Register  
 
About Us | Help | Sign in
 
   

STEP III 2007 question 14 solution

From The Student Room

TSR Wiki > Study Help > Subjects and Revision > Mathematics > STEP > STEP III 2007 question 14 solution


(i) Consider the circle C of radius R (<=1) at the origin. The chance of a dart hitting C is \frac{\pi R^2}{\pi} = 
R^2. So the chance of n darts all hitting C is R^{2n}. But this is exactly the chance of the maximum radius r being \leq R. So P(r\leq R) = R^{2n}. So r has p.d.f. \frac{\text{d}}{\text{d} R}R^{2n} = 2nR^{2n-1}. Then the expected area is \displaystyle \int_0^1 2nR^{2n-1} \pi R^2 \text{ d} R = \frac{\pi n}{n+1}.

In the case of the n-1 nearest darts, the chance of n-1 darts lying within a radius R is p(n darts lie within) + p(n-1 darts lie within) = R^{2n}+ nR^{2n-2}(1-R^2) = nR^{2n-2}-(n-1)R^{2n}. Again differentiate to get the p.d.f. 2n(n-1)R^{2n-3}-2n(n-1)R^{2n-1} = 2n(n-1)(R^{2n-3}-R^{2n-1}).

Then the expected area is \displaystyle 2n(n-1)\pi\int_0^1 (R^{2n-3} -R^{2n-1})R^2 \text{ d} R = n(n-1)\pi \left[\frac{1}{n}-\frac{1}{n+1}\right] = \frac{n-1}{n+1}\pi.

(ii) Now define S to be the square with sides 2M centered at the origin. The chance of a dart hitting S is M^2, so as in (i) we find the minimum square side m has p.d.f. 2nR^{2n-1}. Then the expected area is \displaystyle \int_0^1 2nR^{2n-1} 4M^2 \text{ d} R = \frac{4n}{n+1}.

(iii) Clearly greater, since the alllowed 'target' for the square dartboard includes the entirety of the 'target' for the circular one. For each dart, there's a chance \pi/4 of hitting the target in (i) with the same expectation, and a chance 1-\pi/4 of going outside the target in (i) with a resultant expected area > \pi (which is greater than the expected area in (i)), so the expectation for the area must be greater than in (i). Similarly for considering [i]n[/i] darts.

Comment: At first, I took (iii) to be exclusively considering the square dartboard and comparing the area of the bounding circle to the area of the bounding square (which is a complete misread of the question - doh!). But it's quite an interesting question: it's clear the square is better for large n (since you will typically end up with at least one point near the corners of the dartboard, and bounding the corners gives a square of area 4 but a circle of area 2\pi), and simulation (or calculus) shows it's also better for small n. But I don't see a way of doing this without calculation. (Of course, there may not be one - it was only misreading the question that had me trying to find one after all!).

Solution by DFranklin.

collapse
Recent Threads
 
collapse Is It Much Harder To Get A 1st or 2:1 From Imperial?
started by: Snobpence17
replies: 7
last post: 1 Minute Ago
collapse Chances
started by: billykwok
forum: Medicine
replies: 4
last post: 1 Minute Ago
collapse Integration
started by: RiverDog
forum: Maths
replies: 0
last post: 1 Minute Ago
collapse Would you personally shop at Lidl or Aldi?
started by: burninginme
replies: 41
last post: 1 Minute Ago
 
Article Updates