SOLUTIONS FOR ADMISSIONS TEST IN

MATHEMATICS, JOINT SCHOOLS AND COMPUTER SCIENCE

WEDNESDAY 5 NOVEMBER 2008
Mark Scheme:

Each part of Question 1 is worth four marks which are awarded solely for the correct answer.

Each of Questions 2-7 is worth 15 marks

QUESTION 1:

A. As y = 2x3 − 6x2 + 5x − 7 then

y ′ = 6x2 − 12x + 5.

The quadratic y ′ has discriminant 122 − 4 × 6 × 5 = 24 > 0 and hence the equation y ′ = 0 has
two distinct real roots. The answer is (c).

B. As π < 10 then
L = log10 π < 1.
So
 3

√ √ 1 1 2
log10 (π2 ) = 2L > L × L = L; = L−3 > 1; √ = > 2.
log10 π log10 π L

The answer is (a).

C. We will write c = cos θ and s = sin θ for ease of notation. Eliminating y from the simulta-
neous equations
cx − sy = 2, sx + cy = 1;
we get  
2c + s = c (cx − sy) + s (sx + cy) = c2 + s2 x = x
and similarly eliminating x we find
 
c − 2s = (−s) (cx − sy) + c (sx + cy) = s2 + c2 y = y.

Hence the equations care solvable for any value of θ. The answer is (a).

D. By the remainder theorem when a polynomial p (x) is divided by x − 1 then the remainder
is p (1). So the required remainder here is
50
1 + 3 + 5 + 7 + · · · + 99 = (1 + 99) = 2500
2
as the series is an arithmetic progression. The answer is (b).

1
 3 4
E. The highest power of x in (2x6 + 7) is x18 and in (3x8 − 12) is x32 so the highest power in
5 5 4
[· · · ]5 is (x32 ) = x160 . The highest power of x in (3x5 − 12x2 ) is x25 and in (x7 + 6) is x28 ,
6
so that the highest power of x in [. . .] is (x28 ) = x168 . Thus the highest power of x in {. . .}3
6
3
is (x168 ) = x504 . The answer is (d).
1
F. Suppose that, when the trapezium rule is used to estimate the integral 0 f (x) dx, an
overestimate of E is produced. If the same number of intervals are used in the following
calculations then: 
1
(a) to estimate 0 2f (x) dx an overestimate of 2E will be produced, as the relevant graphs
have been stretched by a factor of 2 and all areas doubled;
1
(b) to estimate 0 (f (x) − 1) dx an overestimate of E will be produced, as the relevant
graphs have been translated
2 down by 1 and all areas remain the same;
(c) to estimate 1 f (x − 1) dx an overestimate of E will be produced, as the relevant graphs
have been translateright by 1 and all areas remain the same;
1
(d) to estimate 0 (1 − f (x)) dx an underestimate of E will be produced, as the relevant
graphs have been reflected in the x-axis — turning the overestimate to an underestimate — and
translated up by 1, which changes nothing with regard to areas. The answer is (d).

G. As 4x − x2 − 5 = − (x − 2)2 − 1, then y = (4x − x2 − 5)

−1
is always negative and has a
minimum value at x = 2. The answer is (c).

H. If we set y = 3x then the equation 9x − 3x+1 = k now reads

y 2 − 3y − k = 0.

This has solutions √

3± 9 + 4k
y=
2
x
which are real when k
−9/4. As y = 3 then we further need that y > 0 for x to be real, but
this is not a problem as the larger root is clearly positive. The answer is (a).

I. We have

S (1) + S (2) + S (3) + · · · + S (99) = S (00) + S (01) + · · · + S (99)

and in the 100 two-digit numbers 00, . . . , 99 there are twenty 0s, twenty 1s, . . . , twenty 9s. So
10
S (1) + S (2) + S (3) + · · · + S (99) = 20 × (0 + 1 + · · · + 9) = 20 × (0 + 9) = 900
2
and the answer is (c).

J. Note that
(3 + cos x)2
(3 − 1)2 = 4; 4 − 2 sin8 x  4.
So the equation will hold only when cos x = −1 and sin x = 0. In the range 0  x < 2π this
only occurs at x = π. The answer is (b).

2
2. (i) [2 marks] A fairly obvious pair (x1 , y1 ) that satisfy (x1 )2 − 2 (y1 )2 = 1 is x1 = 3 and
y1 = 2.

(ii) [6 marks] Note

(xn+1 )2 − 2 (yn+1 )2 = (xn )2 − 2 (yn )2

⇐⇒ (3xn + 4yn )2 − 2 (axn + byn )2 = (xn )2 − 2 (yn )2
   
⇐⇒ 8 − 2a2 (xn)2 + (24 − 4ab) xn yn + 18 − 2b2 (yn )2 = 0

In order to have (xn+1 )2 − 2 (yn+1 )2 = (xn )2 − 2 (yn )2 we need

2a2 = 8, 4ab = 24, 2b2 = 18.

We further require that a, b > 0. We see that a = 2 and b = 3 solve all three equations.

(iii) [4 marks] Starting with x1 = 3, y1 = 2 we find:

x1 = 3, y1 = 2;
x2 = 3 × 3 + 4 × 2 = 17, y2 = 2 × 3 + 3 × 2 = 12;
x3 = 3 × 17 + 4 × 12 = 99, y3 = 2 × 17 + 3 × 12 = 70.

So X = 99 and Y = 70 is such a pair.

(iv) [3 marks] For the generated sequences, (xn ) , (yn ) , we have

(xn)2 − 2 (yn )2 = 1 for each n.

Also the integers xn and yn are getting increasingly larger because of how they are defined in
(ii). So
 2
xn 1
−2 = ≈ 0 for large n,
yn (yn )2
√
and xn /yn ≈ 2 as xn and yn are both positive.

3
3. (i) [3 marks] (A) is −f (x) ; (B) is f (−x) ; (C) is f (x − 1) .
2 2 2 2
(ii) [9 marks] As 22x−x = 2 × 2−(x−1) then the graph of y = 22x−x is the graph of y = 2−x
translated to the right by 1 and stretched parallel to the y-axis by a factor of 2.
2.0

1.5

1.0

0.5

-2 -1 1 2 3
−x2 2x−x2
Plots of y = 2 and y = 2 .
2 2
(iii) [3 marks] c = 21 . The graph of 2−(x−c) is the graph of 2−x translated c to the right.
The integral I (c) represents the area under the graph between 0  x  1. As the graph is
symmetric/even and decreasing away from 0 then this area is maximised by having the apex
half way along the interval 0  x  1, i.e. at x = 1/2 which occurs when c = 12 .

4
4. (i) [4 marks] We can complete the squares in x2 − px + y 2 − qy = 0 to get
 p 2  q 2 p2 + q 2
x− + y− = (1)
2 2 4
which is the equation of the circle with centre: (p/2, q/2) and area: π (p2 + q 2 ) /4. Either by
checking the original question, or the rearranged one, we can see that

 0 at (0, 0) ,
x2 − px + y 2 − qy = p2 − p2 + 0 = 0 at (p, 0) ,

0 + q 2 − q 2 = 0 at (0, q) .

(ii) [5 marks] The area of OP Q is pq/2. So

   
area of circle C π (p2 + q 2 ) 1 π (p2 + q 2 )
= pq = .
area of triangle OP Q 4 2 2pq

Note
π (p2 + q2 )

π ⇐⇒ p2 + q 2
2pq ⇐⇒ (p − q)2
0,
2pq
proving the required inequality.

(iii) [6 marks] Rearranging

 2  
π (p2 + q 2 ) p p
= 2π ⇐⇒ p2 + q 2 = 4pq ⇐⇒ −4 + 1 = 0,
2pq q q

which is a quadratic equation in p/q, and so

√
p 4 ± 16 − 4 √
= = 2 ± 3.
q 2

Now p/q = tan OQP, q/p = tan OP Q and so

√ √ 
{tan OQP, tan OP Q} = 2 − 3, 2 + 3

with the order depending on whether p < q or p > q.

 √   √ 
[It happens that arctan 2 − 3 = π/12 and arctan 2 + 3 = 5π/12, but appreciation of
this was not expected.]

5
5. (i) [3 marks] After the first/second/third students have gone by the doors look like:

Locker 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Student 1 O O O O O O O O O O O O O O
Student 2 O C O C O C O C O C O C O C
Student 3 O C C C O O O C C C O O O C

We can see that the lockers now repeat in a pattern OCCCOO every 6 lockers. As 1000 =
166 × 6 + 4 we have 166 repeats of this pattern and 4 remaining lockers that go OCCC. So there
are 166 × 3 = 498 closed lockers amongst the complete cycles and 3 further in the incomplete
cycle. That is, there are 501 closed lockers in all.

(ii) [4 marks] After the fourth student has gone by we have the following:

Locker 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Student 3 O C C C O O O C C C O O O C
Student 4 O C C O O O O O C C O C O C

with the pattern repeating every 12 lockers in the form OCCOOOOOCCOC. Each cycle con-
tains 5 closed and 7 open doors. Now 1000 = 83 × 12 + 4 and so we have 83 × 5 = 415 closed
lockers amongst the complete cycles and 2 further amongst the incomplete cycle OCCO. In all
then there are 417 closed lockers.

(iii) [4 marks] Locker 100 starts off closed (as all lockers do) and then its state is altered by
every nth student where n is a factor of 100, i.e. by students 1, 2, 4, 5, 10, 20, 25, 50, 100. So 9
students change the state and as this is odd then overall the state will have been changed to
open.

(iv) [4 marks] Locker 1000 starts off closed (as all lockers do) and then its state is altered by
every nth student where n divides 1000 and n  100, i.e. by 1, 2, 4, 5, 8, 10, 20, 25, 40, 50, 100.
So 11 students change the door’s state and as this is odd then overall the state will again have
been changed to open.

6
6. (i) [5 marks] We have six possibilities:

A-B-C = St-L-Sw, St-Sw-L, L-St-Sw, L-Sw-St, Sw-L-St, Sw-St-L.

The statement "I am the liar" cannot be made by St or L; this excludes the first four possibilities
above.

The second statement "A is the liar" excludes Sw-St-L and so we are left with Sw-L-St. Answer:
B is the Liar.

(The third statement is not actually needs but doesn’t contradict the Sw-L-St arrangement.)

(ii) [5 marks] We have six possibilities:

P-Q-R = S-L-C, S-C-L, L-S-C, L-C-S, C-L-S, C-S-L.

One of these statements is from a saint and so true. This means that the Liar has to follow the
Saint in cyclic order and this means the only remaining possibilities are

P-Q-R = S-L-C, L-C-S, C-S-L.

In the first two cases the Contrarian follows the Liar and so tells the truth. But this contradicts
the actual statements so the only possibility remaining is C-S-L. Answer: R is the Liar.

(iii) [5 marks] We have six possibilities:

X-Y-Z = S-L-C, S-C-L, L-S-C, L-C-S, C-L-S, C-S-L.

We will take these case by case:

• S-L-C: As the Contrarian is following the Liar, statement 3 had to be true but isn’t in
this case.

• L-C-S: As the Contrarian is following the Liar, statement 2 had to be true but isn’t in
this case.

• C-S-L: As the Contrarian is following the Liar, statement 4 had to be true but isn’t in
this case.

• S-C-L: In this case, statement 4 is a lie and so the Contrarian would tell the truth in
Statement 5 but doesn’t.

• C-L-S: The Contrarian tells the truth to begin contradicting his nature.

• L-S-C: This is the only remaining case and is consistent.

Answer: X is the Liar.

7
7. (i) [3 marks] The empty word has zero length which is even. If a new word is formed by Rule
2 then aW b will have the same parity of length as W had. Also if U and V are even-length
words then so will be UV . So new words formed from words of even length will themselves be
even.

(ii) [5 marks]
Length 0 words: ∅.
Length 2 words: ab.
Length 4 words: abab, aabb
Length 6 words: ababab, abaabb, aabbab, aababb, aaabbb

(iii) [3 marks] In ∅ there are the same number of as and bs, namely none. If W has the same
number then so will aW b, formed by Rule 2. Also if U and V each have the same number of as
and bs then so will U V . So new words formed by Rules 2 and 3 always have the same property.

(iv) [4 marks] A word of the form aW bW ′ will be of length 2n + 2 if

length (W ) + length (W ′ ) = 2n.

So if W has length 2k  2n then W ′ has length 2 (n − k). There are Ck words of the former
length and Cn−k of the latter length. So we may generate Ck Cn−k such words of length 2n + 2
in this manner for each k. That is,
n
Ck Cn−k
k=0

in all. Further, because the uniqueness of form in the given hint, all words of length 2n + 2 are
counted amongst these words and none are doubly counted. That is
n

Cn+1 = Ck Cn−k
k=0