CTK Exchange
Front Page
Movie shortcuts
Personal info
Awards
Reciprocal links
Privacy Policy

Interactive Activities

Cut The Knot!
MSET99 Talk
Games & Puzzles
Arithmetic/Algebra
Geometry
Probability
Eye Opener
Analog Gadgets
Inventor's Paradox
Did you know?...
Proofs
Math as Language
Things Impossible
My Logo
Math Poll
Other Math sit's
Guest book
News sit's

Recommend this site

Manifesto: what CTK is about |Store| Search CTK Buying a book is a commitment to learning Table of content Things you can find on CTK Chronology of updates Email to Cut The Knot Recommend this page

CTK Exchange

Subject: "probability"     Previous Topic | Next Topic
Printer-friendly copy     Email this topic to a friend    
Conferences The CTK Exchange Middle school Topic #75
Reading Topic #75
Amit Chauhan
guest
Dec-04-02, 09:17 AM (EST)
 
"probability"
 
   A stick of length l is divided randomly into 3 parts. What is the probability that a triangle can be made with the three parts ?


  Alert | IP Printer-friendly page | Edit | Reply | Reply With Quote | Top
alexb
Charter Member
897 posts
Dec-04-02, 09:52 AM (EST)
Click to EMail alexb Click to send private message to alexb Click to view user profileClick to add this user to your buddy list  
1. "RE: probability"
In response to message #0
 
   Search this site or other places for "Barycentric coordinates".


  Alert | IP Printer-friendly page | Edit | Reply | Reply With Quote | Top
vvangogh
guest
Dec-18-02, 05:21 PM (EST)
 
2. "RE: probability"
In response to message #0
 
   my strictly amateur guess would go as follows;
replace the stick with line segment AD. Now select at random 1 point (B)on line segment AD. Now we have 3 possibilities.
1. AB > BD
2. AB < BD
3. AB = BD
In an infinite number of selections, avg(AB) = avg(BD).

Now select a 2nd random point (C) on line segment AD.
Clearly it must fall between A and B or between B and D.
So,
Since avg(AB) = avg(BD), in an infinite number of selections, point C should fall between A and B 50% of the time, and between B and D 50% of the time.

Now, for every case, we have AD = AB+BC+CD or AD = AC+CB+BD.
Also we can only make a triangle if no segment >1/2 AD.

In virtually every case, AB<>BD, so either AB or BD >1/2 AD. So, if point C falls equally on either side of point B, it must, as the number of trials approaches infinity, fall half the time on the shorter segment, leaving the other segment >1/2 AD. Therefore I would conclude the answer to be 1/2, or .5.

I would appreciate if someone could confirm this result, or demonstrate an error in my logic.
Thank You.


  Alert | IP Printer-friendly page | Edit | Reply | Reply With Quote | Top
alexb
Charter Member
897 posts
Dec-18-02, 05:23 PM (EST)
Click to EMail alexb Click to send private message to alexb Click to view user profileClick to add this user to your buddy list  
3. "RE: probability"
In response to message #2
 
   Do compare your derivation with

https://www.cut-the-knot.com/triangle/geoprobability.shtml


  Alert | IP Printer-friendly page | Edit | Reply | Reply With Quote | Top
Order
Member since Dec-11-02
Dec-19-02, 01:25 AM (EST)
Click to EMail Order Click to send private message to Order Click to add this user to your buddy list  
4. "RE: probability"
In response to message #2
 
   Sorry, but your logic does have a flaw. You correctly state that the probability that point C lies in AB is 50%. However, the probability that point C lies in AB *given that AB is the shorter segment* is considerably smaller.

To save space, let E denote the event that AB is the shorter segment. Then we want to find P(C in AB | E). From the rules of probability, we know that this is P(C in AB and E) / P(E). Now the probability of E is just 50% since as you point out AB and BD are equally likely to be the shorter segment.

The probability that C is in AB and E occurs is the probability that B lies to the left of the midpoint of AD and so does C and C < B. Since B and C are chosen independently, the probability that both lie to the left of the midpoint of AD is just (1/2)(1/2) = 1/4. Furthermore, once we condition on B and C being left of the midpoint of AD, there is a 50% chance that B < C and 50% chance that C > B. So the total probability that C lies in AB and E occurs is (1/2)(1/2)(1/2) = 1/8. Thus P(C in AB | E) = (1/8) / (1/2) = 1/4.

Whew! Of course, that doesn't solve Amit's orginal problem, it just shows why your argument doesn't quite work.

To solve the orignal problem, first let me copy the Description of the problem from the CTK page https://www.cut-the-knot.com/triangle/geoprobability.shtml:
Two points are selected at random on a straight line segment of length 1. What is the probability that a triangle can be constructed out of thus obtained three segments?

To be precise, I'll mean that the two points are selected uniformly at random from 0 to 1. I'll call the two points B and C as you did. The only way that the resulting three pieces do not make a triangle is if one of the pieces has length at least 1/2 (this is equation (2) of the CTK Barycentric solution.) There are three ways this can happen. Either (B < 1/2 and C < 1/2) or (B > 1/2 and C > 1/2) or (|B - C| > 1/2). These are disjoint events (that is, only one can happen) and so to find the total probability, we add the probabilities of the three events. The probability that B < 1/2 and C < 1/2 is just (1/2)(1/2) = (1/4), and the probability that B> 1/2 and C > 1/2 is also (1/2)(1/2) = (1/4).

To find the probability of the last event, picture a square with B plotted on the x-axis and C plotted on the y-axis. Then for |B - C| > 1/2, we must have either B - C > 1/2 or C - B > 1/2. The event that (B - C > 1/2) is the lower right hand corner of the square, and (C - B > 1/2) is the upper left hand corner. Each of these are triangles, and so we can find their total area to be 1/4.

Hence there is a 1/4 + 1/4 + 1/4 = 3/4 chance that the resulting three piece DO NOT form a triangle. Hence the chance that they do is just 1 - 3/4 = 1/4, the same probability as found on the Barycentric page.

Mark


  Alert | IP Printer-friendly page | Edit | Reply | Reply With Quote | Top

Conferences | Forums | Topics | Previous Topic | Next Topic

You may be curious to visit the old CTK Exchange archive.

|Front page| |Contents| |Store|

Copyright © 1996-2018 Alexander Bogomolny

[an error occurred while processing this directive]
 Advertise