Hard Candy Mixer

Standard memberRemoved
Posers and Puzzles 19 Apr '21 12:09
  1. R
    Standard memberRemoved
    Joined
    10 Dec '06
    Moves
    8528
    19 Apr '21 12:09
    Jar 𝐀 has 800 red candies
    Jar 𝐁 has 600 green candies

     𝟏) Move one cup of red candies from jar 𝐀 to jar 𝐁 and mix.

     𝟐) Move one cup of mixed candies from jar 𝐁 to jar 𝐀 and mix.

     πŸ‘) Move one cup of candies from each of jar 𝐀 and jar 𝐁 to a bowl and mix.

     πŸ’) Randomly select one candy from the bowl.

    Is the selected candy from the bowl more likely to be red or green?

    For clarity, "one cup" effectively means the same number of candies are transferred at each step.
  2. RSA
    Joined
    20 Oct '16
    Moves
    11569
    19 Apr '21 20:393 edits
    @joe-shmo said
    Jar 𝐀 has 800 red candies
    Jar 𝐁 has 600 green candies

     𝟏) Move one cup of red candies from jar 𝐀 to jar 𝐁 and mix.

     𝟐) Move one cup of mixed candies from jar 𝐁 to jar 𝐀 and mix.

     πŸ‘) Move one cup of candies from each of jar 𝐀 and jar 𝐁 to a bowl and mix.

     πŸ’) Randomly select one candy from the bowl.

    Is the selected candy from the bowl more likely to be red or g ...[text shortened]...
    For clarity, "one cup" effectively means the same number of candies are transferred at each step.
    Well, red should always be more likely, but the proportion of red in the final bowl should increase with the "cup size".

    We can use the following expected value function:

    nM/N

    Where:
    n is the cup size
    M is the number of red or green candies in the jar
    N is the total number of candies in the jar

    This finds the mean number of the candies of your choice that will be in each transfer. Using the means eliminates the issue of random variation for the purposes of this question. Obviously you don't need the expected value function for the first transfer (when all the candies will be red).

    Using that function it can be shown that when a cup is set to the minimum size (1), the expected value of red candies in the final bowl is 2405/2404 and the expected value of green candies is 2403/2404, which gives probability of 2405/4808 for red.

    When the cup size is set to the maximum (600), the expected value of red candies in the final bowl is 675 versus 525 for green, so a probability of 9/16.

    Edit: Fixed some typos (wrote this on my phone in bed - sorry!)
  3. R
    Standard memberRemoved
    Joined
    10 Dec '06
    Moves
    8528
    20 Apr '21 14:241 edit
    @ashiitaka said
    Well, red should always be more likely, but the proportion of red in the final bowl should increase with the "cup size".

    We can use the following expected value function:

    nM/N

    Where:
    n is the cup size
    M is the number of red or green candies in the jar
    N is the total number of candies in the jar

    This finds the mean number of the candies o ...[text shortened]... green, so a probability of 9/16.

    Edit: Fixed some typos (wrote this on my phone in bed - sorry!)
    Correct!

    I chose to follow the the expected value of green candy in each cup.

    let:

      π‘₯ = cup "size"
      𝑃( 𝐺_a ) = probability of drawing a green candy in jar 𝐀 after each step
      𝑃( 𝐺_b ) = probability of drawing a green candy in jar 𝐁 after each step

    𝟏) Move one cup of red candies from jar 𝐀 to jar 𝐁 and mix.

             𝑃( 𝐺_b ) = 600 / ( 600 + π‘₯ )

    𝟐) Move one cup of mixed candies from jar 𝐁 to jar 𝐀and mix.

    We are again moving back π‘₯ candies into jar 𝐀, so there will be 800 candies in the jar.

    The expected value for the number of greens being mixed in from the transfer:

             𝐸( 𝐺_b ) = π‘₯ ‧ 600 / ( 600 + π‘₯ )

          𝑃( 𝐺_a ) = 𝐸( 𝐺_b ) / 800 = 3/4 ‧ π‘₯ / ( 600 + π‘₯ )

    πŸ‘) Move one cup of candies from each of jar 𝐀 and jar 𝐁 to a bowl and mix.

            𝐸( 𝐺_a ) = π‘₯ ‧ ( 3/4 ‧ π‘₯ / ( 600 + π‘₯ ) )

             𝐸( 𝐺_b ) = π‘₯ ‧ 600 / ( 600 + π‘₯ )

    πŸ’) Randomly select one candy from the bowl.

    Finally the probability of drawing a Green out of the bowl is then:

           𝑃( 𝐺_bowl ) = ( 𝐸( 𝐺_a ) + 𝐸( 𝐺_b ) ) / ( 2π‘₯ )

                 = 1/2 ( 3/4 π‘₯ + 600 ) / ( π‘₯ + 600 )

    We can see that for all π‘₯ > 0, 3/4 π‘₯ + 600 < π‘₯ + 600 implies that ;

                𝑃( 𝐺_bowl ) < 1/2

    Thus its is more likely we draw red from the bowl .

Cookies help us deliver our Services. By using our Services or clicking I agree, you agree to our use of cookies. Learn More.I Agree