Ranking throws

Jason throws two darts at a dartboard, aiming for the center. The second dart lands farther from the center than the first. He throws a third dart. What's the probability the third is also farther from the center than the first?

What do you think?
The 2nd dart is farther from center than the 1st. P(3rd dart is also farther than the 1st)?

Enumerating all rankings

Label the quality of each throw A (best), B (middle), C (worst). There are 6 possible orderings:

Ranking1st2nd3rd2nd worse than 1st?3rd worse than 1st?
1ABCyesyes
2ACByesyes
3BACno
4BCAyesno
5CABno
6CBAno

All 6 rankings are equally likely. The condition "2nd is worse than 1st" eliminates rankings 3, 5, and 6. Of the remaining three (rankings 1, 2, and 4), the 3rd dart is worse than the 1st in two of them.

P(3rd worse than 1st2nd worse than 1st)=23P(\text{3rd worse than 1st} \mid \text{2nd worse than 1st}) = \frac{2}{3}

Dart Ranking

Jason throws 3 darts. The 2nd is farther from center than the 1st. What's the probability the 3rd is also farther than the 1st?

A faster argument

The enumeration approach works for 3 darts but gets unwieldy for more. There's a better way to think about it.

The independence argument
The 1st dart is the best among the first 2\text{The 1st dart is the best among the first 2}
This is what '2nd is worse than 1st' means.
Step 1 of 4

The event "3rd dart is the best of all three" is independent of the relative ordering of the first two. Whether the 1st beat the 2nd or not, the 3rd still has a 1/3 chance of being the overall best. This independence is what makes the generalization work.

The general case

Jason throws nn darts. Each subsequent dart lands farther from the center than the first. He throws one more. What's the probability it's also farther than the first?

Ranking independence

The event that the (n+1)(n+1)-th throw is the best among all n+1n+1 throws is independent of the relative ordering of the first nn throws. Each throw has probability 1/(n+1)1/(n+1) of being the overall best, regardless of any conditioning on the order of the others.

P(last dart worse than 1stall others worse than 1st)=11n+1=nn+1P(\text{last dart worse than 1st} \mid \text{all others worse than 1st}) = 1 - \frac{1}{n+1} = \frac{n}{n+1}

For n=2n = 2 (the original problem): P=2/3P = 2/3. For n=5n = 5: P=5/6P = 5/6. As nn grows, you become more confident that the first dart is the best, but there's always a 1/(n+1)1/(n+1) chance the next one beats it.

What do you think?
Jason throws 10 darts. Each one after the first is farther from center. He throws an 11th. P(it's also farther than the 1st)?

Practice

3 darts, the 2nd is closer to center than the 1st. P(3rd is also closer than the 1st)? (decimal like 0.67)
1/3