Serving the Quantitative Finance Community

 
User avatar
e04bf048
Topic Author
Posts: 0
Joined: June 11th, 2009, 9:31 pm

3 dice

June 15th, 2009, 3:44 pm

right, I heard this before that if you have 3 dice and you play a game with someone and you choose 1 of the 3 die and the other person chooses 1 of the other 2 that he has a greater expected return (given that you get paid the amount of the number you roll).why?the only way i can reason it is that i'm multiplying by 1/3 (choosing 1 out of 3 die) where he is only multiplying by 1/2. Am i right in this even though i know dice are memoryless and it shouldn't matter which dice you have.
 
User avatar
karakfa
Posts: 0
Joined: May 25th, 2002, 5:05 pm

3 dice

June 15th, 2009, 3:57 pm

This works only for special dice. I guess it was on Scientific American some years ago. You can design the die such thata wins against b, b wins against c, and c wins against a. Note that the expected values will be the same but it's not the same as winning against.Exercise to you is to set up the die, hint use distinct numbers to eliminate ties.
 
User avatar
karakfa
Posts: 0
Joined: May 25th, 2002, 5:05 pm

3 dice

June 15th, 2009, 8:58 pm

More hints:For a 3 sided die (a rounded-off triangular prism will do) use these values:a: 1,6,8b: 2,4,9c: 3,5,7here on average b beats a, c beats b, and a beats c.
 
User avatar
wileysw
Posts: 7
Joined: December 9th, 2006, 6:13 pm

3 dice

June 17th, 2009, 5:36 am

just to add some reference: this set of dice is so-called nontransitive dice