A frog hops along the number line according to the following rules.
It starts at .
If it is at , then it moves to with probability and it disappears with probability .
For or if it is at then it moves to with probability it moves to with probability and it disappears with probability .
What is the probability that the frog reaches
- A)
- B)
- C)
- D)
- E)
Answer
E
Key insight
Let p1, p2, p3 be the chances of reaching 4 from 1, 2, 3. Each is a quarter of the chances one step down and one step up.
Solution
The frog can wander back and forth many times before reaching , so listing paths would never end. Instead, give each starting spot its own unknown.
- Name the unknowns. Let be the chance of eventually reaching if the frog is now at . Then , and we want .
2. Look one hop ahead.
- From : half the time it moves to , otherwise it disappears. So .
- From : a quarter of the time it goes to , a quarter of the time to , and otherwise it disappears. So .
- In the same way, and .
3. Write everything using , working upward.
- Multiply the second equation by : . So .
- Multiply the third by : . So .
4. Use the last equation to find . Multiply it by : . Substitute:
- Finish. .
The answer is .
Why this works
What happens next depends only on where the frog is now, not on how it got there. So one unknown for each spot, plus "what can the next hop do?", replaces an endless list of back-and-forth paths with four short equations. Writing each unknown as a multiple of from the bottom up leaves one equation with one unknown at the end.
Alternative approach
Climb one step at a time. Let be the chance that a frog at ever reaches .
- From there is no way back, so .
- From : it steps up right away with chance ; or it steps down (chance ), must climb back to (chance ), and then is in the same spot as before (chance ). So , which rearranges to .
This gives , , . To reach the frog must climb every step, so
The trap
Counting only the straight path 0 → 1 → 2 → 3 → 4, which has probability 1/2 × (1/4)³ = 1/128. The frog can also step back and try again, and those paths are what change 128 into 97.
Concise solution
For let be the probability that the frog ever reaches , given that it is sitting at . Then , and disappearing contributes . Each hop has probability at least of ending the walk, so the frog is certain to disappear or reach eventually and these probabilities are well defined.
Condition on the next hop:
Rather than eliminate blindly, express everything in terms of , working up the line.
From the second equation, , so
From the third, , so
From the fourth, , so
Finally .
The answer is .
Review in the AoPS books
- Introduction to Counting & Probability, Chapter 7: Introduction to Probability
- Introduction to Algebra, Chapter 5: Multi-Variable Linear Equations (§5.2, §5.3, §5.6)
- Intermediate Counting & Probability, Chapter 10: Recursion
- Introduction to Algebra, Chapter 21: Sequences & Series (§21.5)
Techniques
Define states/recurrence and iterate · Collapse a sum or product by cancellation