Sixth Term Examination Paper [STEP]

Mathematics 3 [9475]

2021

Examiner’s Report

Mark Scheme
STEP MATHEMATICS 3
2021
Examiner’s Report
 STEP 3 2021

Introduction

The total entry was a marginal increase from that of 2019, that of 2020 having been artificially
reduced. Comfortably more than 90% attempted one of the questions, four others were very
popular, and a sixth was attempted by 70%. Every question was attempted by at least 10% of the
candidature.
85% of candidates attempted no more than 7 questions, though very nearly all the candidates made
genuine attempts on at most six questions (the extra attempts being at times no more than labelling
a page or writing only the first line or two).
Generally, candidates should be aware that when asked to “Show that” they must provide enough
working to fully substantiate their working, and that they should follow the instructions in a question,
so if it says “Hence”, they should be using the previous work in the question in order to complete the
next part. Likewise, candidates should be careful when dividing or multiplying, that things are positive,
or at other times non-zero.
Question 1

This was the most popular question by a fair margin, being attempted by 93%, and equally was
comfortably the most successful with a mean mark of slightly over 15/20. Generally, most found the
equation of the normal in part (i) correctly, though the more successful candidates simplified their
answer sensibly at this point and similarly with other results in the question. A number of
candidates forgot the negative sign when obtaining a perpendicular gradient and merely attempted
to use the reciprocal. Most used implicit differentiation in order to arrive at an expression for the
gradient of the tangent to the second curve in part (i), though parametric differentiation was
probably simpler. There was an equal split between those that obtained the equation of the tangent
to the second curve and demonstrated that it was the same as that for the normal to the first curve,
and those that demonstrated that the point given parametrically was on the normal and that the
gradient of the normal and the tangent were the same.
In part (ii), surprisingly, some candidates made errors with the initial differentiation. Those that
simplified their equation of the normal profited from the easier working, whichever way they then
tried to obtain the perpendicular distance. About three quarters of the candidates found this
distance by first finding the intersection of the normal with a perpendicular line through the origin.
However, using the formula for the perpendicular distance of a point from a line was simpler. A
range of other methods for this distance were seen; briefly, these were (a) simple trigonometry
having sketched the normal, the axes and line’s intercepts, (b) expressing the normal equation as the
scalar product of vectors, (c) minimising by differentiation, or completing the square, of the distance
of a general point on the normal from the origin or (d) by equating two expressions for the area of
the triangle formed by the normal and the two axes. Errors in this part arose from unsimplified
working complicating the issue (as already mentioned), overlooking the modulus sign in the distance
formula, or calculating the distance from the origin to a point on the curve. The final requirement
for the equation of a curve to which the normal found is a tangent was either not spotted by some
candidates who had otherwise answered the question perfectly, or the requirement was
overlooked.
Question 2

This was the fourth most popular question being attempted by very nearly four fifths of the
candidates. It was the third most successful with a mean mark of just over 9/20, though very few
achieved full marks. With four “Show that”s, marks were frequently lost for lack of proper
justification, and with inequalities to demonstrate involving fractional quantities, positivity was often
not considered, let alone proved, or stated as relevant. Even if candidates stated det(M)=0, which
they sometimes didn’t, only a minority of candidates realised that they had to justify using det(M)=0,
and of these only some could do so convincingly; there were a number of incorrect arguments used.
𝑎𝑎2 𝑏𝑏 2
Some candidates sacrificed marks by, for example, attempting to show that (𝑏𝑏−𝑐𝑐)2 + (𝑐𝑐−𝑎𝑎)2 +
𝑐𝑐 2
(𝑎𝑎−𝑏𝑏)2
≥ 2 via a purely algebraic approach rather than using the result just found (i.e. ignoring the
“hence”).

For the very last part of the question, candidates used a variety of methods in order to explain why
𝑥𝑥 + 𝑦𝑦 + 𝑧𝑧 > 2. The most common method was to express the sum in terms of a, b and c and then
show that this was greater than 2, but approaches using the AM-GM inequality or by splitting into
different cases were sometimes used successfully.
Question 3

Whilst this was the second most popular question, being attempted by 84%, it was the fifth most
successful with a mean mark a little below 8/20. Most candidates scored full marks for successfully
obtaining the first result of part (i), and many gained nearly full credit for obtaining the second result
of that part. As is nearly always true, the rule of thumb that it is usually easier to prove that
something is greater than (or less than) zero applied here, and so those that considered
1
(𝐼𝐼𝑛𝑛+1 + 𝐼𝐼𝑛𝑛−1 ) − 𝐼𝐼𝑛𝑛 (and a similar expression for part (ii)) generally fared better. A small, but not
2
insignificant number of candidates solved part (ii) by a direct method and were generally successful
if they did so. Common errors when considering inequalities were failure to fully justify positivity of
integrals in both parts, incorrect flows of logic, obtaining weak rather than strict inequalities, and
stating inequalities that were inconsistent with the claimed ranges of validity. Otherwise, use of
induction or integration by parts caused difficulties, and a number expected, when replicating the
first part of working in (ii) from part (i), that there would again be an equation, and overlooked the
extra term that arose in (ii). Some did not understand that 𝑠𝑠𝑠𝑠𝑠𝑠 𝑥𝑥 𝑐𝑐𝑐𝑐𝑐𝑐 β ≤ 1 in (ii).
Question 4

Comfortably the least popular Pure question on the paper, it was attempted by just very slightly
more than a third of the candidates, which made it almost exactly the same popularity as the most
popular Probability and Statistics question. With a mean score of less than 7/20, it was seventh
most successful. Those candidates who engaged with the given definition of projection and followed
the structure of the question generally did correct calculations of dot products and recognised the
relevance of their calculations. Several candidates assumed properties of a projection, not realising
that the purpose of this question was to prove properties of a projection given only a single
definition. Many of these implicitly made the assumptions when drawing geometric diagrams and
arguing geometrically.
Question 5

A handful of candidates more attempted this question than question 2, but with marginally less
success than question 4. Nearly every candidate obtained the very first result and many then
1
obtained 𝑎𝑎 = from considering the discriminant. Finding the other values of 𝑎𝑎 (1 and 5) caused
2
many candidates difficulty which could have been overcome had they considered equating
𝑑𝑑𝑑𝑑
expressions for . In the diagrams, the curve representing the second equation was often drawn
𝑑𝑑𝑑𝑑
as an ellipse, or with cusps rather than smooth indentations. On the other hand, touching points
were usually well drawn. It seemed that many appreciated that the curves had symmetry but
seldom referred to this in their justification. Similarly, many might have earned credit, but didn’t,
for indicating values of 𝑟𝑟 for important points such as where the curves met the initial line or the line
perpendicular to it. Few candidates found the angles of the cusp in the first two cases (especially
1 1
with struggling to deal with arccos �− �, as opposed to arccos �− � ).
4 2
Question 6

The seventh most popular question, it was attempted by almost 70% of candidates. However, it was
fourth most successful with a mean just short of 8/20. Most candidates successfully differentiated
𝑓𝑓𝛼𝛼 correctly to obtain the required result. Many then sketched a shifted arctan graph but frequently
failed to appreciate that there were two branches to the curve with a discontinuity at 𝑥𝑥 = tan 𝛼𝛼 ,
𝜋𝜋 𝜋𝜋
and also often forgot that the range of the function is �− , � . In addition, few candidates labelled
2 2
all the requisite values of intercepts, the discontinuity, the asymptote, and the range on the axes.
Few consequently sketched 𝑓𝑓𝛼𝛼 (𝑥𝑥 ) − 𝑓𝑓𝛽𝛽 (𝑥𝑥 ) correctly. In part (ii), many candidates incorrectly
manipulated the negative sign when differentiating g, which then meant that although they
𝜋𝜋 3𝜋𝜋
sketched the section of the graph for < 𝑥𝑥 < , they did not wonder why the negative sign arose
2 2
and hence failed to sketch the two constant segments of the function.
Question 7

This was the least successfully attempted Pure question with a mean score under 6/20. It was less
than 4% more popular than question 4. The first part of this question was generally well attempted,
with a significant number of candidates being able to correctly verify the algebraic identity utilising a
number of different approaches. There were some very neat solutions, but candidates who
multiplied throughout by the complex conjugate and managed to keep track of the ensuing algebra
were also often successful. Candidates must make sure that when they are trying to show a given
result that they fully justify their solution – in this case some candidates missed out several steps of
working and so did not gain full credit. Many candidates recognised that the form of z meant that
the number was purely imaginary, but only a few candidates gained full credit for this part of the
question with many omitting the modulus signs on the cot term for the modulus, or omitting the
second possible angle. Some candidates were confused by the angles present in the given form of z
and gave the argument as . In part (ii), the approach using the result from part (i) often
did not score full marks due to the fact candidates would divide by quantities without explaining why
they were non-zero. Some attempted this question with vector methods without clearly setting up
that they were treating a,b as vectors rather than complex numbers. They were often unclear as to
whether they were actually considering vectors, or considering complex numbers, which was
particularly apparent in attempts to take the dot product of vectors without including the “dot”
symbol. A number of candidates attempted to work out the gradients of the two line segments and
show they multiplied to give -1: unfortunately, none recognised that a number of special cases were
not taken care of with this method (cases where the lines were horizontal and vertical) and so did
not score highly. Some candidates took a geometrical approach which needed to be fully explained
to be convincing. For part (iii), many were more successful than for (ii): they recognised that part (ii)
could be applied to give the result, and those who did generally gained full, or nearly full, credit.
Vector approaches and considering the gradients of the line segments were used again in this part,
with some candidates repeating the work they had done in the previous part, with the same pitfalls.
Many omitted the case “if b+c=0 then h=a”. Part (iv) was not attempted by a significant proportion.
Of those who did attempt it, a significant number gained full credit. The most common mistake for
this part of the question was candidates giving the transformation as “reflection through a point”,
which did not gain them credit as this is not considered to be a “Single transformation” as requested
(each point is reflected through a different line). Another common mistake was the miscalculation of
the midpoint of AQ as (b+c+d-a)/2 or as (a+b+c+d)/4.
Question 8

Fifth most popular (77%), this was fourth least successful with a mean mark of six and a half. There
were very few perfect attempts and a sizeable number of attempts failed to get any marks.
Induction in both parts (i) and (iii) was generally executed very well, however marks were
frequently lost for logical imprecision. A very common cause of lost marks was a lack of care with
inequalities involving potentially negative numbers. In part (i), almost no candidates noticed that
squaring the inequality required noting the non-negativity of the lower bound. Many candidates
also had trouble with the base case, some because they were mistakenly thinking 40 = 0. In part
(ii), many candidates lost marks when attempting to show that the sequence |𝑥𝑥𝑛𝑛 | remains
bounded in the case |𝑎𝑎| < 2, by not excluding the possibility that 𝑥𝑥2 goes below −2 and hence
diverges to positive infinity. Another common error in part (ii) was failing to make the link to the
inequality in part (i). Many candidates tried to show divergence to infinity by showing that the
sequence was increasing. In part (iii) most candidates worked back from the required result to
find a suitable value for 𝑎𝑎. The inductive calculation was generally performed well, however
plenty of candidates failed to show that their value of 𝑎𝑎 worked and was greater than 2. When
solving equations, it should either be checked that all the steps are reversible (in this case they
were not because of a possible division by zero) or that the claimed solution does in fact work.
Most attempts at the final section on convergence were informal but successful.
Question 9

Just over a fifth attempted this but it had the dubious distinction of being the least successful
question with a mean score a little over 4/20. There were a number of alternative methods used for
the first result, and those that were successful usually applied the sine rule or dropped
perpendiculars. However, some candidates drew a triangle with angles found and wrote down sine
or cosine rules with no indication of how they were to be combined thus earning very little credit.
Candidates who understood the concept of restitution were usually able to complete the second
part of the question without any problems. Many candidates failed on the last part of the question
by trying to give verbose intuition-based arguments instead of finding a third restitution equation.
Question 10

Whilst this was the least popular question, being attempted by a tenth of the candidature, it was the
sixth most successful with a mean over 7/20. Part (i) was successfully attempted by many
candidates, by correctly finding the coordinates of the particle and then using differentiation and
Pythagoras to find the speed as required. In part (ii), many understood that they could use
conservation of energy even though they failed to justify it. Many used the appropriate circular
motion formula in part (iii), but then stumbled as they lacked justification of the evaluation of their
constant of integration, or the choice of sign when taking the square root. Quite a few struggled to
find the link between 𝑏𝑏 − 𝑎𝑎θ and the given answer, and some attempted to jump to the given
answer!
Question 11

Comfortably the most popular applied question on the paper attracting slightly more than a third of
candidates, it was the second most successful on the whole paper with a mean of 11/20. The quality
of attempts for this question was high, with many candidates scoring full or close to full marks.
Almost all candidates attempting it dealt with part (i) successfully. However, in part (ii) candidates
often made incorrect conditioning arguments. The most common errors were computing
𝑃𝑃(𝑍𝑍 < 𝑧𝑧|𝑌𝑌 = 𝑛𝑛) rather than 𝑃𝑃(𝑍𝑍 < 𝑧𝑧) and confusing 𝑃𝑃(𝑍𝑍 < 𝑧𝑧|𝑌𝑌 = 𝑛𝑛) with 𝑃𝑃(𝑍𝑍 < 𝑧𝑧 𝑎𝑎𝑎𝑎𝑎𝑎 𝑌𝑌 = 𝑛𝑛). In
part (iii), most candidates suitably obtained a probability density function for 𝑍𝑍, but there were
several computational mistakes in the integration by parts to evaluate the expectation. The
independence argument in part (iv) was largely well executed, even when candidates had been
unsuccessful in answering parts (ii) and (iii) of the question.
Question 12

A sixth of candidates attempted this, making it the second least popular, and it was the third least
successful with a mean just shy of six and a half. Very few candidates obtained full marks for the
very first part of the question, showing that 𝑋𝑋12 and 𝑋𝑋23 are independent; the most common error
being to check only that 𝑃𝑃 (𝑋𝑋12 = 1, 𝑋𝑋23 = 1) = 𝑃𝑃 (𝑋𝑋12 = 1)𝑃𝑃(𝑋𝑋23 = 1), rather than all four
possible cases for the different values of the two random variables. However, in general, candidates
engaged well with the combinatorial aspect of this part and provided sound methods for counting
pairs of indices in order to obtain the mean and variance, though many did not use the fact that for
independent random variables, 𝑉𝑉𝑉𝑉𝑉𝑉(∑𝑖𝑖 𝑋𝑋𝑖𝑖 ) = ∑𝑖𝑖 𝑉𝑉𝑉𝑉𝑉𝑉(𝑋𝑋𝑖𝑖 ). However, part (ii) was consistently well
executed, with most candidates that attempted it being successful. In part (iii), establishing non-
independence was well executed, and again, as in part (i), candidates provided sound methods for
counting pairs of indices.
STEP MATHEMATICS 3
2021
Mark Scheme
 𝑑𝑑𝑑𝑑
1. (i) 𝑥𝑥 = −4 cos 3 𝑡𝑡 so = 12 cos 2 𝑡𝑡 sin 𝑡𝑡 M1
𝑑𝑑𝑑𝑑

𝑑𝑑𝑑𝑑
𝑦𝑦 = 12 sin 𝑡𝑡 − 4 sin3 𝑡𝑡 so = 12 cos 𝑡𝑡 − 12 sin2 𝑡𝑡 cos 𝑡𝑡 = 12 cos 𝑡𝑡 (1 − sin2 𝑡𝑡) = 12 cos 3 𝑡𝑡
𝑑𝑑𝑑𝑑
M1
𝑑𝑑𝑑𝑑 12 cos3 𝑡𝑡
So = = cot 𝑡𝑡 A1
𝑑𝑑𝑑𝑑 12 cos2 𝑡𝑡 sin 𝑡𝑡

Thus the equation of the normal at (−4 cos 3 𝜑𝜑 , 12 sin 𝜑𝜑 − 4 sin3 𝜑𝜑 ) is

1
𝑦𝑦 − (12 sin 𝜑𝜑 − 4 sin3 𝜑𝜑) = − (𝑥𝑥 − −4 cos 3 𝜑𝜑)
cot 𝜑𝜑
M1 A1ft

This simplifies to 𝑥𝑥 sin 𝜑𝜑 + 𝑦𝑦 cos 𝜑𝜑 = 12 sin 𝜑𝜑 cos 𝜑𝜑 − 4 sin3 𝜑𝜑 cos 𝜑𝜑 − 4 sin 𝜑𝜑 cos 3 𝜑𝜑

That is 𝑥𝑥 sin 𝜑𝜑 + 𝑦𝑦 cos 𝜑𝜑 = 8 sin 𝜑𝜑 cos 𝜑𝜑 A1 (6)

Alternative simplification 𝑥𝑥 tan 𝜑𝜑 + 𝑦𝑦 = 8 sin 𝜑𝜑

𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑
For 𝑥𝑥 = 8 cos 3 𝑡𝑡 , = −24 cos 2 𝑡𝑡 sin 𝑡𝑡 and for 𝑦𝑦 = 8 sin3 𝑡𝑡 , = 24 sin2 𝑡𝑡 cos 𝑡𝑡
𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑

𝑑𝑑𝑑𝑑 24 sin2 𝑡𝑡 cos 𝑡𝑡

So = = − tan 𝑡𝑡 M1 A1ft
𝑑𝑑𝑑𝑑 −24 cos2 𝑡𝑡 sin 𝑡𝑡
2 2
Thus the equation of the tangent to 𝑥𝑥 3 + 𝑦𝑦 3 = 4 at (8 cos 3 𝜑𝜑 , 8 sin3 𝜑𝜑 ) is

𝑦𝑦 − 8 sin3 𝜑𝜑 = − tan 𝜑𝜑 (𝑥𝑥 − 8 cos 3 𝜑𝜑)

M1

This simplifies to

𝑥𝑥 sin 𝜑𝜑 + 𝑦𝑦 cos 𝜑𝜑 = 8 sin3 𝜑𝜑 cos 𝜑𝜑 + 8 sin 𝜑𝜑 cos 3 𝜑𝜑 = 8 sin 𝜑𝜑 cos 𝜑𝜑 (sin2 𝜑𝜑 + cos 2 𝜑𝜑)
That is 𝑥𝑥 sin 𝜑𝜑 + 𝑦𝑦 cos 𝜑𝜑 = 8 sin 𝜑𝜑 cos 𝜑𝜑 as required. A1 (4)

Alternative 1

the normal is a tangent to the second curve if it has the same gradient and the point
(8 cos 3 𝜑𝜑 , 8 sin3 𝜑𝜑 ) lies on the normal. M1

Gradient working as before M1A1ft

Substitution 𝑥𝑥 sin 𝜑𝜑 + 𝑦𝑦 cos 𝜑𝜑 = 8 sin 𝜑𝜑 cos 3 𝜑𝜑 + 8 sin3 𝜑𝜑 cos 𝜑𝜑 = 8 sin 𝜑𝜑 cos 𝜑𝜑 (sin2 𝜑𝜑 +
cos 2 𝜑𝜑) = 8 sin 𝜑𝜑 cos 𝜑𝜑 as required or 𝑥𝑥 tan 𝜑𝜑 + 𝑦𝑦 = 8 sin 𝜑𝜑 cos 𝜑𝜑 (sin2 𝜑𝜑 + cos 2 𝜑𝜑) A1

Alternative 2
2 −1 2 −1 𝑑𝑑𝑑𝑑
𝑥𝑥 3 + 𝑦𝑦 3 =0
3 3 𝑑𝑑𝑑𝑑
M1
 𝑑𝑑𝑑𝑑
(ii) 𝑥𝑥 = cos 𝑡𝑡 + 𝑡𝑡 sin 𝑡𝑡 so = − sin 𝑡𝑡 + 𝑡𝑡 cos 𝑡𝑡 + sin 𝑡𝑡 = 𝑡𝑡 cos 𝑡𝑡
𝑑𝑑𝑑𝑑

𝑑𝑑𝑑𝑑
𝑦𝑦 = sin 𝑡𝑡 − 𝑡𝑡 cos 𝑡𝑡 so = cos 𝑡𝑡 − cos 𝑡𝑡 + 𝑡𝑡 sin 𝑡𝑡 = 𝑡𝑡 sin 𝑡𝑡 M1
𝑑𝑑𝑑𝑑

𝑑𝑑𝑑𝑑
So = tan 𝑡𝑡 A1
𝑑𝑑𝑑𝑑

Thus the equation of the normal at (cos 𝜑𝜑 + 𝜑𝜑 sin 𝜑𝜑 , sin 𝜑𝜑 − 𝜑𝜑 cos 𝜑𝜑 ) is

𝑦𝑦 − (sin 𝜑𝜑 − 𝜑𝜑 cos 𝜑𝜑) = − cot 𝜑𝜑�𝑥𝑥 − (cos 𝜑𝜑 + 𝜑𝜑 sin 𝜑𝜑)�

M1 A1ft

This simplifies to 𝑥𝑥 cos 𝜑𝜑 + 𝑦𝑦 sin 𝜑𝜑 = 1 A1 (5)

Alternatives which can be followed through to perpendicular distance step, or alternative method #
are

𝑥𝑥 + 𝑦𝑦 tan 𝜑𝜑 = sec 𝜑𝜑 and 𝑥𝑥 cot 𝜑𝜑 + 𝑦𝑦 = csc 𝜑𝜑

−1
The distance of (0,0) from 𝑥𝑥 cos 𝜑𝜑 + 𝑦𝑦 sin 𝜑𝜑 = 1 is � �=1
�cos2 𝜑𝜑+sin2 𝜑𝜑

M1 A1ft A1

Alternatively, the perpendicular to 𝑥𝑥 cos 𝜑𝜑 + 𝑦𝑦 sin 𝜑𝜑 = 1 through (0,0) is

𝑦𝑦 cos 𝜑𝜑 − 𝑥𝑥 sin 𝜑𝜑 = 0 , and these two lines meet at (cos 𝜑𝜑 , sin 𝜑𝜑 )

M1 A1ft

which is a distance �cos 2 𝜑𝜑 + sin2 𝜑𝜑 = 1 from (0,0) . A1

So the curve to which this normal is a tangent is a circle centre (0,0) , radius 1 which is thus
𝑥𝑥 2 + 𝑦𝑦 2 = 1 M1 A1 (5)
 1 −𝑥𝑥 𝑥𝑥 𝑎𝑎 𝑎𝑎 − (𝑏𝑏 − 𝑐𝑐 )𝑥𝑥 𝑎𝑎 − 𝑎𝑎 0
2. (i) � 𝑦𝑦 1 −𝑦𝑦� �𝑏𝑏 � = �𝑏𝑏 − (𝑐𝑐 − 𝑎𝑎)𝑦𝑦� = �𝑏𝑏 − 𝑏𝑏 � = �0� as required. M1 A1*
−𝑧𝑧 𝑧𝑧 1 𝑐𝑐 𝑐𝑐 − (𝑎𝑎 − 𝑏𝑏)𝑧𝑧 𝑐𝑐 − 𝑐𝑐 0
𝑎𝑎 0 0
−1
As a, b and c are distinct, they cannot all be zero. If 𝑀𝑀 exists �𝑏𝑏 � = 𝑀𝑀−1 �0� = �0� which is a
𝑐𝑐 0 0
contradiction.
1 −𝑥𝑥 𝑥𝑥
So, 𝑀𝑀−1 does not exist and thus 𝑑𝑑𝑑𝑑𝑑𝑑 � 𝑦𝑦 1 −𝑦𝑦� = 0 , M1
−𝑧𝑧 𝑧𝑧 1
i.e. 1 − 𝑥𝑥𝑥𝑥𝑥𝑥 + 𝑥𝑥𝑥𝑥𝑥𝑥 + 𝑦𝑦𝑦𝑦 + 𝑧𝑧𝑧𝑧 + 𝑥𝑥𝑥𝑥 = 0 , (Sarus)

or 1(1 + 𝑦𝑦𝑦𝑦) − −𝑥𝑥 (𝑦𝑦 − 𝑦𝑦𝑦𝑦) + 𝑥𝑥 (𝑦𝑦𝑦𝑦 + 𝑧𝑧) = 0 (by co-factors) M1

which simplifies to

𝑦𝑦𝑦𝑦 + 𝑧𝑧𝑧𝑧 + 𝑥𝑥𝑥𝑥 = −1 A1 * (5)

(𝑥𝑥 + 𝑦𝑦 + 𝑧𝑧)2 ≥ 0

So 𝑥𝑥 2 + 𝑦𝑦 2 + 𝑧𝑧 2 + 2𝑦𝑦𝑦𝑦 + 2𝑧𝑧𝑧𝑧 + 2𝑥𝑥𝑥𝑥 ≥ 0 M1

and so 𝑥𝑥 2 + 𝑦𝑦 2 + 𝑧𝑧 2 ≥ 2 A1* (2)

2 −𝑥𝑥 −𝑥𝑥 𝑎𝑎 2𝑎𝑎 − (𝑏𝑏 + 𝑐𝑐 )𝑥𝑥 2𝑎𝑎 − 2𝑎𝑎 0

(ii) �−𝑦𝑦 2 −𝑦𝑦� �𝑏𝑏 � = �2𝑏𝑏 − (𝑐𝑐 + 𝑎𝑎)𝑦𝑦� = �2𝑏𝑏 − 2𝑏𝑏 � = �0�
−𝑧𝑧 −𝑧𝑧 2 𝑐𝑐 2𝑐𝑐 − (𝑎𝑎 + 𝑏𝑏)𝑧𝑧 2𝑐𝑐 − 2𝑐𝑐 0

B1 M1 A1
2 −𝑥𝑥 −𝑥𝑥 𝑎𝑎 0
As a, b and c are positive, they cannot all be zero. Thus as �−𝑦𝑦 2 −𝑦𝑦� �𝑏𝑏 � = �0� ,
−𝑧𝑧 −𝑧𝑧 2 𝑐𝑐 0
2 −𝑥𝑥 −𝑥𝑥
as in part (i), 𝑑𝑑𝑑𝑑𝑑𝑑 �−𝑦𝑦 2 −𝑦𝑦� = 0 ,
−𝑧𝑧 −𝑧𝑧 2
i.e. 8 − 𝑥𝑥𝑥𝑥𝑥𝑥 − 𝑥𝑥𝑥𝑥𝑥𝑥 − 2𝑦𝑦𝑦𝑦 − 2𝑧𝑧𝑧𝑧 − 2𝑥𝑥𝑥𝑥 = 0 , that is M1 A1

𝑥𝑥𝑥𝑥𝑥𝑥 + 𝑦𝑦𝑦𝑦 + 𝑧𝑧𝑧𝑧 + 𝑥𝑥𝑥𝑥 = 4 A1* (6)

(𝑥𝑥 + 1)(𝑦𝑦 + 1)(𝑧𝑧 + 1) = 𝑥𝑥𝑥𝑥𝑥𝑥 + 𝑧𝑧𝑧𝑧 + 𝑥𝑥𝑥𝑥 + 𝑥𝑥 + 𝑦𝑦 + 𝑧𝑧 + 1 = 4 + 𝑥𝑥 + 𝑦𝑦 + 𝑧𝑧 + 1 > 5

M1 A1

because as a, b, and c are all positive, so are x, y and z. E1

2𝑎𝑎 2𝑏𝑏 2𝑐𝑐
Thus � + 1� � + 1� � + 1� > 5
𝑏𝑏+𝑐𝑐 𝑐𝑐+𝑎𝑎 𝑎𝑎+𝑏𝑏

Multiplying by (𝑏𝑏 + 𝑐𝑐 )(𝑐𝑐 + 𝑎𝑎)(𝑎𝑎 + 𝑏𝑏) , all three factors of which are positive, gives
(2𝑎𝑎 + 𝑏𝑏 + 𝑐𝑐 )(𝑎𝑎 + 2𝑏𝑏 + 𝑐𝑐 )(𝑎𝑎 + 𝑏𝑏 + 𝑐𝑐 ) > 5(𝑏𝑏 + 𝑐𝑐 )(𝑐𝑐 + 𝑎𝑎)(𝑎𝑎 + 𝑏𝑏) as required. A1* (4)
2𝑎𝑎 2𝑎𝑎 2𝑏𝑏 2𝑐𝑐
𝑥𝑥 = > as a, b, and c are positive, and similarly both, 𝑦𝑦 > and 𝑧𝑧 >
𝑏𝑏+𝑐𝑐 𝑎𝑎+𝑏𝑏+𝑐𝑐 𝑎𝑎+𝑏𝑏+𝑐𝑐 𝑎𝑎+𝑏𝑏+𝑐𝑐

M1
2𝑎𝑎 2𝑏𝑏 2𝑐𝑐 2(𝑎𝑎+𝑏𝑏+𝑐𝑐)
Thus 4 + 𝑥𝑥 + 𝑦𝑦 + 𝑧𝑧 + 1 > 4 + + + +1=4+ +1=7
𝑎𝑎+𝑏𝑏+𝑐𝑐 𝑎𝑎+𝑏𝑏+𝑐𝑐 𝑎𝑎+𝑏𝑏+𝑐𝑐 𝑎𝑎+𝑏𝑏+𝑐𝑐

dM1

and thus following the argument used to obtain the previous result
(2𝑎𝑎 + 𝑏𝑏 + 𝑐𝑐 )(𝑎𝑎 + 2𝑏𝑏 + 𝑐𝑐 )(𝑎𝑎 + 𝑏𝑏 + 𝑐𝑐 ) > 7(𝑏𝑏 + 𝑐𝑐 )(𝑐𝑐 + 𝑎𝑎)(𝑎𝑎 + 𝑏𝑏) as required.

A1* (3)
3. (i)
𝛽𝛽
1 1
(𝐼𝐼𝑛𝑛+1 + 𝐼𝐼𝑛𝑛−1 ) = �(sec 𝑥𝑥 + tan 𝑥𝑥 )𝑛𝑛+1 + (sec 𝑥𝑥 + tan 𝑥𝑥 )𝑛𝑛−1 𝑑𝑑𝑑𝑑
2 2
0

𝛽𝛽
1
= �(sec 𝑥𝑥 + tan 𝑥𝑥 )𝑛𝑛−1 ((sec 𝑥𝑥 + tan 𝑥𝑥 )2 + 1) 𝑑𝑑𝑑𝑑
2
0

M1
𝛽𝛽
1
= �(sec 𝑥𝑥 + tan 𝑥𝑥 )𝑛𝑛−1 (sec 2 𝑥𝑥 + 2 sec 𝑥𝑥 tan 𝑥𝑥 + tan2 𝑥𝑥 + 1) 𝑑𝑑𝑑𝑑
2
0

𝛽𝛽

= �(sec 𝑥𝑥 + tan 𝑥𝑥 )𝑛𝑛−1 (sec 2 𝑥𝑥 + sec 𝑥𝑥 tan 𝑥𝑥 ) 𝑑𝑑𝑑𝑑

0

M1
𝛽𝛽
1 1
= � sec 𝑥𝑥 + tan 𝑥𝑥 � = ((sec 𝛽𝛽 + tan 𝛽𝛽 )𝑛𝑛 − 1)
( ) 𝑛𝑛
𝑛𝑛 0 𝑛𝑛
M1 A1 *A1 (5)

as required.
1 1
(𝐼𝐼𝑛𝑛+1 + 𝐼𝐼𝑛𝑛−1 ) − 𝐼𝐼𝑛𝑛 = (𝐼𝐼𝑛𝑛+1 − 2𝐼𝐼𝑛𝑛 + 𝐼𝐼𝑛𝑛−1 )
2 2
𝛽𝛽
1
= �(sec 𝑥𝑥 + tan 𝑥𝑥 )𝑛𝑛+1 − 2(sec 𝑥𝑥 + tan 𝑥𝑥 )𝑛𝑛 + (sec 𝑥𝑥 + tan 𝑥𝑥 )𝑛𝑛−1 𝑑𝑑𝑑𝑑
2
0

M1
𝛽𝛽
1 2
= �(sec 𝑥𝑥 + tan 𝑥𝑥 )𝑛𝑛−1 �(sec 𝑥𝑥 + tan 𝑥𝑥 ) − 1� 𝑑𝑑𝑑𝑑
2
0

M1 A1
2
�(sec 𝑥𝑥 + tan 𝑥𝑥 ) − 1� > 0 for all x>0
𝜋𝜋
sec 𝑥𝑥 ≥ 1 for 0 ≤ 𝑥𝑥 < and hence for 0 ≤ x < 𝛽𝛽 and similarly tan 𝑥𝑥 ≥ 0 , and thus also
2
(sec 𝑥𝑥 + tan 𝑥𝑥 )𝑛𝑛−1 > 0 . E1

1
Therefore, (𝐼𝐼𝑛𝑛+1 + 𝐼𝐼𝑛𝑛−1 ) − 𝐼𝐼𝑛𝑛 > 0 , A1
2
1 1
and so 𝐼𝐼𝑛𝑛 < (𝐼𝐼𝑛𝑛+1 + 𝐼𝐼𝑛𝑛−1 ) = ((sec 𝛽𝛽 + tan 𝛽𝛽 )𝑛𝑛 − 1) as required. M1 *A1 (7)
2 𝑛𝑛
Alternative 1: it has already been shown that
1 𝛽𝛽
(𝐼𝐼𝑛𝑛+1 + 𝐼𝐼𝑛𝑛−1 ) = ∫0 (sec 𝑥𝑥 + tan 𝑥𝑥 )𝑛𝑛−1 (sec 2 𝑥𝑥 + sec 𝑥𝑥 tan 𝑥𝑥 ) 𝑑𝑑𝑑𝑑
2

𝛽𝛽

= � sec 𝑥𝑥 (sec 𝑥𝑥 + tan 𝑥𝑥 )𝑛𝑛 𝑑𝑑𝑑𝑑

0

which is greater than 𝐼𝐼𝑛𝑛 as the expression being integrated is greater than (sec 𝑥𝑥 + tan 𝑥𝑥 )𝑛𝑛 because
sec 𝑥𝑥 > 0 over this domain.

Alternative 2:-
𝛽𝛽

𝐼𝐼𝑛𝑛+1 − 𝐼𝐼𝑛𝑛 = �(sec 𝑥𝑥 + tan 𝑥𝑥 )𝑛𝑛 (sec 𝑥𝑥 + tan 𝑥𝑥 − 1) 𝑑𝑑𝑑𝑑

0

𝛽𝛽

𝐼𝐼𝑛𝑛 − 𝐼𝐼𝑛𝑛−1 = �(sec 𝑥𝑥 + tan 𝑥𝑥 )𝑛𝑛−1 (sec 𝑥𝑥 + tan 𝑥𝑥 − 1) 𝑑𝑑𝑑𝑑

0

M1 A1 A1

For 0 < x < 𝛽𝛽 , sec 𝑥𝑥 > 1 , tan 𝑥𝑥 > 0 so sec 𝑥𝑥 + tan 𝑥𝑥 > 1 E1 and thus 𝐼𝐼𝑛𝑛+1 − 𝐼𝐼𝑛𝑛 > 𝐼𝐼𝑛𝑛 − 𝐼𝐼𝑛𝑛−1 A1
1 1
and so 𝐼𝐼𝑛𝑛 ≤ (𝐼𝐼𝑛𝑛+1 + 𝐼𝐼𝑛𝑛−1 ) = ((sec 𝛽𝛽 + tan 𝛽𝛽 )𝑛𝑛 − 1) M1 *A1 (7)
2 𝑛𝑛

1 1 𝛽𝛽
(ii)
2
(𝐽𝐽𝑛𝑛+1 + 𝐽𝐽𝑛𝑛−1 ) =
2
∫0 (sec 𝑥𝑥 cos 𝛽𝛽 + tan 𝑥𝑥 )𝑛𝑛+1 + (sec 𝑥𝑥 cos 𝛽𝛽 + tan 𝑥𝑥 )𝑛𝑛−1 𝑑𝑑𝑑𝑑
𝛽𝛽
1
= �(sec 𝑥𝑥 cos 𝛽𝛽 + tan 𝑥𝑥 )𝑛𝑛−1 ((sec 𝑥𝑥 cos 𝛽𝛽 + tan 𝑥𝑥 )2 + 1) 𝑑𝑑𝑑𝑑
2
0

M1
𝛽𝛽
1
= �(sec 𝑥𝑥 cos 𝛽𝛽 + tan 𝑥𝑥 )𝑛𝑛−1 (sec 2 𝑥𝑥 cos 2 𝛽𝛽 + 2 sec 𝑥𝑥 cos 𝛽𝛽 tan 𝑥𝑥 + tan2 𝑥𝑥 + 1) 𝑑𝑑𝑑𝑑
2
0

𝛽𝛽
1
= �(sec 𝑥𝑥 cos 𝛽𝛽 + tan 𝑥𝑥 )𝑛𝑛−1 (sec 2 𝑥𝑥 (1 − sin2 𝛽𝛽 ) + 2 sec 𝑥𝑥 cos 𝛽𝛽 tan 𝑥𝑥 + tan2 𝑥𝑥 + 1) 𝑑𝑑𝑑𝑑
2
0

𝛽𝛽

= �(sec 𝑥𝑥 cos 𝛽𝛽 + tan 𝑥𝑥 )𝑛𝑛−1 �(sec 2 𝑥𝑥 + sec 𝑥𝑥 cos 𝛽𝛽 tan 𝑥𝑥 ) − sec 2 𝑥𝑥 sin2 𝛽𝛽�𝑑𝑑𝑑𝑑
0

M1
𝛽𝛽
𝛽𝛽
1
�(sec 𝑥𝑥 cos 𝛽𝛽 + tan 𝑥𝑥 )𝑛𝑛−1 (sec 2 𝑥𝑥 + sec 𝑥𝑥 cos 𝛽𝛽 tan 𝑥𝑥 )𝑑𝑑𝑑𝑑 = � (sec 𝑥𝑥 cos 𝛽𝛽 + tan 𝑥𝑥 )𝑛𝑛 �
𝑛𝑛 0
0
 M1
1
= ((1 + tan 𝛽𝛽 )𝑛𝑛 − cos 𝑛𝑛 𝛽𝛽 )
𝑛𝑛
A1
𝛽𝛽

�(sec 𝑥𝑥 cos 𝛽𝛽 + tan 𝑥𝑥 )𝑛𝑛−1 sec 2 𝑥𝑥 sin2 𝛽𝛽 𝑑𝑑𝑑𝑑 > 0

0

by a similar argument to part (i), namely sec 2 𝑥𝑥 sin2 𝛽𝛽 > 0 for any x, and sec 𝑥𝑥 cos 𝛽𝛽 + tan 𝑥𝑥 > 0
𝜋𝜋
as sec 𝑥𝑥 > 0 and tan 𝑥𝑥 ≥ 0 for 0 ≤ x < 𝛽𝛽 < E1
2
1 1
Hence (𝐽𝐽𝑛𝑛+1 + 𝐽𝐽𝑛𝑛−1 ) < ((1 + tan 𝛽𝛽 )𝑛𝑛 − cos 𝑛𝑛 𝛽𝛽 ) A1
2 𝑛𝑛

But
𝛽𝛽
1 1 2
(𝐽𝐽𝑛𝑛+1 + 𝐽𝐽𝑛𝑛−1 ) − 𝐽𝐽𝑛𝑛 == �(sec 𝑥𝑥 cos 𝛽𝛽 + tan 𝑥𝑥 )𝑛𝑛−1 �(sec 𝑥𝑥 cos 𝛽𝛽 + tan 𝑥𝑥 ) − 1� 𝑑𝑑𝑑𝑑 > 0
2 2
0

M1
1 1
as before, and thus 𝐽𝐽𝑛𝑛 < (𝐽𝐽𝑛𝑛+1 + 𝐽𝐽𝑛𝑛−1 ) < ((1 + tan 𝛽𝛽 )𝑛𝑛 − cos 𝑛𝑛 𝛽𝛽 ) as required. *A1 (8)
2 𝑛𝑛
4. (i)
1 1
𝒎𝒎. 𝒂𝒂 = (𝒂𝒂 + 𝒃𝒃). 𝒂𝒂 = (1 + 𝒂𝒂. 𝒃𝒃) = 𝑚𝑚 cos 𝛼𝛼 where 𝛼𝛼 is the non-reflex angle between a and m
2 2
1 1
𝒎𝒎. 𝒃𝒃 = (𝒂𝒂 + 𝒃𝒃). 𝒃𝒃 = (1 + 𝒂𝒂. 𝒃𝒃) = 𝑚𝑚 cos 𝛽𝛽 where 𝛼𝛼 is the non-reflex angle between 𝒃𝒃 and m
2 2

M1 A1

Thus cos 𝛼𝛼 = cos 𝛽𝛽 and so 𝛼𝛼 = 𝛽𝛽 as for 0 ≤ 𝜏𝜏 ≤ 𝜋𝜋 , there is only one value of 𝜏𝜏 for any given
value of cos 𝜏𝜏 . E1 (3)

(ii) 𝒂𝒂𝟏𝟏 . 𝒄𝒄 = (𝒂𝒂 − (𝒂𝒂. 𝒄𝒄)𝒄𝒄). 𝒄𝒄 = 𝒂𝒂. 𝒄𝒄 − 𝒂𝒂. 𝒄𝒄 𝒄𝒄. 𝒄𝒄 = 0 as required. *B1

𝒂𝒂. 𝒄𝒄 = cos 𝛼𝛼 , 𝒃𝒃. 𝒄𝒄 = cos 𝛽𝛽 , 𝒂𝒂. 𝒃𝒃 = cos 𝜃𝜃

𝒂𝒂𝟏𝟏 = 𝒂𝒂 − (𝒂𝒂. 𝒄𝒄)𝒄𝒄 and 𝒃𝒃𝟏𝟏 = 𝒃𝒃 − (𝒃𝒃. 𝒄𝒄)𝒄𝒄

|𝒂𝒂𝟏𝟏 |𝟐𝟐 = 𝒂𝒂𝟏𝟏 . 𝒂𝒂𝟏𝟏 = (𝒂𝒂 − (𝒂𝒂. 𝒄𝒄)𝒄𝒄). (𝒂𝒂 − (𝒂𝒂. 𝒄𝒄)𝒄𝒄) = 𝒂𝒂. 𝒂𝒂 − 2𝒂𝒂. 𝒄𝒄 𝒂𝒂. 𝒄𝒄 + 𝒂𝒂. 𝒄𝒄 𝒂𝒂. 𝒄𝒄 𝒄𝒄. 𝒄𝒄

= 1 − 2 cos2 𝛼𝛼 + cos 2 𝛼𝛼 = sin2 𝛼𝛼

M1

and so. as 𝛼𝛼 is acute, |𝒂𝒂𝟏𝟏 | = sin 𝛼𝛼 as required. *A1

𝒂𝒂𝟏𝟏 . 𝒃𝒃𝟏𝟏 = (𝒂𝒂 − (𝒂𝒂. 𝒄𝒄)𝒄𝒄). (𝒃𝒃 − (𝒃𝒃. 𝒄𝒄)𝒄𝒄) = 𝒂𝒂. 𝒃𝒃 − 2(𝒂𝒂. 𝒄𝒄)(𝒃𝒃. 𝒄𝒄) + (𝒂𝒂. 𝒄𝒄)(𝒃𝒃. 𝒄𝒄)(𝒄𝒄. 𝒄𝒄)
= cos 𝜃𝜃 − cos 𝛼𝛼 cos 𝛽𝛽

M1 A1

but also, 𝒂𝒂𝟏𝟏 . 𝒃𝒃𝟏𝟏 = sin 𝛼𝛼 sin 𝛽𝛽 cos 𝜑𝜑 B1 M1

and hence,
cos 𝜃𝜃 − cos 𝛼𝛼 cos 𝛽𝛽
cos 𝜑𝜑 =
sin 𝛼𝛼 sin 𝛽𝛽
as required. *A1 (8)
1 1 1
(iii) 𝒎𝒎𝟏𝟏 = 𝒎𝒎 − (𝒎𝒎. 𝒄𝒄)𝒄𝒄 = (𝒂𝒂 + 𝒃𝒃) − � (𝒂𝒂 + 𝒃𝒃). 𝒄𝒄� 𝒄𝒄 = (𝒂𝒂𝟏𝟏 + 𝒃𝒃𝟏𝟏 ) B1
2 2 2

𝒎𝒎𝟏𝟏 bisects the angle between 𝒂𝒂𝟏𝟏 and 𝒃𝒃𝟏𝟏 if and only if
𝒎𝒎𝟏𝟏 . 𝒂𝒂𝟏𝟏 𝒎𝒎𝟏𝟏 . 𝒃𝒃𝟏𝟏
=
sin 𝛼𝛼 sin 𝛽𝛽
M1

Thus, multiplying through by 2 sin 𝛼𝛼 sin 𝛽𝛽 ,

(𝒂𝒂𝟏𝟏 + 𝒃𝒃𝟏𝟏 ). 𝒂𝒂𝟏𝟏 sin 𝛽𝛽 = (𝒂𝒂𝟏𝟏 + 𝒃𝒃𝟏𝟏 ). 𝒃𝒃𝟏𝟏 sin 𝛼𝛼

A1

(sin2 𝛼𝛼 + 𝒂𝒂𝟏𝟏 . 𝒃𝒃𝟏𝟏 ) sin 𝛽𝛽 = (sin2 𝛽𝛽 + 𝒂𝒂𝟏𝟏 . 𝒃𝒃𝟏𝟏 ) sin 𝛼𝛼

 M1 A1

So
(𝒂𝒂𝟏𝟏 . 𝒃𝒃𝟏𝟏 − sin 𝛼𝛼 sin 𝛽𝛽 )(sin 𝛼𝛼 − sin 𝛽𝛽 ) = 0

A1

and thus, sin 𝛼𝛼 = sin 𝛽𝛽 in which case 𝛼𝛼 = 𝛽𝛽 as both angles are acute, *A1

or cos 𝜃𝜃 − cos 𝛼𝛼 cos 𝛽𝛽 = sin 𝛼𝛼 sin 𝛽𝛽 , meaning that cos 𝜃𝜃 = cos 𝛼𝛼 cos 𝛽𝛽 + sin 𝛼𝛼 sin 𝛽𝛽 = cos(𝛼𝛼 − 𝛽𝛽 )

M1 *A1 (9)
5. (i) The curves meet when 𝑎𝑎 + 2 cos 𝜃𝜃 = 2 + cos 2𝜃𝜃

That is, 𝑎𝑎 + 2 cos 𝜃𝜃 = 2 + 2 cos 2 𝜃𝜃 − 1 or as required, B1 2 cos 2 𝜃𝜃 − 2 cos 𝜃𝜃 + 1 − 𝑎𝑎 = 0

1
The curves touch if this quadratic has coincident roots, M1 i.e. if 4 − 8(1 − 𝑎𝑎) = 0 ⇒ 𝑎𝑎 = , *A1
2
or if cos 𝜃𝜃 = ±1 , M1 in which cases 𝑎𝑎 = 1 A1 or 𝑎𝑎 = 5 . A1 (6)

Alternatively, for the curves to touch, they must have the same gradient, so differentiating,

−2 sin 𝜃𝜃 = −2 sin 2𝜃𝜃 = −4 sin 𝜃𝜃 cos 𝜃𝜃

M1

in which case, either sin 𝜃𝜃 = 0 giving cos 𝜃𝜃 = ±1 , M1 in which cases 𝑎𝑎 = 1 A1 or 𝑎𝑎 = 5 , A1 or

1 1
cos 𝜃𝜃 = in which case 𝑎𝑎 = . *A1 (6)
2 2

1 1 𝜋𝜋 3 𝜋𝜋
(ii) If 𝑎𝑎 = then at points where they touch, cos 𝜃𝜃 = so 𝜃𝜃 = ± and thus � , ± �. M1A1
2 2 3 2 3

5
𝑟𝑟 = 𝑎𝑎 + 2 cos 𝜃𝜃 is symmetrical about the initial line which it intercepts at � , 0 � and has a cusp at
2
1 1 𝜋𝜋
�0 , ± cos −1 �− � � . It passes through � , ± � and only exists for
4 2 2

1 1
− cos −1 �− � < 𝜃𝜃 < cos −1 �− � .
4 4

𝑟𝑟 = 2 + cos 2𝜃𝜃 is symmetrical about both the initial line, and its perpendicular. It passes through
𝜋𝜋
(3,0) , (3, 𝜋𝜋) , and �1 , ± �
2

Sketch G6 (8)

(iii) If 𝑎𝑎 = 1 , then the curves meet where 2 cos 2 𝜃𝜃 − 2 cos 𝜃𝜃 = 0 , i.e. cos 𝜃𝜃 = 1 at (3,0) where
𝜋𝜋
they touch, and cos 𝜃𝜃 = 0 at �1, ± �
2

𝑟𝑟 = 𝑎𝑎 + 2 cos 𝜃𝜃 is symmetrical about the initial line which it intercepts at (3 , 0 ) and has a cusp at
1 2𝜋𝜋 𝜋𝜋
�0 , ± cos −1 �− � � = �0 , ± � . It passes through �1 , ± � and only exists for
2 3 2

2𝜋𝜋 2𝜋𝜋
− < 𝜃𝜃 < .
3 3

Sketch G3

If 𝑎𝑎 = 5 , then the curves meet where 2 cos 2 𝜃𝜃 − 2 cos 𝜃𝜃 − 4 = 0 , i.e. only cos 𝜃𝜃 = −1 at (3, 𝜋𝜋)
where they touch, as cos 𝜃𝜃 ≠ 2 .

𝑟𝑟 = 𝑎𝑎 + 2 cos 𝜃𝜃 is symmetrical about the initial line which it intercepts at (7 , 0 ) and (3 , 𝜋𝜋 ) . It

𝜋𝜋
also passes through �5 , ± � .
2

Sketch G3 (6)
6. (i)
𝑥𝑥 tan 𝛼𝛼 + 1
𝑓𝑓𝛼𝛼 (𝑥𝑥 ) = tan−1 � �
tan 𝛼𝛼 − 𝑥𝑥
1 (tan 𝛼𝛼 − 𝑥𝑥 ) tan 𝛼𝛼 + (𝑥𝑥 tan 𝛼𝛼 + 1)
𝑓𝑓′𝛼𝛼 (𝑥𝑥 ) = 2
𝑥𝑥 tan 𝛼𝛼 + 1 (tan 𝛼𝛼 − 𝑥𝑥 )2
1+� �
tan 𝛼𝛼 − 𝑥𝑥
M1 A1

tan2 𝛼𝛼 + 1
=
(tan 𝛼𝛼 − 𝑥𝑥 )2 + (𝑥𝑥 tan 𝛼𝛼 + 1)2

sec 2 𝛼𝛼 sec 2 𝛼𝛼 1
= = =
tan 𝛼𝛼 + 𝑥𝑥 + 𝑥𝑥 tan 𝛼𝛼 + 1 sec 𝛼𝛼 (1 + 𝑥𝑥 ) 1 + 𝑥𝑥 2
2 2 2 2 2 2

M1 M1 *A1 (5)

as required.

Alternative
𝑥𝑥 tan 𝛼𝛼 + 1
𝑓𝑓𝛼𝛼 (𝑥𝑥 ) = tan−1 � �
tan 𝛼𝛼 − 𝑥𝑥
𝑥𝑥 + cot 𝛼𝛼
= tan−1 � �
1 − 𝑥𝑥 cot 𝛼𝛼
𝜋𝜋
tan(tan−1 𝑥𝑥 ) + tan � − 𝛼𝛼�
= tan −1 � 2 �
𝜋𝜋
1 − tan(tan 𝑥𝑥 ) tan � − 𝛼𝛼�
−1
2
M1 A1
𝜋𝜋
= tan−1 �tan �tan−1 𝑥𝑥 + − 𝛼𝛼��
2
M1
𝜋𝜋 𝜋𝜋
= tan−1 𝑥𝑥 + − 𝛼𝛼 if this is less than , i.e. if 𝑥𝑥 < tan 𝛼𝛼
2 2
𝜋𝜋
or = tan−1 𝑥𝑥 − − 𝛼𝛼 if 𝑥𝑥 > tan 𝛼𝛼 M1
2
𝑑𝑑 1
So 𝑓𝑓′𝛼𝛼 (𝑥𝑥 ) = (tan−1 𝑥𝑥 ) = *A1 (5)
𝑑𝑑𝑑𝑑 1+𝑥𝑥 2

Thus 𝑓𝑓𝛼𝛼 (𝑥𝑥 ) = tan−1 𝑥𝑥 + 𝑐𝑐

1 𝜋𝜋
𝑓𝑓𝛼𝛼 (0) = tan−1 � � = tan−1 (cot 𝛼𝛼 ) = − 𝛼𝛼
tan 𝛼𝛼 2
𝑓𝑓𝛼𝛼 (𝑥𝑥 ) = 0 when 𝑥𝑥 = − cot 𝛼𝛼
𝜋𝜋 𝜋𝜋
There is a discontinuity at 𝑥𝑥 = tan 𝛼𝛼 , with 𝑓𝑓𝛼𝛼 (𝑥𝑥 ) approaching from below and − from above.
2 2

As 𝑥𝑥 → ±∞ , 𝑓𝑓𝛼𝛼 (𝑥𝑥 ) → tan−1 (− tan 𝛼𝛼 ) = −𝛼𝛼

 𝜋𝜋 𝜋𝜋
So 𝑓𝑓𝛼𝛼 (𝑥𝑥 ) = tan−1 𝑥𝑥 + − 𝛼𝛼 for 𝑥𝑥 < tan 𝛼𝛼 and 𝑓𝑓𝛼𝛼 (𝑥𝑥 ) = tan−1 𝑥𝑥 − − 𝛼𝛼 for 𝑥𝑥 > tan 𝛼𝛼
2 2

Sketch G1 G1 G1 (3)

𝑦𝑦 = 𝑓𝑓𝛼𝛼 (𝑥𝑥 ) − 𝑓𝑓𝛽𝛽 (𝑥𝑥 ) =

𝜋𝜋 𝜋𝜋
� − 𝛼𝛼� − � − 𝛽𝛽� = β − α for 𝑥𝑥 < tan 𝛼𝛼
2 2

𝜋𝜋 𝜋𝜋
�− − 𝛼𝛼� − � − 𝛽𝛽� = β − α − π for tan 𝛼𝛼 < 𝑥𝑥 < tan 𝛽𝛽
2 2

𝜋𝜋 𝜋𝜋
and �− − 𝛼𝛼� − �− − 𝛽𝛽� = β − α for 𝑥𝑥 > tan 𝛽𝛽
2 2

Sketch G1 G1 G1 (3)

(ii) 𝑔𝑔(𝑥𝑥 ) = tanh−1 (sin 𝑥𝑥 ) − sinh−1 (tan 𝑥𝑥 )

1 1
𝑔𝑔′ (𝑥𝑥 ) = 2
cos 𝑥𝑥 − sec 2 𝑥𝑥
1 − sin 𝑥𝑥 √1 + tan2 𝑥𝑥
M1 A1 A1

cos 𝑥𝑥 sec 2 𝑥𝑥 sec 2 𝑥𝑥

= − = sec 𝑥𝑥 − = 2 sec 𝑥𝑥
cos 2 𝑥𝑥 |sec 𝑥𝑥| − sec 𝑥𝑥
M1 *A1 (5)
𝜋𝜋 3𝜋𝜋
as required, for sec 𝑥𝑥 < 0 , i.e. for < 𝑥𝑥 < .
2 2

(For sec 𝑥𝑥 > 0 , 𝑔𝑔′ (𝑥𝑥 ) = 0 )

Sketch G1 G1 G1 G1 (4)
7.
𝑒𝑒 𝑖𝑖𝑖𝑖 + 𝑒𝑒 𝑖𝑖𝑖𝑖
𝑧𝑧 =
𝑒𝑒 𝑖𝑖𝑖𝑖 − 𝑒𝑒 𝑖𝑖𝑖𝑖
cos 𝜃𝜃 + 𝑖𝑖 sin 𝜃𝜃 + cos 𝜑𝜑 + 𝑖𝑖 sin 𝜑𝜑
=
cos 𝜃𝜃 + 𝑖𝑖 sin 𝜃𝜃 − cos 𝜑𝜑 − 𝑖𝑖 sin 𝜑𝜑
M1
𝜃𝜃 + 𝜑𝜑 𝜃𝜃 − 𝜑𝜑 𝜃𝜃 + 𝜑𝜑 𝜃𝜃 − 𝜑𝜑
2 cos cos + 2𝑖𝑖 sin cos
= 2 2 2 2
𝜃𝜃 + 𝜑𝜑 𝜃𝜃 − 𝜑𝜑 𝜃𝜃 + 𝜑𝜑 𝜃𝜃 − 𝜑𝜑
−2 sin sin + 2𝑖𝑖 cos sin
2 2 2 2
M1 A1 A1
𝜃𝜃 − 𝜑𝜑 𝜃𝜃 + 𝜑𝜑 𝜃𝜃 + 𝜑𝜑
2 cos �cos + 𝑖𝑖 sin �
= 2 2 2
𝜃𝜃 − 𝜑𝜑 𝜃𝜃 + 𝜑𝜑 𝜃𝜃 + 𝜑𝜑
2 sin �𝑖𝑖 cos − sin �
2 2 2
𝜃𝜃 − 𝜑𝜑
= −𝑖𝑖 cot
2
𝜑𝜑 − 𝜃𝜃
= 𝑖𝑖 cot
2
*A1 (5)

as required.

Alternatively,
𝜃𝜃−𝜑𝜑 𝜃𝜃−𝜑𝜑 𝜃𝜃 − 𝜑𝜑
𝑖𝑖� � −𝑖𝑖� �
𝑒𝑒 𝑖𝑖𝑖𝑖 + 𝑒𝑒 𝑖𝑖𝑖𝑖 𝑒𝑒 2 + 𝑒𝑒 2 2 cos
𝑧𝑧 = 𝑖𝑖𝑖𝑖 = 𝜃𝜃−𝜑𝜑 = 2 = −𝑖𝑖 cot 𝜃𝜃 − 𝜑𝜑 = 𝑖𝑖 cot 𝜑𝜑 − 𝜃𝜃
𝑒𝑒 − 𝑒𝑒 𝑖𝑖𝑖𝑖
𝑖𝑖� � −𝑖𝑖�
𝜃𝜃−𝜑𝜑
� 𝜃𝜃 − 𝜑𝜑 2 2
𝑒𝑒 2 − 𝑒𝑒 2 2𝑖𝑖 sin
2
M1 M1 A1 A1 *A1 (5)
𝜃𝜃 − 𝜑𝜑
|𝑧𝑧| = �cot �
2
M1 A1
𝜋𝜋
|arg 𝑧𝑧| =
2
𝜋𝜋 3𝜋𝜋
[or arg 𝑧𝑧 = 𝑜𝑜𝑜𝑜 ]
2 2

M1 A1 (4)

(ii) Let 𝑎𝑎 = 𝑒𝑒 𝑖𝑖𝑖𝑖 and 𝑏𝑏 = 𝑒𝑒 𝑖𝑖𝑖𝑖 M1 then 𝑥𝑥 = 𝑎𝑎 + 𝑏𝑏 = 𝑒𝑒 𝑖𝑖𝑖𝑖 + 𝑒𝑒 𝑖𝑖𝑖𝑖 and 𝐴𝐴𝐴𝐴 = 𝑏𝑏 − 𝑎𝑎 = 𝑒𝑒 𝑖𝑖𝑖𝑖 − 𝑒𝑒 𝑖𝑖𝑖𝑖

𝑥𝑥 𝑒𝑒 𝑖𝑖𝑖𝑖 + 𝑒𝑒 𝑖𝑖𝑖𝑖
arg 𝑥𝑥 − arg 𝐴𝐴𝐴𝐴 = arg = arg 𝑖𝑖𝑖𝑖
𝐴𝐴𝐴𝐴 𝑒𝑒 − 𝑒𝑒 𝑖𝑖𝑖𝑖
𝜋𝜋
so using (i), |arg 𝑥𝑥 − arg 𝐴𝐴𝐴𝐴| = A1 and thus OX and AB are perpendicular, since 𝑥𝑥 = 𝑎𝑎 + 𝑏𝑏 ≠ 0
2
and 𝑎𝑎 ≠ 𝑏𝑏 as A and B are distinct. E1 (3)
Alternative:- 0, 𝑎𝑎, 𝑎𝑎 + 𝑏𝑏, 𝑏𝑏 define a rhombus OAXB as |𝑎𝑎| = |𝑏𝑏| = 1. Diagonals of a rhombus are
perpendicular (and bisect one another).

(iii) ℎ = 𝑎𝑎 + 𝑏𝑏 + 𝑐𝑐 so 𝐴𝐴𝐴𝐴 = 𝑎𝑎 + 𝑏𝑏 + 𝑐𝑐 − 𝑎𝑎 = 𝑏𝑏 + 𝑐𝑐 and 𝐵𝐵𝐵𝐵 = 𝑐𝑐 − 𝑏𝑏 and thus

𝐴𝐴𝐴𝐴 𝑏𝑏 + 𝑐𝑐
=
𝐵𝐵𝐵𝐵 𝑐𝑐 − 𝑏𝑏
B1

as 𝑐𝑐 − 𝑏𝑏 ≠ 0

From (ii),
𝐴𝐴𝐴𝐴 𝜋𝜋
�arg �=
𝐵𝐵𝐵𝐵 2
so BC is perpendicular to AH E1

unless 𝑏𝑏 + 𝑐𝑐 = 0 E1 in which case ℎ = 𝑎𝑎 E1 (4)

(iv) 𝑝𝑝 = 𝑎𝑎 + 𝑏𝑏 + 𝑐𝑐 𝑞𝑞 = 𝑏𝑏 + 𝑐𝑐 + 𝑑𝑑 𝑟𝑟 = 𝑐𝑐 + 𝑑𝑑 + 𝑎𝑎 𝑠𝑠 = 𝑑𝑑 + 𝑎𝑎 + 𝑏𝑏
𝑎𝑎+𝑞𝑞 𝑎𝑎+𝑏𝑏+𝑐𝑐+𝑑𝑑
The midpoint of AQ is = and so by its symmetry it is also the midpoint of BR, CS, and
2 2
DP, B1 E1

and thus ABCD is transformed to PQRS by a rotation of 𝜋𝜋 radians about midpoint of AQ. E1 B1 (4)

Alternatively, ABCD is transformed to PQRS by an enlargement scale factor -1 , centre of

enlargement midpoint of AQ.
8. (i) Suppose 𝑥𝑥𝑘𝑘 ≥ 2 + 4𝑘𝑘−1 (𝑎𝑎 − 2) for some particular integer k (and this is positive as 𝑎𝑎 > 2 )

E1

Then 𝑥𝑥𝑘𝑘+1 = 𝑥𝑥𝑘𝑘 2 − 2 ≥ [2 + 4𝑘𝑘−1 (𝑎𝑎 − 2)]2 − 2 = 4 + 4𝑘𝑘 (𝑎𝑎 − 2) + 42𝑘𝑘−2 (𝑎𝑎 − 2)2 − 2

= 2 + 4𝑘𝑘 (𝑎𝑎 − 2) + 42𝑘𝑘−2 (𝑎𝑎 − 2)2

> 2 + 4𝑘𝑘 (𝑎𝑎 − 2)

M1 A1

which is the required result for 𝑘𝑘 + 1 .

For 𝑛𝑛 = 1 , 2 + 4𝑛𝑛−1 (𝑎𝑎 − 2) = 2 + 𝑎𝑎 − 2 = 𝑎𝑎 so in this case, 𝑥𝑥𝑛𝑛 = 2 + 4𝑛𝑛−1 (𝑎𝑎 − 2) B1 and thus
by induction 𝑥𝑥𝑛𝑛 ≥ 2 + 4𝑛𝑛−1 (𝑎𝑎 − 2) for positive integer n. E1 (5)

(ii) If |𝑥𝑥𝑘𝑘 | ≤ 2 , then 0 ≤ |𝑥𝑥𝑘𝑘 |2 ≤ 4 , so −2 ≤ |𝑥𝑥𝑘𝑘 |2 − 2 ≤ 2 , that is −2 ≤ 𝑥𝑥𝑘𝑘+1 ≤ 2 . M1A1

If |𝑎𝑎| ≤ 2 , |𝑥𝑥1 | ≤ 2 and thus by induction −2 ≤ 𝑥𝑥𝑛𝑛 ≤ 2 , that is 𝑥𝑥𝑛𝑛 ↛ ∞ E1

Whether 𝑎𝑎 = ±𝛼𝛼 , 𝑥𝑥2 would equal the same value, namely 𝛼𝛼 2 − 2 . E1

So to consider |𝑎𝑎| ≥ 2 , we only need consider 𝑎𝑎 > 2 to discuss the behaviour of all terms after the
first. Therefore, from part (i), we know 𝑥𝑥𝑛𝑛 ≥ 2 + 4𝑛𝑛−1 (|𝑎𝑎| − 2) for 𝑛𝑛 ≥ 2 , and thus 𝑥𝑥𝑛𝑛 → ∞ as
𝑛𝑛 → ∞ ; B1 hence we have shown 𝑥𝑥𝑛𝑛 → ∞ as 𝑛𝑛 → ∞ if and only if |𝑎𝑎| ≥ 2 . (5)

(iii)
𝐴𝐴𝑥𝑥1 𝑥𝑥2 ⋯ 𝑥𝑥𝑘𝑘
𝑦𝑦𝑘𝑘 =
𝑥𝑥𝑘𝑘+1
𝐴𝐴𝑥𝑥1 𝑥𝑥2 ⋯ 𝑥𝑥𝑘𝑘+1 𝑥𝑥𝑘𝑘+1 2
𝑦𝑦𝑘𝑘+1 = = 𝑦𝑦
𝑥𝑥𝑘𝑘+2 𝑥𝑥𝑘𝑘+2 𝑘𝑘
M1

Suppose that
�𝑥𝑥𝑘𝑘+1 2 − 4
𝑦𝑦𝑘𝑘 =
𝑥𝑥𝑘𝑘+1
for some positive integer k, E1 then

𝑥𝑥𝑘𝑘+1 2 �𝑥𝑥𝑘𝑘+1 2 − 4 𝑥𝑥𝑘𝑘+1 �𝑥𝑥𝑘𝑘+1 2 − 4

𝑦𝑦𝑘𝑘+1 = =
𝑥𝑥𝑘𝑘+2 𝑥𝑥𝑘𝑘+1 𝑥𝑥𝑘𝑘+2

As 𝑥𝑥𝑘𝑘+2 = 𝑥𝑥𝑘𝑘+1 2 − 2 , 𝑥𝑥𝑘𝑘+1 = �𝑥𝑥𝑘𝑘+2 + 2 , and �𝑥𝑥𝑘𝑘+1 2 − 4 = �𝑥𝑥𝑘𝑘+2 − 2 ,

and thus,
�𝑥𝑥𝑘𝑘+2 + 2�𝑥𝑥𝑘𝑘+2 − 2 �𝑥𝑥𝑘𝑘+2 2 − 4
𝑦𝑦𝑘𝑘+1 = =
𝑥𝑥𝑘𝑘+2 𝑥𝑥𝑘𝑘+2
M1 A1

which is the required result for 𝑘𝑘 + 1 .

 𝐴𝐴𝑥𝑥1
𝑦𝑦1 =
𝑥𝑥2
and also we wish to have
�𝑥𝑥2 2 − 4
𝑦𝑦1 =
𝑥𝑥2
M1

then 𝐴𝐴𝑥𝑥1 = �𝑥𝑥2 2 − 4 , that is 𝐴𝐴2 𝑥𝑥1 2 = 𝑥𝑥2 2 − 4 , and as 𝑥𝑥1 = 𝑎𝑎 , 𝑥𝑥2 = 𝑥𝑥1 2 − 2 = 𝑎𝑎2 − 2

so
𝐴𝐴2 𝑎𝑎2 = (𝑎𝑎2 − 2)2 − 4 = 𝑎𝑎4 − 4𝑎𝑎2 , 𝐴𝐴2 = 𝑎𝑎2 − 4 , and thus 𝑎𝑎 = √𝐴𝐴2 + 4 , as 𝑎𝑎 ≠ 0 nor
−√𝐴𝐴2 + 4 because 𝑎𝑎 > 2. A1 E1

So as the result is true for 𝑦𝑦1 , and we have shown it to be true for 𝑦𝑦𝑘𝑘+1 if it is true for 𝑦𝑦𝑘𝑘 , it is true
by induction for all positive integer 𝑛𝑛 that
�𝑥𝑥𝑛𝑛+1 2 − 4
𝑦𝑦𝑛𝑛 =
𝑥𝑥𝑛𝑛+1
E1 (8)

As 𝑎𝑎 > 2 from (ii) 𝑥𝑥𝑛𝑛 → ∞ as 𝑛𝑛 → ∞ M1 and thus using result just proved, 𝑦𝑦𝑛𝑛 → 1 as 𝑛𝑛 → ∞ ,
i.e. the sequence converges. *A1 (2)
9.

Using the sine rule, from triangle PQR

𝑃𝑃𝑃𝑃 𝑃𝑃𝑃𝑃
=
sin 𝜃𝜃 sin �2𝜋𝜋 − 𝜑𝜑�
3
M1 A1

From triangle PQC

𝑃𝑃𝑃𝑃 𝑎𝑎 − 𝑥𝑥
𝜋𝜋 = 2𝜋𝜋
sin
3 sin � 3 − 𝜃𝜃�

A1

From triangle PBR

𝑃𝑃𝑃𝑃 𝑥𝑥
𝜋𝜋 = sin 𝜑𝜑
sin
3
A1

Eliminating PR and PQ between these three equations

𝜋𝜋 2𝜋𝜋 2𝜋𝜋 𝜋𝜋
𝑥𝑥 sin sin � − 𝜑𝜑� sin � − 𝜃𝜃� = sin 𝜑𝜑 sin 𝜃𝜃 (𝑎𝑎 − 𝑥𝑥 ) sin
3 3 3 3
M1 A1

Hence
√3 1 √3 1
𝑥𝑥 � cos 𝜑𝜑 + sin 𝜑𝜑� � cos 𝜃𝜃 + sin 𝜃𝜃� = (𝑎𝑎 − 𝑥𝑥 ) sin 𝜑𝜑 sin 𝜃𝜃
2 2 2 2

giving
�√3 cot 𝜑𝜑 + 1��√3 cot 𝜃𝜃 + 1�𝑥𝑥 = 4(𝑎𝑎 − 𝑥𝑥 )

as required. M1 *A1 (8)

If the ball has speed 𝑣𝑣1 moving from P to Q, speed 𝑣𝑣2 moving from Q to R, and speed 𝑣𝑣3 moving
from R to P,
2𝜋𝜋 𝜋𝜋
then CLM at Q parallel to CA gives 𝑣𝑣1 cos � − 𝜃𝜃� = 𝑣𝑣2 cos and NELI perpendicular to CA gives
3 3
2𝜋𝜋 𝜋𝜋 2𝜋𝜋 𝜋𝜋
𝑒𝑒𝑣𝑣1 sin � − 𝜃𝜃� = 𝑣𝑣2 sin , and dividing these gives 𝑒𝑒 tan � − 𝜃𝜃� = tan
3 3 3 3

M1 A1

and similarly,
𝜋𝜋
CLM at R parallel to AB gives 𝑣𝑣2 cos = 𝑣𝑣3 cos 𝜑𝜑 and NELI perpendicular to AB gives
3
𝜋𝜋 𝜋𝜋
𝑒𝑒𝑣𝑣2 sin = 𝑣𝑣3 sin 𝜑𝜑 , and dividing these gives 𝑒𝑒 tan = tan 𝜑𝜑 . A1
3 3

2𝜋𝜋 𝜋𝜋 −√3−tan 𝜃𝜃
𝑒𝑒 tan � − 𝜃𝜃� = tan yields 𝑒𝑒 = √3 M1 which simplifies to
3 3 1−√3 tan 𝜃𝜃
𝑒𝑒�√3 + tan 𝜃𝜃� = √3 �√3 tan 𝜃𝜃 − 1� , or in turn, (3 − 𝑒𝑒) tan 𝜃𝜃 = √3(1 + 𝑒𝑒) and so
(3−𝑒𝑒)
cot 𝜃𝜃 = A1
√3(1+𝑒𝑒)

𝜋𝜋 1
𝑒𝑒 tan = tan 𝜑𝜑 yields cot 𝜑𝜑 = A1
3 𝑒𝑒√3

Substituting these two expressions into the first result of the question,
1 (3 − 𝑒𝑒)
� + 1� � + 1� 𝑥𝑥 = 4(𝑎𝑎 − 𝑥𝑥 )
𝑒𝑒 (1 + 𝑒𝑒)

M1

This simplifies to
1 + 𝑒𝑒 4
𝑥𝑥 = 4(𝑎𝑎 − 𝑥𝑥 )
𝑒𝑒 1 + 𝑒𝑒
that is
𝑥𝑥 = 𝑒𝑒(𝑎𝑎 − 𝑥𝑥 )

so
𝑎𝑎𝑎𝑎
𝑥𝑥 =
1 + 𝑒𝑒
as required. *A1 (8)
2𝜋𝜋
To continue the motion at P, then similarly to before, the third impact gives 𝑒𝑒 tan � − 𝜑𝜑� = tan 𝜃𝜃
3

M1

So
−√3 − tan 𝜑𝜑 √3(𝑒𝑒 + 1)
tan 𝜃𝜃 = 𝑒𝑒 = 𝑒𝑒
1 − √3 tan 𝜑𝜑 3𝑒𝑒 − 1

and thus, using the previously found result for cot 𝜃𝜃

(3 − 𝑒𝑒) 3𝑒𝑒 − 1
=
√3(1 + 𝑒𝑒) √3(𝑒𝑒 + 1)𝑒𝑒
M1 A1

That is 𝑒𝑒(3 − 𝑒𝑒) = 3𝑒𝑒 − 1 , that is 𝑒𝑒 2 = 1 and as 𝑒𝑒 ≥ 0 , 𝑒𝑒 = 1 (and not -1) *B1 (4)
10. (i) At time t, the point where the string is tangential to the cylinder, M1 say T is at
(𝑎𝑎 cos 𝜃𝜃, 𝑎𝑎 sin 𝜃𝜃 ) , A1 the piece of string that remains straight is of length 𝑏𝑏 − 𝑎𝑎θ, M1 , the vector
−sin 𝜃𝜃
representing the string is thus (𝑏𝑏 − 𝑎𝑎θ) � � dM1 A1 so the particle is at the point
cos 𝜃𝜃
(𝑎𝑎 cos 𝜃𝜃 − (𝑏𝑏 − 𝑎𝑎θ) sin 𝜃𝜃 , 𝑎𝑎 sin 𝜃𝜃 + (𝑏𝑏 − 𝑎𝑎θ) cos 𝜃𝜃 ) . M1 A1 (7)

𝑥𝑥̇ = −𝑎𝑎𝜃𝜃̇ sin 𝜃𝜃 − (𝑏𝑏 − 𝑎𝑎θ)𝜃𝜃̇ cos 𝜃𝜃 + 𝑎𝑎𝜃𝜃̇ sin 𝜃𝜃 = −(𝑏𝑏 − 𝑎𝑎θ)𝜃𝜃̇ cos 𝜃𝜃

𝑦𝑦̇ = 𝑎𝑎𝜃𝜃̇ cos 𝜃𝜃 − (𝑏𝑏 − 𝑎𝑎θ)𝜃𝜃̇ sin 𝜃𝜃 − 𝑎𝑎𝜃𝜃̇ cos 𝜃𝜃 = −(𝑏𝑏 − 𝑎𝑎θ)𝜃𝜃̇ sin 𝜃𝜃
M1 A1

2 2
Thus the speed is ��(𝑏𝑏 − 𝑎𝑎θ)𝜃𝜃̇ cos 𝜃𝜃� + �(𝑏𝑏 − 𝑎𝑎θ)𝜃𝜃̇ sin 𝜃𝜃� = (𝑏𝑏 − 𝑎𝑎θ)𝜃𝜃̇ as required. M1 A1 (4)

(ii) The only horizontal force on the particle is the tension in the string, which is perpendicular to the
velocity at any time, so kinetic energy is conserved. E1 Therefore,
1 2 1
𝑚𝑚 �(𝑏𝑏 − 𝑎𝑎θ)𝜃𝜃̇� = 𝑚𝑚𝑢𝑢2
2 2
M1

and so, as (𝑏𝑏 − 𝑎𝑎θ)𝜃𝜃̇ and 𝑢𝑢 are both positive (𝑏𝑏 − 𝑎𝑎θ)𝜃𝜃̇ = 𝑢𝑢 *A1 (3)

(iii) The tension in the string, using instantaneous circular motion, at time t is
𝑚𝑚𝑢𝑢2
(𝑏𝑏 − 𝑎𝑎θ)
M1 A1

As (𝑏𝑏 − 𝑎𝑎θ)𝜃𝜃̇ = 𝑢𝑢 , integrating with respect to t,

𝑎𝑎𝜃𝜃 2
𝑏𝑏𝑏𝑏 − = 𝑢𝑢𝑢𝑢 + 𝑐𝑐
2
M1

but when 𝑡𝑡 = 0 , 𝜃𝜃 = 0 so 𝑐𝑐 = 0 . M1 A1
𝑎𝑎𝜃𝜃2
Thus, 𝑏𝑏𝑏𝑏 − = 𝑢𝑢𝑢𝑢
2

i.e.

2𝑏𝑏𝑏𝑏 𝑏𝑏 2 𝑏𝑏 2 2𝑢𝑢𝑢𝑢 𝑏𝑏 2 − 2𝑎𝑎𝑎𝑎𝑎𝑎

𝜃𝜃 2 − + 2= 2− =
𝑎𝑎 𝑎𝑎 𝑎𝑎 𝑎𝑎 𝑎𝑎2
Alternatively, integrating (𝑏𝑏 − 𝑎𝑎θ)𝜃𝜃̇ = 𝑢𝑢 with respect to t,

(𝑏𝑏 − 𝑎𝑎θ)2
− = 𝑢𝑢𝑢𝑢 + 𝑘𝑘
2𝑎𝑎
M1
 𝑏𝑏 2
When 𝑡𝑡 = 0 , 𝜃𝜃 = 0 so 𝑘𝑘 = − M1 A1
2𝑎𝑎

(𝑏𝑏 − 𝑎𝑎θ)2 𝑏𝑏 2 𝑏𝑏 2 − 2𝑎𝑎𝑎𝑎𝑎𝑎

= − 𝑢𝑢𝑢𝑢 =
2𝑎𝑎 2𝑎𝑎 2𝑎𝑎
Thus, taking positive roots,

𝑏𝑏 − 𝑎𝑎θ √𝑏𝑏 2 − 2𝑎𝑎𝑎𝑎𝑎𝑎

=
𝑎𝑎 𝑎𝑎
Hence, the tension is

𝑚𝑚𝑢𝑢2
√𝑏𝑏 2 − 2𝑎𝑎𝑎𝑎𝑎𝑎
*A1 (6)
11. (i)
𝑛𝑛+1
𝑛𝑛+1
𝑃𝑃(𝑌𝑌 = 𝑛𝑛) = 𝑃𝑃 (𝑛𝑛 ≤ 𝑋𝑋 < 𝑛𝑛 + 1) = � 𝜆𝜆 𝑒𝑒 −𝜆𝜆𝜆𝜆 𝑑𝑑𝑑𝑑 = �− 𝑒𝑒 −𝜆𝜆𝜆𝜆 �𝑛𝑛 = − 𝑒𝑒 −𝜆𝜆(𝑛𝑛+1) + 𝑒𝑒 −𝜆𝜆𝜆𝜆
𝑛𝑛

M1 M1

= �1 − 𝑒𝑒 −𝜆𝜆 �𝑒𝑒 −𝜆𝜆𝜆𝜆

as required. *A1 (3)

(ii)
∞ ∞ 𝑟𝑟+𝑧𝑧 ∞
−𝜆𝜆𝜆𝜆 𝑟𝑟+𝑧𝑧
𝑃𝑃(𝑍𝑍 < 𝑧𝑧) = � 𝑃𝑃(𝑟𝑟 ≤ 𝑋𝑋 < 𝑟𝑟 + 𝑧𝑧) = � � 𝜆𝜆 𝑒𝑒 𝑑𝑑𝑑𝑑 = ��− 𝑒𝑒 −𝜆𝜆𝜆𝜆 �𝑟𝑟
𝑟𝑟=0 𝑟𝑟=0 𝑟𝑟 𝑟𝑟=0

M1 M1
∞ ∞
−𝜆𝜆(𝑟𝑟+𝑥𝑥) −𝜆𝜆𝜆𝜆
= ��− 𝑒𝑒 + 𝑒𝑒 � = ��1 − 𝑒𝑒 −𝜆𝜆𝜆𝜆 �𝑒𝑒 −𝜆𝜆𝜆𝜆
𝑟𝑟=0 𝑟𝑟=0

M1 A1

1 1 − 𝑒𝑒 −𝜆𝜆𝜆𝜆
= �1 − 𝑒𝑒 −𝜆𝜆𝜆𝜆 � =
1 − 𝑒𝑒 −𝜆𝜆 1 − 𝑒𝑒 −𝜆𝜆
using sum of an infinite GP with magnitude of common ratio less than one.

M1 *A1 (6)
1−𝑒𝑒 −𝜆𝜆𝜆𝜆 𝑑𝑑 1−𝑒𝑒 −𝜆𝜆𝜆𝜆 𝜆𝜆𝜆𝜆 −𝜆𝜆𝜆𝜆
(iii) As 𝑃𝑃(𝑍𝑍 < 𝑧𝑧) = , 𝑓𝑓𝑍𝑍 (𝑧𝑧) = � �= M1
1−𝑒𝑒 −𝜆𝜆 𝑑𝑑𝑑𝑑 1−𝑒𝑒 −𝜆𝜆 1−𝑒𝑒 −𝜆𝜆

so
1 1
𝜆𝜆𝜆𝜆 −𝜆𝜆𝜆𝜆 1 1
𝐸𝐸 (𝑍𝑍) = � 𝑧𝑧 −𝜆𝜆
𝑑𝑑𝑑𝑑 = −𝜆𝜆
� �−𝑧𝑧𝑒𝑒 −𝜆𝜆𝜆𝜆 �0 + � 𝑒𝑒 −𝜆𝜆𝜆𝜆 𝑑𝑑𝑑𝑑�
1 − 𝑒𝑒 1 − 𝑒𝑒
0 0

M1 M1
1
1 −𝜆𝜆
𝑒𝑒 −𝜆𝜆𝜆𝜆 1 −𝜆𝜆
𝑒𝑒 −𝜆𝜆 1
= �−𝑒𝑒 − � � � = �−𝑒𝑒 − + �
1 − 𝑒𝑒 −𝜆𝜆 𝜆𝜆 0 1 − 𝑒𝑒 −𝜆𝜆 𝜆𝜆 𝜆𝜆

A1

1 𝑒𝑒 −𝜆𝜆
= −
𝜆𝜆 1 − 𝑒𝑒 −𝜆𝜆
or alternatively
1 �1 − (𝜆𝜆 + 1)𝑒𝑒 −𝜆𝜆 �
𝜆𝜆 1 − 𝑒𝑒 −𝜆𝜆

A1 (5)
(iv)
𝑃𝑃 (𝑌𝑌 = 𝑛𝑛 𝑎𝑎𝑎𝑎𝑎𝑎 𝑧𝑧1 < 𝑍𝑍 < 𝑧𝑧2 ) = 𝑃𝑃(𝑛𝑛 + 𝑧𝑧1 < 𝑋𝑋 < 𝑛𝑛 + 𝑧𝑧2 )
𝑛𝑛+𝑧𝑧2
𝑛𝑛+𝑧𝑧2
= � 𝜆𝜆 𝑒𝑒 −𝜆𝜆𝜆𝜆 𝑑𝑑𝑑𝑑 = �− 𝑒𝑒 −𝜆𝜆𝜆𝜆 �𝑛𝑛+𝑧𝑧 = − 𝑒𝑒 −𝜆𝜆(𝑛𝑛+𝑧𝑧2 ) + 𝑒𝑒 −𝜆𝜆(𝑛𝑛+𝑧𝑧1 ) = 𝑒𝑒 −𝜆𝜆𝜆𝜆 �𝑒𝑒 −𝜆𝜆𝑧𝑧1 − 𝑒𝑒 −𝜆𝜆𝑧𝑧2 �
1
𝑛𝑛+𝑧𝑧1

M1 A1

𝑃𝑃(𝑌𝑌 = 𝑛𝑛 𝑎𝑎𝑎𝑎𝑎𝑎 𝑧𝑧1 < 𝑍𝑍 < 𝑧𝑧2 ) = 𝑒𝑒 −𝜆𝜆𝜆𝜆 �𝑒𝑒 −𝜆𝜆𝑧𝑧1 − 𝑒𝑒 −𝜆𝜆𝑧𝑧2 �

1 − 𝑒𝑒 −𝜆𝜆𝑧𝑧2 1 − 𝑒𝑒 −𝜆𝜆𝑧𝑧1
= �1 − 𝑒𝑒 −𝜆𝜆 �𝑒𝑒 −𝜆𝜆𝜆𝜆 � − �
1 − 𝑒𝑒 −𝜆𝜆 1 − 𝑒𝑒 −𝜆𝜆

M1 A1

= 𝑃𝑃 (𝑌𝑌 = 𝑛𝑛) × 𝑃𝑃( 𝑧𝑧1 < 𝑍𝑍 < 𝑧𝑧2 ) M1

so Y and Z are independent. E1 (6)

12. (i)
1 5 1 5
𝑃𝑃(𝑋𝑋12 = 1) = , 𝑃𝑃(𝑋𝑋12 = 0) = , 𝑃𝑃(𝑋𝑋23 = 1) = , 𝑃𝑃(𝑋𝑋23 = 0) =
6 6 6 6

If 𝑋𝑋23 = 1 , then players 2 and 3 score the same as one another. In that case, 𝑋𝑋12 = 1 would mean
1
that player 1 also obtained that same score so 𝑃𝑃(𝑋𝑋12 = 1|𝑋𝑋23 = 1) = = 𝑃𝑃 (𝑋𝑋12 = 1).
6

If 𝑋𝑋23 = 1 , 𝑋𝑋12 = 0 would mean that player 1 obtained a different score so

5
𝑃𝑃 (𝑋𝑋12 = 0|𝑋𝑋23 = 1) = = 𝑃𝑃(𝑋𝑋12 = 0)
6
If 𝑋𝑋23 = 0 , then players 2 and 3 score differently to one another. In that case, 𝑋𝑋12 = 1 would mean
1
that player 1 also obtained the same score as player 2 so 𝑃𝑃(𝑋𝑋12 = 1|𝑋𝑋23 = 0) = = 𝑃𝑃(𝑋𝑋12 = 1)
6

If 𝑋𝑋23 = 0 , 𝑋𝑋12 = 0 would mean that player 1 obtained a different score to player 2 so
5
𝑃𝑃 (𝑋𝑋12 = 0|𝑋𝑋23 = 0) = = 𝑃𝑃(𝑋𝑋12 = 0)
6
Hence 𝑋𝑋12 is independent of 𝑋𝑋23 . M1 A1 (2)

Alternatively,

𝑋𝑋12 𝑋𝑋23
1 1 requires players 2 and 3 to both score same as player 1 so
1 1 1
𝑃𝑃(𝑋𝑋12 = 1 𝑎𝑎𝑎𝑎𝑎𝑎 𝑋𝑋23 = 1 ) = = × = 𝑃𝑃 (𝑋𝑋12 = 1) × 𝑃𝑃 (𝑋𝑋23 = 1)
36 6 6
1 0 requires player 2 to score the same as player as player 1, and player 3 score differently so
5 1 5
𝑃𝑃(𝑋𝑋12 = 1 𝑎𝑎𝑎𝑎𝑎𝑎 𝑋𝑋23 = 0 ) = = × = 𝑃𝑃 (𝑋𝑋12 = 1) × 𝑃𝑃 (𝑋𝑋23 = 0)
36 6 6
0 1 requires players 2 and 3 to score the same as one another, and player 1 score differently so
5 5 1
𝑃𝑃(𝑋𝑋12 = 0 𝑎𝑎𝑎𝑎𝑎𝑎 𝑋𝑋23 = 1 ) = = × = 𝑃𝑃 (𝑋𝑋12 = 0) × 𝑃𝑃 (𝑋𝑋23 = 1)
36 6 6
0 0 requires both player 1 and 3 to score differently to player 2 so
25 5 5
𝑃𝑃(𝑋𝑋12 = 0 𝑎𝑎𝑛𝑛𝑑𝑑 𝑋𝑋23 = 0 ) = = × = 𝑃𝑃 (𝑋𝑋12 = 0) × 𝑃𝑃 (𝑋𝑋23 = 0)
36 6 6
Hence 𝑋𝑋12 is independent of 𝑋𝑋23 . M1 A1 (2)

If total score is T , then

𝑇𝑇 = � 𝑋𝑋𝑖𝑖𝑖𝑖
𝑖𝑖<𝑗𝑗

M1
so
1 5 𝑛𝑛(𝑛𝑛 − 1)
𝐸𝐸 (𝑇𝑇) = 𝐸𝐸 �� 𝑋𝑋𝑖𝑖𝑖𝑖 � = � 𝐸𝐸�𝑋𝑋𝑖𝑖𝑖𝑖 � = 𝑛𝑛𝐶𝐶2 𝐸𝐸 (𝑋𝑋12 ) = 𝑛𝑛𝐶𝐶2 �1 × +0× �=
6 6 12
𝑖𝑖<𝑗𝑗 𝑖𝑖<𝑗𝑗

M1 A1

1 5 12
𝑉𝑉𝑉𝑉𝑉𝑉(𝑇𝑇) = 𝑉𝑉𝑉𝑉𝑉𝑉 �� 𝑋𝑋𝑖𝑖𝑖𝑖 � = � 𝑉𝑉𝑉𝑉𝑉𝑉�𝑋𝑋𝑖𝑖𝑖𝑖 � = 𝑛𝑛𝐶𝐶2 𝑉𝑉𝑉𝑉𝑉𝑉(𝑋𝑋12 ) = 𝑛𝑛𝐶𝐶2 �12 × + 02 × − �
6 6 6
𝑖𝑖<𝑗𝑗 𝑖𝑖<𝑗𝑗

M1
5𝑛𝑛(𝑛𝑛 − 1)
=
72
A1 (5)

(ii)
𝑉𝑉𝑉𝑉𝑉𝑉(𝑌𝑌1 + 𝑌𝑌2 + … + 𝑌𝑌𝑚𝑚 ) = 𝐸𝐸 ((𝑌𝑌1 + 𝑌𝑌2 + … + 𝑌𝑌𝑚𝑚 )2 ) − [𝐸𝐸 (𝑌𝑌1 + 𝑌𝑌2 + … + 𝑌𝑌𝑚𝑚 )]2

= 𝐸𝐸�𝑌𝑌1 2 + 𝑌𝑌2 2 + … + 𝑌𝑌𝑚𝑚 2 + 2𝑌𝑌1 𝑌𝑌2 + 2𝑌𝑌1 𝑌𝑌3 + … + 2𝑌𝑌𝑛𝑛−1 𝑌𝑌𝑛𝑛 � − [𝐸𝐸 (𝑌𝑌1 ) + 𝐸𝐸 (𝑌𝑌2 ) + … + 𝐸𝐸 (𝑌𝑌𝑚𝑚 )]2
𝑚𝑚 𝑚𝑚−1 𝑚𝑚
2
= 𝐸𝐸 �� 𝑌𝑌𝑖𝑖 � + 2𝐸𝐸 � � � 𝑌𝑌𝑖𝑖 𝑌𝑌𝑗𝑗 � − (0 + 0 + ⋯ + 0)2
𝑖𝑖=1 𝑖𝑖=1 𝑗𝑗=𝑖𝑖+1

𝑚𝑚 𝑚𝑚−1 𝑚𝑚
2
= � 𝐸𝐸�𝑌𝑌𝑖𝑖 � + 2 � � 𝐸𝐸�𝑌𝑌𝑖𝑖 𝑌𝑌𝑗𝑗 �
𝑖𝑖=1 𝑖𝑖=1 𝑗𝑗=𝑖𝑖+1

M1 *A1 (2)

(iii)
1 1 1
𝑃𝑃(𝑍𝑍12 = 1) = × =
2 6 12
If 𝑍𝑍23 = 1 then player 2 has rolled an even score and player 3 has scored the same so, in this case,
1
for 𝑍𝑍12 = 1 , require player 1 to roll the score that player has so 𝑃𝑃(𝑍𝑍12 = 1|𝑍𝑍23 = 1) = .
6

Therefore, 𝑃𝑃(𝑍𝑍12 = 1) ≠ 𝑃𝑃 (𝑍𝑍12 = 1|𝑍𝑍23 = 1) and thus 𝑍𝑍12 and 𝑍𝑍23 are not independent.

Alternatively,
1 1
𝑃𝑃(𝑍𝑍12 = 1) = , 𝑃𝑃(𝑍𝑍23 = 1) =
12 12

For 𝑍𝑍12 = 1 and 𝑍𝑍23 = 1 we require all three players to score the same even number so
3 1 1 1 1 1
𝑃𝑃(𝑍𝑍12 = 1 𝑎𝑎𝑎𝑎𝑎𝑎 𝑍𝑍23 = 1 ) = × × = ≠ × = 𝑃𝑃(𝑍𝑍12 = 1) × 𝑃𝑃(𝑍𝑍23 = 1)
6 6 6 72 12 12
and thus they are not independent. M1 A1 (2)

Using part (ii), let 𝑌𝑌1 = 𝑍𝑍12 , let 𝑌𝑌2 = 𝑍𝑍13 , … let 𝑌𝑌𝑚𝑚 = 𝑍𝑍(𝑛𝑛−1)𝑛𝑛
𝑛𝑛(𝑛𝑛−1)
(and with 𝑚𝑚 = 𝑛𝑛𝐶𝐶2 = ).
2
 1 1 5 1
𝑃𝑃(𝑍𝑍12 = 1) = , 𝑃𝑃(𝑍𝑍12 = −1) = , 𝑃𝑃(𝑍𝑍12 = 0) = so 𝐸𝐸 (𝑍𝑍12 ) = 0 and 𝐸𝐸�𝑍𝑍12 2 � = and
12 12 6 6
likewise for all other Z (Y!). B1 B1

If total score is U , then

𝑈𝑈 = � 𝑍𝑍𝑖𝑖𝑖𝑖
𝑖𝑖<𝑗𝑗

so

𝐸𝐸 (𝑈𝑈) = 𝐸𝐸 �� 𝑍𝑍𝑖𝑖𝑖𝑖 � = � 𝐸𝐸�𝑍𝑍𝑖𝑖𝑖𝑖 � = 0

𝑖𝑖<𝑗𝑗 𝑖𝑖<𝑗𝑗

B1

which means we can apply the result of (ii).

If 𝑍𝑍12 = 1 then 𝑍𝑍13 = 1 or 𝑍𝑍13 = 0

If 𝑍𝑍12 = −1 then 𝑍𝑍13 = −1 or 𝑍𝑍13 = 0

Otherwise 𝑍𝑍12 = 0
1 1 1
So 𝐸𝐸 (𝑍𝑍12 𝑍𝑍13 ) = 1 × 1 × + −1 × −1 × = M1 A1
72 72 36

So
𝑛𝑛(𝑛𝑛 − 1) 1 𝑛𝑛−1 1 𝑛𝑛(𝑛𝑛 − 1) 𝑛𝑛(𝑛𝑛 − 1)(𝑛𝑛 − 2)
𝑉𝑉𝑉𝑉𝑉𝑉(𝑈𝑈) = × + 2 × 𝑛𝑛 × 𝐶𝐶2 × = +
2 6 36 12 36
M1 M1 A1

𝑛𝑛(𝑛𝑛 − 1) 𝑛𝑛(𝑛𝑛 − 1)(𝑛𝑛 + 1) 𝑛𝑛(𝑛𝑛2 − 1)

= �3 + (𝑛𝑛 − 2)� = =
36 36 36
*A1 (9)
Cambridge Assessment Admissions Testing offers a range of tests to support selection and
recruitment for higher education, professional organisations and governments around the world.
Underpinned by robust and rigorous research, our assessments include:

 assessments in thinking skills

 admissions tests for medicine and healthcare
 behavioural styles assessment
 subject-specific admissions tests.

We are part of a not-for-profit department of the University of Cambridge.

Cambridge Assessment
Admissions Testing
The Triangle Building
Shaftesbury Road
Cambridge
CB2 8EA
United Kingdom

Admissions tests support:

www.admissionstestingservice.org/help