11/05 Weekly Problem

Daily Problem 11/05

2006 Alabama ARML TST Problem 15

Ying lives on Strangeland, a tiny planet with 4 little cities that are each 100 miles apart from each other.

One day, Ying begins driving from her home city of Viavesta to the city of Havennew, which takes her about an hour. When she gets to Havennew, she decides she wants to go straight to another city on Strangeland, so she randomly chooses one of the other three cities (possibly Viavesta), and starts driving there.

Ying drives like this for most of the day, making 8 total trips between cities on Strangeland, choosing randomly where to drive to next from each stop. She then stops at her final city of destination, digs a hole, and buries her car.

Let p be the probability Ying buried her car in Viavesta and let q be the probability she buried it in Havennew. Find the value of p + q.

Saen’s Answer

Prerequisites

Let’s denote the cities as A, B, V, and H and establish what we know already.

  • We know that the second journey ends in H.
  • On this basis, it is impossible for the third journey to end in H, since we are travelling from H.
  • We have 4 cities in total, and from each city, we could travel to each of the other 3.
  • On this basis, for each journey point, there is a \frac{1}{3} chance of travelling to each other city. For instance, if we are at A, then we can either travel to B, H or V.
  • The probability of being in any of those cities after the third journey is \frac{1}{3}.

Finding the probability distribution of H across multiple journeys.

The journey will follow the prototype V-H-X-X-X-X-X-X-H.

X here simply denotes an unknown city, although it’s important to note that each X must be different to the X preceding it. Also note how there are 8 dashes representing 8 journeys, so we actually have a total of 9 cities in our path.

At first glance it seems like we could just distribute the probability over \frac{1}{3} for each movement. But this wouldn’t work since we have the first and second steps predetermined.

In order to work out the probability of being in H after each movement, we need to take the probability of not being in H from the previous movement, and then distribute this over 3 branches.

Since the last known point is city H, we should try and work out the probability of being in city H after each move.

  • We are predetermined to be in H after the first movement, as shown in the prototype.
  • This means that the probability of being in H after the second movement is 0, since we are departing city H.
  • From this point onwards the exact movement pattern is indeterminate. After the third move, we have a \frac{1}{3} chance of being in H.
  • After the fourth move is where it starts to get a little more involved. We will need to get the probability that we were not in H beforehand (since we can’t travel from H to H) and then apply 3 possible branches to this probability. The probability of not being in H before the move is \frac{2}{3}, and we have a \frac{1}{3} chance of moving anywhere else except this city. This gives us a \frac{2}{3} \times \frac{1}{3} = \frac{2}{9} chance of being in H.
  • The step we did above now needs to be repeated, until we reach the end of the chain.

Each time we are essentially performing the following steps:

  • calculate the probability of not being in H.
  • divide this by 3 to account for the branching distribution.

Algebraic Formation of Distribution

We can write this algebraically.

If X_n is the probability of being in H after n journeys, then

Xn+1=1.0−Xn3X_{n+1} = \frac{1.0 - X_n}{3}Xn+1​=31.0−Xn​​

We are making 8 journeys, so we will calculate the value of X_8.

X_1 1
X_2 0
X_3 \frac{1 - 0}{3} = \frac{1}{3}
X_4 \frac{1 - \frac{1}{3}}{3} = \frac{2}{9}
X_5 \frac{1 - \frac{2}{9}}{3} = \frac{7}{27}
X_6 \frac{1 - \frac{7}{27}}{3} = \frac{20}{81}
X_7 \frac{1 - \frac{20}{81}}{3} = \frac{61}{243}
X_8 \frac{1 - \frac{61}{243}}{3} = \frac{182}{729}

By following the distribution appropriately, we can see that there are 182 possibilities out of 729 where we will end up in city H. Therefore, probability p = \frac{182}{729}.

The Second Probability

Keep in mind that the probability of each city being chosen per-turn is equally distributed. What isn’t equally distributed is the probability of a single city at a single point overall, since our starting data of V and H is predetermined.

Now we can use this to work out probability q.

If X_8 is \frac{182}{729}, then the probability of being in any of the other three cities is 1 - \frac{182}{729} which is \frac{547}{729}.

Final Steps

We can now distribute this across three choices, giving us \frac{547}{729} / 3 which is \frac{547}{2187}. Therefore probability q = \frac{547}{2187}.

Our final step for this question is to add these together to find probability p + q.
p + q = \frac{182}{729} + \frac{547}{2187} = \boxed{\frac{1093}{2187}}

Our final answer is therefore \boxed{\frac{1093}{2187}}.