Bay 12 Games Forum

Please login or register.

Login with username, password and session length
Advanced search  
Pages: 1 [2]

Author Topic: The Maths Thread  (Read 6130 times)

Virex

  • Bay Watcher
  • Subjects interest attracted. Annalyses pending...
    • View Profile
Re: The Maths Thread
« Reply #15 on: March 19, 2012, 02:43:18 pm »

Yeah, you're right. I'm going to spend some more time in the corner over here...
Logged

Another

  • Bay Watcher
    • View Profile
Re: The Maths Thread
« Reply #16 on: March 19, 2012, 04:11:04 pm »

What if f(x)=k*exp(-x^2/(2s))?
It's integral is bounded, so wouldn't any such tower converge to a finite number?
Was there a requirement to converge to non-zero values and can't it be satisfied with appropriate k and s?
Logged

MagmaMcFry

  • Bay Watcher
  • [EXISTS]
    • View Profile
Re: The Maths Thread
« Reply #17 on: March 19, 2012, 05:01:12 pm »

What if f(x)=k*exp(-x^2/(2s))?
It's integral is bounded, so wouldn't any such tower converge to a finite number?
Was there a requirement to converge to non-zero values and can't it be satisfied with appropriate k and s?
That will work, but can you prove it?

Yeah, you're right. I'm going to spend some more time in the corner over here...
No you won't.

In the meantime, enjoy some real math (not the stuff people do at school):

1. If two opposite corner squares of a chess board are removed, is it possible to dissect the rest of the chess board into thirty-one congruent dominos?

2. Ten bowls are standing in a circle. One of them contains two beans, the rest are empty. You may repeatedly choose two adjacent bowls and put one bean into each bowl. Can you reach a state where every bowl contains the same amount of beans?

3. Prove that in every triangle, the ___ intersect in one point:
  a) angle bisectors
  b) medians
  c) altitudes

4. Let k be a circle, and let AB be one of the circle's chords. Let k' be a circle that touches the interior of k in C and AB in D. Prove that CD bisects the angle ACB.

5. Given 1000 red points and 1000 blue points on the plane, of which no three are collinear, is it always possible to draw 1000 non-intersecting line segments onto the plane so that every point lies on a segment with a point of the other color?

6. There are five different colors of gumball in a gumball machine. How many gumballs must you buy to ensure that you have three of the same color?

7. Given a set of 101 distinct integers between 1 and 200, are there necessarily two different integers in the set so that one is a multiple of the other?

If you (that means you) have answers, please send me a PM instead of posting them here. I'll put your results into this post, and I'll post complete answers to the questions next Monday.


Teaser question (not a part of the test, available for public discussion):

You are given a pencil and a straightedge with two marks. How many line segments must you draw to construct three points that are the vertices of a regular triangle?
Note that you may not move the straightedge and the pencil at the same time, and that the straightedge may only be moved a finite number of times (in short: you may not use the two marks as a compass).
Logged

Virex

  • Bay Watcher
  • Subjects interest attracted. Annalyses pending...
    • View Profile
Re: The Maths Thread
« Reply #18 on: March 19, 2012, 05:09:31 pm »

What if f(x)=k*exp(-x^2/(2s))?
It's integral is bounded, so wouldn't any such tower converge to a finite number?
Was there a requirement to converge to non-zero values and can't it be satisfied with appropriate k and s?
That will work, but can you prove it?
If the integral is bounded then the upper and lower values could tend to -inf and inf and the total integral would still be bound. I can't prove off-hand that it would also converge though, there may be some values of a for which f(x)=k*exp(-x^2/(2s)) oscillates, although that seems unlikely.
Logged

MagmaMcFry

  • Bay Watcher
  • [EXISTS]
    • View Profile
Re: The Maths Thread
« Reply #19 on: March 19, 2012, 05:24:27 pm »

Actually, you can choose any integrable f(x) with |f(x)| < 1, and the resulting sequence will converge for any a. So if you choose k from (-1, 1), the resulting sequence will converge, but feel free to investigate the problem for other values of k.
Logged

MagmaMcFry

  • Bay Watcher
  • [EXISTS]
    • View Profile
Re: The Maths Thread
« Reply #20 on: March 20, 2012, 04:06:11 pm »

Teaser question hint: Try to construct some easier stuff first. Can you construct parallel and perpendicular lines and angle bisectors? Can you bisect (trisect, ...) line segments? Can you construct a translation (rotation) of a point, given direction and length (direction, angle and origin)? Does the term "Spiral of Theodorus" sound helpful?

Kickstarter: To draw a parallel line to a line g through a point P not on g, do the following:
Choose a point B on g. Mark the two points A and C on g that are one ruler mark distance (RMD) away from B (not g).
Choose a line h through B that intersects AP and CP in Q and R resp., with A != Q != R != C.
Now let S be the intersection of AR and CQ. PS is parallel to g.

Can someone prove why this works?
« Last Edit: March 21, 2012, 03:45:27 pm by MagmaMcFry »
Logged

MadocComadrin

  • Bay Watcher
  • A mysterious laboratory goblin!
    • View Profile
Re: The Maths Thread
« Reply #21 on: March 21, 2012, 02:49:30 pm »

I got a line PS that is perpendicular to g following those instructions. I might have read them wrong. Let me look again.

Edit: yep. Misread something that got me off on the wrong proof. :X

Nonetheless, somethings still not quite right, I'm getting either a perpendicular line or nothing of importance.
« Last Edit: March 21, 2012, 03:15:15 pm by MadocComadrin »
Logged

MagmaMcFry

  • Bay Watcher
  • [EXISTS]
    • View Profile
Re: The Maths Thread
« Reply #22 on: March 21, 2012, 03:48:34 pm »

I fixed a typo on line 2 of the instructions, maybe that's the problem.
ABC should be three points on g with |AB| = |BC|.
Logged

Spectr

  • Bay Watcher
  • Watching Bays
    • View Profile
Re: The Maths Thread
« Reply #23 on: March 21, 2012, 09:18:16 pm »

I found that this video helps me quite a bit with the maths.
Logged
Word

darkrider2

  • Bay Watcher
    • View Profile
Re: The Maths Thread
« Reply #24 on: March 21, 2012, 09:27:37 pm »

I found that this video helps me quite a bit with the maths.

After seeing this I spent about four hours trying to figure out the name for the circle but came up with nothing. :\
Logged

MagmaMcFry

  • Bay Watcher
  • [EXISTS]
    • View Profile
Re: The Maths Thread
« Reply #25 on: April 02, 2012, 02:09:16 pm »

Nobody tried my quiz. :(

Answers:

1. A complete chess board contains an equal number of black and white squares. The two opposite corner squares have the same color, so if they're removed, the new shape contains different numbers of black and white squares. But every domino contains exactly one black and one white square, so after cutting out 30 dominos from the new shape, you are left with two squares of the same color that are therefore not adjacent.

2. Color the bowls alternately red and blue so that the bowl initially containing two bowls is red. Let R denote the number of beans in the red bowls, and let B denote the number of beans in the blue bowls. Then every move increases both R and B by one, so R-B always remains constant, no matter what you do. But in the starting configuration, R-B is 2, and in an end configuration, R-B is 0. So you can't reach an end configuration.

3.
a) Let s_a, s_b, and s_c denote the three angle bisectors of the triangle. Then s_a is the location of all the points inside the triangle with equal distance to b and c, and so on.
Let d(P,g) denote the distance between a point P and a line g.
Now let I denote the intersection of s_b and s_c. I lies on s_b, so d(I,c) = d(I,a). But I also lies on s_c, so d(I,c) = d(I,a) = d(I_b). Therefore I also lies on s_a and is therefore the common point of s_a, s_b, and s_c.

b) Let D, E and F be the midpoints of BC, CA and AB. Then AF = FB, BD = DC, CE = EA. Then AF/FB * BD/DC * CE/EA = 1*1*1 = 1, and Ceva's theorem says that in this case AD, BE and CF intersect in one point.


c) Let AD, BE and CF denote the three altitudes of the triangle. Thales' theorem says that ABDE, BCEF and CAFD are cyclic quadrilaterals. The inscribed angle theorem now says that <EDA = <EBA = <EBF = <ECF = <ACF = <ADF, therefore DA bisects <EDF. Similarly, EB bisects <FED and FC bisects <DFE. So AD, BE and CF are the angle bisectors of DEF. They intersect, see part 3a).

4. Let p be the uniform scaling mapping k' onto k, so that p(C) = C. Now p(AB) is parallel to AB, and also a tangent to k in p(D). Therefore |p(D)A| = |p(D)b|. Using the inscribed angle theorem, it follows that <DCB = <p(D)CB = <ACp(D) = <ACD, therefore CD bisects ACB.

5. Let M be the set of configurations C of 1000 segments that have a red and a blue end each so that every point is an end of a segment in C, let D be the one with the shortest total segment length. Let us assume that there exist two segments PR and QS of D that intersect in T, where P and Q are blue and R and S are red. Now let D' be the configuration that is identical to D, except that D' contains PS and QR instead of PR and QS. Now D' is also in M, and the total segment length of D' is smaller than that of D, which is a contradiction to the assumption that D is the "shortest" configuration in M.
Therefore D can have no intersections.

6. Let us assume that you bought some gumballs but don't have three of the same color. Then you must have at most two of every color, therefore you cannot have bought more than 10 gumballs.
Therefore it follows that if you buy 11 gumballs, you must have three of the same color.

7. Let us define 100 groups labeled 1, 3, 5, 7, ..., 199. Let us divide the set of integers from 1 to 200 into those groups according to their largest odd factor. So if a number n is in a group labeled g, then n is equal to g times a power of 2. So if you pick two numbers out of the same group, then their quotient is a power of 2, therefore one is a multiple of the other.
Now if you pick 101 distinct integers between 1 and 200, you are bound to pick two who are in the same group. Of those two, one is a multiple of the other.

Teaser question: Come on, discuss it already.
Logged

lorb

  • Bay Watcher
    • View Profile
Re: The Maths Thread
« Reply #26 on: April 02, 2012, 02:54:47 pm »

Nobody tried my quiz. :(

I would have tried, if the answers weren't already given. Maybe i'll try some generalizations of those things.

Edit:
2. Color the bowls alternately red and blue so that the bowl initially containing two bowls is red. Let R denote the number of beans in the red bowls, and let B denote the number of beans in the blue bowls. Then every move increases both R and B by one, so R-B always remains constant, no matter what you do. But in the starting configuration, R-B is 2, and in an end configuration, R-B is 0. So you can't reach an end configuration.
Replacing 2 by n does not change anything, so this will work with any (non-zero) amount of beans initially in one bowl.
Also it's obvious that this proofs holds for any number of pots as long as it is possible to color them alternately when they are aligned in a circle, so any even number. So wen can generalize the initial question to:

2*X bowls are standing in a circle. One of them contains N beans, the rest are empty. You may repeatedly choose two adjacent bowls and put one bean into each bowl. Can you reach a state where every bowl contains the same amount of beans? X and N are any non-zero natural numbers. (I'm excluding negative bowls and beans but i suppose it works out the same [if we can figure out how to color a negative amount of bowls xD] )
« Last Edit: April 02, 2012, 03:26:47 pm by lorb »
Logged
Please be gracious in judging my english. (I am not a native speaker/writer.)
"This tile is supported by that wall."
Pages: 1 [2]