec 7, 2016, 2:51 AM
Reason: Subject name
Reason: Subject name
New Blog
by Tommy2000, Nov 21, 2016, 4:46 AM
OK shameless advertising: https://miniluigi251.wordpress.com/
I can't guarantee activity nor quality of content. It's more like a place for me throw memories and things I'd like to remember. Read if you want, or don't; I'm fine with either.
I can't guarantee activity nor quality of content. It's more like a place for me throw memories and things I'd like to remember. Read if you want, or don't; I'm fine with either.
CSS Attempt
by Tommy2000, Aug 28, 2015, 4:04 PM
Oops, I kinda wanted to change the color of stuff, but I'm not too good at CSS yet 
Oh well, let me know if you like it or if it's too bright (which I expect it is, but idk)
BTW: In case you couldn't tell, I wasn't joking about pink being my favorite color
EDIT: I changed the title to "Blazer". However, this is NOT something related to 420, my school mascot is actually the "Blazer"

Oh well, let me know if you like it or if it's too bright (which I expect it is, but idk)
BTW: In case you couldn't tell, I wasn't joking about pink being my favorite color

EDIT: I changed the title to "Blazer". However, this is NOT something related to 420, my school mascot is actually the "Blazer"

This post has been edited 2 times. Last edited by Tommy2000, Aug 28, 2015, 4:07 PM
27. 2014 ISL C2
by Tommy2000, Aug 25, 2015, 5:28 PM
We have
sheets of paper, with the number
written on each of them. We perform the following operation. In every step we choose two distinct sheets; if the numbers on the two sheets are
and
, then we erase these numbers and write the number
on both sheets. Prove that after
steps, the sum of the numbers on all the sheets is at least
.
Hint
Solution







Hint
Look for a monovariant!
Solution
Notice that the product always increases by at least
times since
. The initial product is
, so after
steps, the product is at least
. Now, by AM-GM, we have
.





![$\sum \ge 2^m \sqrt[2^m]{\prod}=2^m*\sqrt[2^m]{2^{m*2^m}}=4^m$](http://latex.artofproblemsolving.com/d/2/8/d28401e173136828207612f89eb1bcf3a55defea.png)
This post has been edited 1 time. Last edited by Tommy2000, Aug 25, 2015, 5:28 PM
Reason: typo
Reason: typo
Archives






Shouts
Submit
47 shouts
Contributors
Tags
About Owner
- Posts: 715
- Joined: Jan 12, 2013
Blog Stats
- Blog created: Jul 21, 2015
- Total entries: 43
- Total visits: 5583
- Total comments: 58
Search Blog