Stephen SikloS

Advanced Problems in
Mathematics
Preparing for University
 To access digital resources including:
blog posts
videos
online appendices

and to purchase copies of this book in:

hardback
paperback
ebook editions

Go to:
[Link]

Open Book Publishers is a non-profit independent initiative.

We rely on sales and donations to continue publishing
high-quality academic works.
Advanced Problems in Mathematics:
Preparing for University

Stephen Siklos
 [Link]

c 2016 Stephen Siklos.

!

This work is licensed under a Creative Commons Attribution 4.0 International license (CC BY 4.0). This
license allows you to share, copy, distribute and transmit the work; to adapt the work and to make
commercial use of the work providing attribution is made to the author (but not in any way that suggests
that they endorse you or your use of the work). Attribution should include the following information:
Stephen Siklos, Advanced Problems in Mathematics: Preparing for University. Cambridge, UK: Open Book
Publishers, 2015. [Link] 10.11647/OBP.0075
Further details about CC BY licenses are available at [Link]

Digital material and resources associated with this volume are available at
[Link]
STEP questions reproduced by kind permission of Cambridge Assessment Group Archives.

This is the first volume of the OBP Series in Mathematics:

ISSN 2397-1126 (Print)

ISSN 2397-1134 (Online)

ISBN Paperback: 9781783741427

ISBN Digital (PDF): 9781783741441
ISBN Digital ebook (epub): 9781783741458
ISBN Digital ebook (mobi): 9781783741465
DOI: 10.11647/OBP.0075

c Diane Potter. Creative Commons Attribution-

Cover image: Paternoster Vents (2012). Photograph !
NonCommercial-NoDerivs CC BY-NC-ND

All paper used by Open Book Publishers is SFI (Sustainable Forestry Initiative) and PEFC (Programme
for the Endorsement of Forest Certification Schemes) Certified.

Printed in the United Kingdom and United States by

Lightning Source for Open Book Publishers
 Contents

About this book ix

STEP 1

Worked Problems 11
Worked problem 1 11
Worked problem 2 15

Problems 19
P1 An integer equation 19
P2 Partitions of 10 and 20 21
P3 Mathematical deduction 23
P4 Divisibility 25
P5 The modulus function 27
P6 The regular Reuleaux heptagon 29
P7 Chain of equations 31
P8 Trig. equations 33
P9 Integration by substitution 35
P10 True or false 37
P11 Egyptian fractions 39
P12 Maximising with constraints 41
P13 Binomial expansion 43
P14 Sketching subsets of the plane 45
P15 More sketching subsets of the plane 47
P16 Non-linear simultaneous equations 49
P17 Inequalities 51
P18 Inequalities from cubics 53
P19 Logarithms 55
P20 Cosmological models 57
P21 Melting snowballs 59
P22 Gregory’s series 61
P23 Intersection of ellipses 63
P24 Sketching x (1 − x)
m n
65
vi Advanced Problems in Mathematics

P25 Inequalities by area estimates 67

P26 Simultaneous integral equations 69
P27 Relation between coefficients of quartic for real roots 71
P28 Fermat numbers 73
P29 Telescoping series 75
P30 Integer solutions of cubics 77
P31 The harmonic series 79
P32 Integration by substitution 81
P33 More curve sketching 83
P34 Trig sum 85
P35 Roots of a cubic equation 88
P36 Root counting 90
P37 Irrationality of e 92
P38 Discontinuous integrands 94
P39 A difficult integral 96
P40 Estimating the value of an integral 98
P41 Integrating the modulus function 100
P42 Geometry 102
P43 The t substitution 104
P44 A differential-difference equation 106
P45 Lagrange’s identity 108
P46 Bernoulli polynomials 110
P47 Vector geometry 112
P48 Solving a quartic 114
P49 Areas and volumes 116
P50 More curve sketching 118
P51 Spherical loaf 120
P52 Snowploughing 122
P53 Tortoise and hare 124
P54 How did the chicken cross the road? 126
P55 Hank’s gold mine 128
P56 A chocolate orange 130
P57 Lorry on bend 132
P58 Fielding 134
P59 Equilibrium of rod of non-uniform density 136
P60 Newton’s cradle 138
P61 Kinematics of rotating target 140
P62 Particle on wedge 142
P63 Sphere on step 144
 Stephen Siklos vii

P64 Elastic band on cylinder 146

P65 A knock-out tournament 148
P66 Harry the calculating horse 150
P67 PIN guessing 152
P68 Breaking plates 154
P69 Lottery 156
P70 Bodies in the fridge 158
P71 Choosing keys 160
P72 Commuting by train 162
P73 Collecting voles 164
P74 Breaking a stick 166
P75 Random quadratics 168

Syllabus 170
 About this book

This book has two aims.

• The general aim is to help bridge the gap between school and university mathematics.
You might wonder why such a gap exists. The reason is that mathematics is taught at school for
various purposes: to improve numeracy; to hone problem-solving skills; as a service for students
going on to study subjects that require some mathematical skills (economics, biology, engineering,
chemistry — the list is long); and, finally, to provide a foundation for the small number of students
who will continue to a specialist mathematics degree. It is a very rare school that can achieve
all this, and almost inevitably the course is least successful for its smallest constituency, the real
mathematicians.

• The more specific aim is to help you to prepare for STEP or other examinations required for uni-
versity entrance in mathematics. To find out more about STEP, read the next section.

It used to be said that mathematics and cricket were not spectator sports; and this is still true of math-
ematics. To progress as a mathematician, you have to strengthen your mathematical muscles. It is not
enough just to read books or attend lectures. You have to work on problems yourself.
One way of achieving the first of the aims set out above is to work on the second, and that is how this
book is structured. It consists almost entirely of problems for you to work on.
The problems are all based on STEP questions. I chose the questions either because they are ‘nice’ —
in the sense that you should get a lot of pleasure from tackling them (I did), or because I felt I had
something interesting to say about them.
The first two problems (the ‘worked problems’) are in a stream of consciousness format. They are in-
tended to give you an idea how a trained mathematician would think when tackling them. This ap-
proach is much too long-winded to sustain for the remainder of the book, but it should help you to see
what sort of questions you should be asking yourself as you work on the later problems.
Each subsequent problem occupies two pages. On the first page is the STEP question, followed by a
comment. The comments may contain hints, they may direct your attention to key points, and they may
include more general discussions. On the next page is a solution; you have to turn over, so that your eye
cannot accidentally fall on a key line of working. The solutions give enough working for you to be able
to read them through and pick up at least the gist of the method; they may not give all the details of the
calculations. For each problem, the given solution is of course just one way of producing the required
result: there may be many other equally good or better ways. Finally, if there is space on the page after
the solution (which is sometimes not the case, especially if diagrams have to be fitted in), there is a
postmortem. The postmortems may indicate what aspects of the solution you should be reviewing and
they may tell you about the ideas behind the problems.
I hope that you will use the comments and solutions as springboards rather than feather beds. You will
only really benefit from this book if you have a good go at each problem before looking at the comment
and certainly before looking at the solution. The problems are chosen so that there is something for
you to learn from each one, and this will be lost to you for ever if you simply read the solution without
thinking about the problem on your own.
x Advanced Problems in Mathematics

I have given each problem a difficulty rating ranging from ! to !!!. Difficulty in mathematics is in the
eye of the beholder: you might find a question difficult simply because you overlooked some key step,
which on another day you would not have hesitated over. You should not therefore be discouraged if
you are stuck on a !-question; though you should probably be encouraged if you get through one of
the rare !!!-questions without mishap.
This book is about depth not breadth. I have not tried to teach you any new topics. Instead, I want
to lead you towards a deeper understanding of the material you already know. I therefore restricted
myself to problems requiring knowledge of the specific and rather limited syllabus that is laid out at the
end of the book. The pure mathematics section corresponds to the syllabus for STEP papers I and II. If
you are studying British A-levels, Scottish Highers, or the International Baccalaureate, for example, you
will be familiar with most of this material. For the mechanics and statistics/probability sections, there
is less agreement about what a core syllabus should be (in the IB there is no mechanics at all), so I gave
myself a freer hand.
Calculators are not required for any of the problems in this book and calculators are not permitted in
STEP examinations. In the early days of STEP, calculators were permitted but they were not required for
any question. It was found that candidates who tried to use calculators sometimes ended up missing
the point of the question or getting a silly answer. My advice is to remove the battery so that you are not
tempted.
I started this section by listing the aims of the book. You may have noticed that teaching you mathemat-
ics is not an aim. I can’t remember where I heard the following rather nice analogy. In 1464, a huge block
of Carrara marble was carefully chosen from a quarry in Tuscany and transported to Florence, where it
lay almost untouched for many years. In 1501 it was given to the sculptor Michelangelo. He worked
hard on it, chipping away and chipping away for three years, until at last, inside the block, he found a
beautiful statue of David. You can see a picture at:
[Link]
And the analogy? I can’t teach you mathematics with this book, but I believe that much hard work on
your part, chipping away at the problems, will eventually reveal the mathematician that is within you.
I hope you enjoy using this book as much as I have enjoyed putting it together.
 STEP

What is STEP?

STEP (Sixth Term Examination Paper) is an examination used by Cambridge University as part of its
procedure for admitting students to study mathematics. Applicants are interviewed in December, and
may then be offered a place conditional on the results of their public examinations (A-level, International
Baccalaureate, etc) and STEP. The examinations are sat in June and offers are confirmed in August when
all the examination results are available.
STEP is used for conditional offers not just by Cambridge, but (at the time of writing) also by Warwick
University for almost all of its Mathematics offers, and to a lesser extent by some other English universi-
ties. Many other university mathematics departments recommend that their applicants practise on the
past papers even if they do not take the examination. In 2015, 4322 scripts were marked, only about
1000 of which were written by students holding an offer from Cambridge.
The first STEPs were taken in 1987, and there was were specimen papers before that from which some
of the questions in this book were drawn. At that time, there were STEPs in many subjects but by 2001
only the mathematics papers remained. The examination has been more or less stable over nearly 30
years: it has not been blown about by the various fads in the public examinations systems that came
and went during that time.
There are three STEPs, called papers I, II and III. Each paper has thirteen questions, including three on
mechanics and two on probability/statistics. Candidates are assessed on six questions only.
The pure mathematics question in Papers I and II are based the core A-level Mathematics syllabus, with
some minor additions, which is listed at the end of this book. The pure mathematics questions in Paper
III are based on a ‘typical’ Further Mathematics mathematics A-level syllabus (at the time of writing,
there is not even a partial core for Further Mathematics A-levels).
There is also no core (at the time of writing) for A-level mechanics and statistics, so the STEP syllabuses
for these areas consist of material that a student with a particular interest might have covered. It has to
be said, though, that the statistics questions are very likely to require knowledge of probability rather
than statistics (for example, there are very few questions on statistical tests of given data). This is because
the underlying theory of statistics is quite difficult, and therefore unsuitable for examining at this level,
whereas the application of statistical tests is rather routine and again unsuitable for examination at this
level.

What is the purpose of STEP?

From the point of view of admissions to a university mathematics course, STEP has three purposes.

• It is used as a hurdle for entrance to university mathematics courses, and sometimes for other
mathematics-based courses. There is strong evidence that success in STEP correlates very well
2 Advanced Problems in Mathematics

with university examination results.1

• It acts as preparation for the university course, because the style of mathematics found in STEP
questions is similar to that of undergraduate mathematics.

• It tests motivation. It is important to prepare for STEP (by working through old papers, for exam-
ple), which can require considerable dedication. Those who are not willing to make the effort are
unlikely to thrive on a difficult university mathematics course.

STEP vs A-level
A-level2 tests mathematical knowledge and technique by asking you to tackle fairly stereotyped prob-
lems. STEP asks you to apply the same knowledge and technique to problems that are, ideally, unfamil-
iar.
Here is an A-level question, in which you follow the instructions in the question:

By using the substitution u = 2x − 1, or otherwise, find

!
2x
dx.
(2x − 1)2

And here, for comparison, is a STEP question, which requires both competence in basic mathematical
techniques and mathematical intuition. Note that help is given for the first integral, so that everyone
starts at the same level. Then, for the second integral, candidates have to show that they understand
why the substitution used in the first part worked, and how it can be adapted.

Use the substitution x = 2 − cos θ to evaluate the integral

! 2 " #1
x−1 2
dx.
3/2 3−x

Show that, for a < b,

! q " #1 √
x−a 2 (b − a)(π + 3 3 − 6)
dx = ,
p b−x 12
where p = (3a + b)/4 and q = (a + b)/2.

The differences between STEP and A-level are:

1. STEP questions are much longer. Candidates completing four questions in three hours will almost
certainly get a grade 1.

2. STEP questions are much less routine.

3. STEP questions may require considerable dexterity in performing mathematical manipulations.

4. Individual STEP questions may require knowledge of several different areas of mathematics (espe-
cially the mechanics and statistics questions, which will often require advanced pure mathematical
techniques).

1
Recent studies comparing rank in STEP with rank in first-year Cambridge mathematics examinations reveal a Spearman corre-
lation coefficient of 0.63, which is very high in comparison with other predictors of university examination results.
2
I use the term ‘A-level’ here as a shorthand for a typical school mathematics examination. The particular examinations you take
may well be very different in style and format but, even if that is the case, I am sure some of what follows will strike a chord
with you.
 Stephen Siklos 3

5. The marks available for each part of the question are not disclosed on the paper.

6. There is a lot of choice on STEP papers (6 questions out of 13).

7. Calculators are not permitted in STEP examinations.

These difference matter, because in mathematics more than in any other subject it is very important to
match the difficulty of the question with the ability of the candidates. For example, you could reasonably
have the question ‘Was Henry VIII a good king?’ on a lower-school history paper, an A-level paper, or as
a PhD topic. The answers would (or should) differ according to the level. On mathematics examination
papers, the question has to be tailored to the level in order to discriminate between the candidates: if it
is too easy, nearly all candidates will score very high marks; if it is too hard, nearly all candidates will
make little progress on any of the questions.

Setting STEP
STEP is produced under the auspices of the Cambridge Assessment examining board. The setting proce-
dure starts 30 months before the date of the examination, when the three examiners (one for each paper,
from schools or universities) are asked to produce a draft paper. The first drafts are then vetted by the
STEP coordinator (me!), who tries to enforce uniformity of difficulty, checks suitability of material and
style, and tries to reduce overlap between the papers. Examiners then produce a second draft, based on
the coordinator’s suggestions. The second drafts are agreed with the coordinator and then circulated
to three moderators (normally school teachers), and to the other examiners, who produce written com-
ments and discuss the drafts in a two-day meeting. The examiners then produce third drafts, taking into
account the consensus at the meetings. These drafts are sent to a vetter, who works through the papers,
pointing out mistakes and infelicities. The resulting draft is checked by a second vetter and finally by
a team of students. At each stage, the drafts are produced camera-ready, using a special mathematical
word-processing package called LaTeX (which is also used to typeset this book).3

STEP Questions
STEP questions do not fall into any one category. Typically, there will be a range of types on each of the
papers. Here are some thoughts, in no particular order.

• My favourite sort of question is in two (or maybe more) parts: in the first part, candidates are
asked to perform some unfamiliar task and are told how to do it (integration using a given substi-
tution, or expressing a quartic as the algebraic sum of two squares, for example); for the later parts,
candidates are expected to demonstrate that they have understood and learned from the first part
by applying the method to a new and perhaps more complicated task.

• Another favourite of mine is the question which has different answers according to the value of a
certain number (or parameter). A common example involves sketching a graph whose shape de-
pends on whether a parameter is positive or negative. Ideally, the different values of the parameter
are not given in the question, and candidate has to identify them for herself or himself.

3
It is freely available, so you might like to try it out. You will probably find it good fun to use, but quite time-consuming and not
really suitable for writing out the solutions of STEP problems.
4 Advanced Problems in Mathematics

• Another good type of question requires candidates to do some preliminary special-case work and
then prove a general result.

• In another type, candidates have to show that they can understand and use new notation or a new
theorem.

• Questions with several unrelated parts (for example, three integrals using different techniques)
are generally avoided; but if they occur, there tends to be a ‘sting in the tail’ involving putting all
the parts together in some way.

• Some questions do not rely on any part of the syllabus: instead they might require ‘common sense’,
involving counting or seeing patterns, or they might involve some aspect of more elementary
mathematics with an unusual slant. Such questions try to test capacity for clear and logical thought
without using much mathematical knowledge (like the calendar question mentioned in the section
on preparation below, or questions concerning islands populated by toads ‘who always tell the
truth’ and frogs ‘who always fib’).

• Some questions are devised to check that you do not simply apply routine methods blindly. For
example, a function might have a maximum value at the end of the interval upon which it is
defined, even though its derivative might be non-zero there. Finding the maximum in such a case
is not simply a question of routine differentiation.

• There are always questions specifically on integration or differentiation, and many others (includ-
ing mechanics and probability) that use calculus as a means to an end.

• Graph-sketching is regarded by mathematicians as a fundamental skill and there are nearly always
questions that require a sketch.

• Basic ideas from analysis, such as

) b
0 " f (x) " k ⇒ 0 " f (x)dx " (b − a)k
a
or
f ! (x) > 0 ⇒ f (b) > f (a) for b > a
or the relationship between an integral and a sum often come up, though knowledge of such
results is never assumed — candidates may be told that they can use the result without proof, or a
sketch ‘proof’ may be asked for.

• As mentioned above, questions on statistical tests are rare, because questions that require real
understanding (rather than ‘cookbook’ methods) tend to be too difficult. More often, the questions
in the Probability and Statistics section are about probability.

• The mechanics questions normally require a firm understanding of the basic principles (when to
apply conservation of momentum and energy, for example) and may well involve a differential
equation. Projectile questions are often set, but are never routine.

Advice to candidates

First appearances
I am often asked whether STEP is ‘difficult’. Of course, it depends on what is meant by ‘difficult’; it is
not difficult compared with the mathematics I do every day. But to be on the safe side, I always answer
‘yes’ before explaining further.
 Stephen Siklos 5

Your first impression on looking at a STEP paper is likely to be that it does indeed look very difficult.
Don’t be discouraged! Its difficult appearance is largely due to it being very different in style from what
you are used to.
At the time of writing, a typical A-level examination lasts 90 minutes and contains 10 compulsory ques-
tions. That is 9 minutes per question. If you are considering studying mathematics at a top university, it
is likely that you will manage to do them all and get them nearly all right in the time available. A STEP
examination lasts 3 hours, and you are only supposed to do six questions in three hours. You are very
likely to get a grade 1 if you manage four questions (not necessarily complete); that means that each
question is designed to take 45 minutes. If you compare a 9 minute question with a 45 minute question,
of course the 45 minute question looks very hard!
You may be put off by the number of subjects covered on the paper. You should not be. STEP is sup-
posed to provide sufficient questions for all candidates, no matter which mathematics syllabus (at the
appropriate level) they have covered. It would be a very exceptional candidate who had the knowledge
required to do all the questions. And there is plenty of choice (6 questions out of 13).
Once you get used to the idea that STEP is very different from A-level, it becomes much less daunting.

Preparation

The best preparation for STEP (apart, of course, from working through this excellent book) is to work
slowly through old papers.4 Hints and answers are available for some years, but you should use these
with discretion: doing a question with hints and answers in front of you is nothing like doing it yourself,
and you may well miss the whole point of the question (which is to make you think about mathematics).
In general, thinking about the problem is much more important than getting the answer.
Should you try to learn up areas of mathematics that are not in your syllabus in preparation for STEP?
The important thing to know is that it is much better to be very good at your syllabus than to have have
a sketchy knowledge of lots of additional topics: depth rather than breadth is what matters. It may
conceivably be worth your while to round out your knowledge of a topic you have already studied to
fit in with the STEP syllabus; it is probably not worth your while to learn a new topic for the purposes
of the exam, though I can think of a couple of exceptions:

• Hyperbolic functions obey simple rules similar to trigonometric functions (in fact, they are the
same functions in the complex plane). If you haven’t come across them, you can easily master
them in a short time and this will open the door to many questions that would otherwise have
been inaccessible to you.

• de Moivre’s theorem (relating to complex numbers) is also very straightforward, though questions
requiring de Moivre are less common.

It is worth emphasising that there is no ‘hidden agenda’: a candidate who does two complete probability
questions and two complete mechanics questions will obtain the same mark and grade as one who does
four complete pure questions.
Just as the examiners have no hidden agenda concerning syllabus, so they have no hidden agenda
concerning your method of answering the question. If you can get to the end of a question correctly you

4
These and the other publications mentioned below are obtainable from the STEP web site [Link]. You
will also find other useful material on the NRICH web site; do google it now if you haven’t already done so.
6 Advanced Problems in Mathematics

will get full marks whatever method you use.5 Some years ago one of the questions asked candidates
to find the day of the week of a given date (say, the 5th of June 1905). A candidate who simply counted
backwards day by day from the date of the exam would have received full marks for that question (but
would not have had time to do any other questions).
You may be worried that the examiners expect some mysterious thing called rigour. Do not worry: STEP
is an exam for schools, not universities, and the examiners understand the difference. Nevertheless, it is
extremely important that you present ideas clearly, and show working at all stages.

Presentation
You should set out your answer legibly and logically (don’t scribble down the first thought that comes
into your head) – this not only helps you to avoid silly mistakes but also signals to the examiner that
you know what you are doing (which can be effective even if you haven’t the foggiest idea what you
are doing).
Examiners are not as concerned with neatness as you might fear. However if you receive complaints
from your teachers that your answers are difficult to follow then you should listen.6 Remember that
more space usually means greater legibility. Try writing on alternate lines (this leaves a blank line for
corrections).
Try to read your answers with a hostile eye. Have you made it clear when you have come to the end
of a particular argument? Try underlining your conclusions. Have you explained what you are trying
to do? For example, if a question asks ‘Is A true?’ try beginning your answer by writing ‘A is true’ —
if you think that it is true — so that the examiner knows which way your argument leads. If you used
some idea (for example, integration by substitution), did you tell the examiner that this is what you
were doing?

What to do if you cannot get started on a problem

Try the following, in order.

• Reread the question to check that you understand what is wanted.

• Reread the question to look for clues – the way it is phrased, or the way a formula is written,
or other relevant parts of the question. (You may think that the setters are trying to set difficult
questions or to catch you out. Usually, nothing could be further from the truth: they are probably
doing all in their power to make it easy for you by trying to tell you what to do).

• Try to work out exactly what it is that you don’t understand.

• Simplify the notation – e.g. by writing out sums explicitly.

• Look at special cases (choose special values which simplify the problem) in order to try to under-
stand why a result is true.

• Write down your thoughts – in particular, try to express the exact reason why you are stuck.

5
Though you must obey instructions in the question: for example, if it says ‘Hence prove ...’, then you must use the previous
result in your proof.
6
Begin rant: I am very surprised at the scrappy and illegible work that I receive from a few of my students. It seems so dis-
respectful to expect me to spend ages trying to decipher their work when they could have spent a little more time making it
presentable, for example by copying it out neatly or writing more slowly. Why is my time less important than theirs? End rant.
 Stephen Siklos 7

• Go on to another question and go back later.

• If you are preparing for the examination (but not in the actual examination!) take a short break.7

• Discuss it with a friend or teacher (again, better not do this in the actual examination) or consult
the hints and answers, but make sure you still think it through yourself.
BUT REMEMBER: following someone else’s solution is not remotely the same thing as doing the
problem yourself. Once you have seen someone else’s solution to a problem, then you are de-
prived, for ever, of much of the benefit that could have come from working it out yourself.
Even if, ultimately, you get stuck on a particular problem, you derive vastly more benefit from
seeing a solution to something with which you have already struggled, than by simply following
a solution to something to which you’ve given very little thought.

What if a problem isn’t coming out?

If you have got started but the answer doesn’t seem to be coming out, then try the following.

• Check your algebra. In particular, make sure that what you have written works in special cases.
For example: if you have written the series for log(1 + x) as

1 − x + 12 x2 − 13 x3 + · · ·

then a quick check will reveal that it doesn’t work for x = 0; clearly, the 1 should not be there.8
A note on the subject of algebra. In many of the problems in this book, the algebra is quite stiff:
you have to go through many lines of calculation before you get to an expression recognisably
close to your target. Really, the only way to manage this efficiently is to check each line carefully
before going on the next line. Otherwise, you can waste hours.9

• Make sure that what you have written makes sense. For example, in a problem which is dimen-
sionally consistent, you cannot add x (with dimension length, say) to x2 or to exp x (which itself
does not make sense — the argument of exp has to be dimensionless). Even if there are no dimen-
sions in the problem, it is often possible to mentally assign dimensions and hence enable a quick
check.
Be wary of applying familiar processes to unfamiliar objects (very easy to do when you are feeling
at sea): for example, it is all too easy, if you are not sure where your solution is going, to solve the
vector equation a.x = 1 by dividing both sides by vector a; a bad idea.

• Analyse exactly what you are being asked to do. Try to understand the hints, explicit and implicit.
Remember to distinguish between terms such as explain/prove/define/etc. (There is essentially
no difference between ‘prove’ and ‘show’: the former tends to be used in more formal situations,
but if you are asked to ‘show’ something, a proper proof is required.)

• Remember that different parts of a question are often linked. There may be guidance in the nota-
tion and choice of names of variables in the question.

7
J. E. Littlewood (1885–1977), distinguished Cambridge mathematician and author of the highly entertaining Mathematicians
Miscellany (Cambridge University Press, 1986) used to work seven days a week until an experiment revealed that when he took
Sundays off the good ideas had a way of coming on Mondays.
8
Another check will reveal that for very small positive x, log is positive (since its argument is bigger than 1) whereas the series
is negative, so there is clearly something else wrong.
9
Which what I normally do; but I am still hoping to heed my own advice in the not too distant future.
8 Advanced Problems in Mathematics

• If you get irretrievably stuck in the exam, state in words what you are trying to do and move on
(at A-level, you don’t get credit for merely stating intentions, but STEP examiners are generally
grateful for any sign of intelligent life).

What to do after completing a question

It is a natural instinct to consider that you have finished with a question once you have got to an answer.
However this instinct should be resisted both from a general mathematical point of view and from the
much narrower view of preparing for an examination. Instead, when you have completed a question
you should stop for a few minutes and think about it. Here is a check list for you to run through.

• Look back over what you have done, checking that the arguments are correct and making sure that
they work for any special cases you can think of. It is surprising how often a chain of completely
spurious arguments and gross algebraic blunders leads to the given answer.

• Check that your answer is reasonable. For example if the answer is a probability p then you should
check that 0 " p " 1. If your answer depends on an integer n, does it behave as it should when
n → ∞? Is it dimensionally correct?
If, in the exam, you find that your answer is not reasonable, but you don’t have time to do anything
about it, then write a brief phrase showing that you understand that your answer is unreasonable
(e.g. ‘This is wrong because mass must be positive’).

• Check that you have used all the information given. In many ways the most artificial aspect of
examination questions is that you are given exactly the amount of information required to answer
the question. If your answer does not use everything you are given then either it is wrong or you
are faced with a very unusual examination question.

• Check that you have understood the point of the question. It is, of course, the case that not all
exam questions have a point, but many do. What idea did the examiners want you to have?
Which techniques did they want you to demonstrate? Is the result of the question interesting in
some way? Does it generalise? If you can see the point of the question would your working show
the point to someone who did not know it in advance?

• Make sure that you are not unthinkingly applying mathematical tools which you do not fully
understand.10

• It is good preparation for the examination to try to see how the problem fits into the wider context
and see if there is a special point which it is intended to illustrate. You may need help with this.

• As preparation for the examination, make sure that you actually understand not only what you
have done, but also why you have done it that way rather than some other way. This is particularly
important if you have had to use a hint or solution.

10
Mathematicians should feel as insulted as engineers by the following joke.

A mathematician, a physicist and an engineer enter a mathematics contest, the first task of which is to prove that all odd
number are prime. The mathematician has an elegant argument: ‘1’s a prime, 3’s a prime, 5’s a prime, 7’s a prime. Therefore,
by mathematical induction, all odd numbers are prime. It’s the physicist’s turn: ‘1’s a prime, 3’s a prime, 5’s a prime, 7’s a
prime, 11’s a prime, 13’s a prime, so, to within experimental error, all odd numbers are prime.’ The most straightforward proof
is provided by the engineer: ‘1’s a prime, 3’s a prime, 5’s a prime, 7’s a prime, 9’s a prime, 11’s a prime . . . ’.
 Stephen Siklos 99

Solution to problem 40
x6
(i) Let f (x) = . Then
(x2 + 1)4

6x5 8x7 2(3 − x2 )x5

f ! (x) = − =
(x2 + 1)4 (x2 + 1)5 (x2 + 1)5

which is positive for 0 < x2 < 3. Therefore, f (x) increases in value from 0 at x = 0 to 1
16 at x = 1.
(ii) Doing the differentiation gives

1 5Ax4 + 3Bx2 + C 6x(Ax5 + Bx3 + Cx) Dx6

≡ − + 2
(x2 + 1) 4 2
(x + 1) 3 2
(x + 1) 4 (x + 1)4

and multiplying by (x2 + 1)4 gives the following identity:

1 ≡ (5Ax4 + 3Bx2 + C)(x2 + 1) − 6x(Ax5 + Bx3 + Cx) + Dx6 (∗)

i.e.
1 ≡ (D − A)x6 + (5A − 3B)x4 + (3B − 5C)x2 + C .

Equating coefficients of the different powers of x on each side of the equivalence sign gives 1 = C,
0 = 3B − 5C, 0 = 5A − 3B, 0 = D − A, so A = 1, B = 53 , C = 1 and D = 1.
(iii) Using the results of parts (i) and (ii), we see that for 0 " x " 1
" 5 5 3 # " 5 5 3 #
d x + 3x + x 1 d x + 3x + x 1
2 3
" 2 4
" 2 3
+ .
dx (x + 1) (x + 1) dx (x + 1) 16

Inequalities can be integrated, so

) 1 " 5 5 3 # ) 1 ) 1 " 5 5 3 # ) 1
d x + 3x + x 1 d x + 3x + x 1
2 3
dx " 2 4
dx " 2 3
dx + dx
0 dx (x + 1) 0 (x + 1) 0 dx (x + 1) 0 16

i.e. $ %1 ) $ 5 %1 & '

1
x5 + (5/3)x3 + x 1 x + (5/3)x3 + x x 1
" dx " +
(x2 + 1)3 0 0
2
(x + 1) 4 2
(x + 1) 3
0 16 0
from which the required result follows immediately.

Post-mortem
Instead of equating coefficients in (∗), you could obtain equations for A , B , C and D by putting four
carefully chosen values of x into the equation. An obvious choice is x = 0, but (thinking flexibly!) you
could try x = i to eliminate terms with factors of x2 + 1 . Then it becomes more difficult to find good
choices.
You might wonder why, in part (ii), the term inside the derivative has only odd powers of x. Would it
not make it more general to include even powers as well? You could in fact include even powers but
you would find that their coefficients would be zero: all the other terms in the equation are even in x,
so the derivative has to be an even function which means that the function being differentiated must be
odd.
Although the idea of this question is good, the final result is a bit feeble. It only gives the value of the
integral to an accuracy of about 16 / 24 which is about 15%. The actual value of the integral can be found
1 11

fairly easily using the substitution x = tan t and is 1148 + 64 π so the inequalities can be used to give
5

(rather bad) estimates for π namely 2.933 " π " 3.733 .

100 Advanced Problems in Mathematics

Problem 41: Integrating the modulus function (!!)

Show that ) ) )
1 0 1
x x
| xe |dx = − xe dx + xex dx
−1 −1 0

and hence evaluate the integral.

Evaluate the following integrals:
) 4
(i) | x3 − 2x2 − x + 2 | dx ;
0

) π
(ii) | sin x + cos x | dx .
−π

2000 Paper I

Comments
The very first part shows you how to do this sort of integral (with mod signs in the integrand) by
splitting up the range of integration at the points where the integrand changes sign. In parts (i) and (ii)
you have to use the technique on different examples.
142 Advanced Problems in Mathematics

Problem 62: Particle on wedge (!!!)

A wedge of mass M rests on a smooth horizontal surface. The face of the wedge is a smooth
plane inclined at an angle α to the horizontal. A particle of mass m slides down the face of the
wedge, starting from rest. At a later time t, the speed V of the wedge, the speed v of the particle
and the angle β of the velocity of the particle below the horizontal are as shown in the diagram.

❩ ✉
❩
❩
❩❆
❆❩❩
✛ ❆❯ ❩
V v ❩
❆ ❩
❩
❆ ❩
❆ α❩
❆β
❆
Let y be the vertical distance descended by the particle. Derive the following results, stating in
(ii) and (iii) the mechanical principles you use:

(i) V sin α = v sin(β − α) ;

(ii) tan β = (1 + k) tan α, where k = m/M ;

(iii) 2gy = v 2 (1 + k cos2 β) .

" #2
dy g tan2 β gK tan2 α
Hence show that 2g y = !
, where g ! = 2 = and K = 1 + k. Deduce
dt tan β + K K tan2 α + 1
that y = 12 g ! t2 .

1998 Paper II

Comments
I was surprised at the difficulty of this problem, compared with a particle on a fixed wedge. There
is an easier method (just using forces and Newton’s second law) but the route suggested here uses
basic principles, the outcomes of which are the numbered parts of the question. For a fixed wedge, the
horizontal component of momentum is not conserved, because of the force required to hold the wedge;
but (i) and (iii) are just what you would use in the fixed case.
Part (ii) is used only (apart from the very last result) to show that the angle β is constant. This is a bit at
first sight surprising: it means that the particle moves in a straight line.
Most of the important intermediate results are given to you, but it is still very good discipline to check
that they hold in special cases: for example, k = 0 corresponding to a massless particle or equivalently
a fixed wedge; and α = 0 or α = 12 π corresponding to a horizontal or vertical wedge face. You should
check that you understand what should happen in these special cases and that your understanding is
consistent with the formulae.
 Stephen Siklos 143

Solution to problem 62
(i) This equation follows immediately from the constraint placed on the particle: it remains in contact
with the plane so components of the velocities of the particle and the wedge perpendicular to the face
must be equal.
(ii) Horizontal momentum is conserved, so M V = mv cos β. Substitution for V /v using (i) gives

V sin(β − α) sin β cos α − sin α cos β sin β

k cos β = = = = − cos β
v sin α sin α tan α
which leads immediately to the required result. Note that the angle β remains constant in the motion.
(iii) Conservation of energy gives

mgy = 12 M V 2 + 12 mv 2 , i.e. kgy = 12 (kv cos β)2 + 12 kv 2 .

Thus 2gy = (k cos2 β + 1)v 2 .

The next step is to write v in terms of y to obtain the differential equation. The vertical component of
velocity is v sin β, so
dy
v sin β = .
dt
Now using the result of (iii) gives
( )
dy 2gy 2gy sin2 β 1 1
= v sin β = sin β = = (2g ! ) 2 y 2 . (∗)
dt 1 + k cos2 β 1 + (K − 1) cos2 β

Squaring gives the required result, and we can use part (ii) to obtain the expression for g ! in terms of α
and K.
Now, at last, we use the hard-won result that β, and hence g ! , is constant. The equation
" #2
dy
= 2g ! y
dt

is exactly the same as the equation of conservation of energy for a particle falling vertically in a gravita-
tional field of strength g ! . For such a particle, we have y = 12 g ! t2 , as required.

Post-mortem
The basic principles behind this solution are conservation of momentum and energy, together with
the constraint that the particle moves on the surface of the face of the wedge. The benefit of using
conservation of energy is that you do not have to worry about the normal reaction force, because it does
no work.
Of course, we could obtain the final result by integrating the differential equation (∗), which takes only
a few lines.
144 Advanced Problems in Mathematics

Problem 63: Sphere on step (!)

A uniform solid sphere of radius a and mass m is drawn very slowly and without slipping
from horizontal ground onto a step of height 12 a by a horizontal force of magnitude F which
is always applied to the highest point of the sphere and is always perpendicular to the vertical
plane which forms the face of the step. Find the maximum value of F in the motion, and prove
√
that the coefficient of friction between the sphere and the edge of the step must exceed 1/ 3.

1997 Paper II

Comments
This is quite straightforward once you have realised that very slowly means so slowly that the sphere can
be considered to be static at each position. You just have to solve a statics problem with the usual tools:
resolving forces and taking moments about suitably chosen points.
It seems odd to have a force that always acts at the highest point of the sphere. The idea is that there is
a string wrapped round the sphere and it is being pulled horizontally to get the sphere up the step.
 Stephen Siklos 145

Solution to problem 63

✬✩
f R
F ✛❅■
❅ ""
✒
❅ "
❅ "
✫✪
❅" θ

❄
mg
Let the angle between the radius to the point of contact with the step and the downward vertical be θ,
as shown. At the point of contact, let the frictional force, which is tangent to the sphere, be f and the
reaction R (along the radius of the sphere).
To investigate F , we take moments about a point cunningly chosen to eliminate other unknown forces.
Since the lines of action both R and f pass through the point of contact of the sphere with the step, we
can eliminate both forces by taking moments about this point:

F a(1 + cos θ) = mga sin θ ,

so
sin θ
F = mg = mg tan 12 θ .
1 + cos θ
This takes its maximum value when θ is largest, i.e. when the sphere just touches the horizontal ground.
√ √
At this position, cos θ = 12 and sin θ = 12 3 and Fmax = mg/ 3.
Taking moments about the centre of the sphere (to eliminate the normal reaction and weight) gives
F = f , and resolving forces parallel to the radius at the point of contact gives

R = F sin θ + mg cos θ.

Now
sin2 θ 1 − cos2 θ
F sin θ = mg = mg = mg(1 − cos θ)
1 + cos θ 1 + cos θ

so R = mg. We therefore need µmg > Fmax , as required.

Post-mortem
Should we have known that R = mg (in the last line of the solution)? It doesn’t look obvious from
the diagram. However, if we follow the principle, used twice already, of taking moments about a point
that eliminates unwanted forces, the result drops out. To relate mg and R directly, we need to eliminate
F and f . We therefore need to take moments about the intersection of the lines of action of these two
forces, which is shown in the diagram (call it P ). Moments must balance about any point: it doesn’t
matter whether the point is inside the body, or outside it as in the case of P . It is easy to see from the
geometry (f and F act along tangents to the sphere) that the distances from P to the lines of action of
mg and R are equal, which means that the two forces are equal.
To be precise about the meaning of slow in this context, you have to compare the dynamic forces con-
nected with the motion of the sphere with the static forces. The motion of the sphere is rotation about
the fixed point of contact with the step. If we assume that the centre moves with constant speed, the
only extra force due to the motion is an additional reaction at the point of contact with the step. This
reaction is centrifugal in nature and so is roughly of the form mv 2 /a, which is to be small compared
with R. The static approximation (‘very slowly’) is therefore valid if v 2 ) ga.
146 Advanced Problems in Mathematics

Problem 64: Elastic band on cylinder (!)

A smooth cylinder with circular cross-section of radius a is held with its axis horizontal. A light
elastic band of unstretched length 2πa and modulus of elasticity λ is wrapped round the circum-
ference of the cylinder, so that it forms a circle in a plane perpendicular to the axis of the cylinder.
A particle of mass m is then attached to the rubber band at its lowest point and released from
rest.

(i) Given that the particle falls to a distance 2a below the axis of the cylinder, but no further,
show that
9πmg
λ= √ .
(3 3 − π)2

(ii) Given instead that the particle reaches its maximum speed at a distance 2a below the axis
of the cylinder, find a similar expression for λ .

2001 Paper I

Comments
This question uses the most basic ideas in mechanics, such as conservation of energy. I included it in
this selection of questions without realising that the properties of stretched strings are not in the syllabus
(which is given in the appendix). However, I didn’t throw it out: it is a nice question and the only two
things you need to know about stretched strings are, for a stretched string of natural (i.e. unstretched)
length l and extended length l + x with modulus of elasticity λ:

λx2
(i) the potential energy stored in the stretched string is ;
2l
λx
(ii) the tension in the stretched string, by Hooke’s law, is .
l
 Stephen Siklos 147

Solution to problem 64
The diagram below shows the system when the particle has fallen a distance a from its initial position.

a a

2a

A bit of geometry (including Pythagoras) on the above diagram shows that the length of the extended
√
band is 2πa − 2a arccos 12 + 2 3 a so the extension of the band is
√ √
−2a arccos 12 + 2 3 a , i.e. 2a( 3 − 13 π) .

(i) At the lowest point, the speed is zero. That suggests using an energy equation, which will involve
speed and displacement (for the potential energy) but not acceleration.
Conserving energy (taking the initial potential energy to be zero), we find that when the particle has
fallen a distance a and has speed v,
√ 1 2
1 2 1 [2a( 3 − 3 π)]
0 = 2 mv + 2 λ − mga .
2πa
At the lowest point, v = 0, which gives the required answer.
(ii) Now we are interested in the point where the speed is greatest, i.e. where the acceleration is zero, so
this time we should use the equation of motion of the particle.
The component of the tension T in the band in the vertical direction, acting upwards on the particle,
when the particle has fallen a distance y is 2T cos θ where sin θ = a/(a + y). Applying Newton’s second
law to the motion of the particle gives

mÿ = 2T cos θ − mg , (∗)

where, by Hooke’s law, T = λ × extension/2πa . At the maximum speed, ÿ = 0. This occurs at y = a, so

(∗) becomes √ √
2a( 3 − 13 π) 3
0 = 2λ − mg .
2πa 2
√
3 πmg
In this case λ = √ .
3 3−π

Post-mortem
The question was relatively simple because you only needed to evaluate energy or tension at very special
points: maximum extension or maximum speed. You could of course have worked from the general
* * + ,
equation of motion (∗), with cos θ = 2ay + y 2 /(a+y) and extension 2 2ay + y 2 −2a arcsin a/(a+y) ,
but you can’t solve the differential equation to find the general motion. It would be possible for small
oscillations.
148 Advanced Problems in Mathematics

Problem 65: A knock-out tournament (!)

A tennis tournament is arranged for 2n players. It is organised as a knockout tournament, so that

only the winners in any given round proceed to the next round. Opponents in each round except
the final are drawn at random, and in any match either player has a probability 12 of winning.
Two players are chosen at random before the start of the first round. Find the probabilities that
they play each other:

(i) in the first round;

(ii) in the final round;

(iii) in the tournament.

1987 Specimen Paper II

Comments
Note that the set-up is not the usual one for a tennis tournament, where the only random element is in
the first round line-up. Two players cannot then meet in the final if they are in the same half of the draw.
Part (i) is straightforward, but parts (ii) and (iii) need a bit of thought. There is a short way and a long
way of tackling these parts, and both have merits.
It is a good plan to check your answers, if possible, by reference to simple special cases where you can
see what the answers should be; n = 1 or n = 2, for example.
Interestingly, the answers are independent of the probability that the players have of winning a match;
the 2s in the answers represent the number of players in each match rather than (the reciprocal of) the
probability that each player has of winning a match. It also does not matter how the draw for each
round is made. This is clear if you use the short method mentioned above.
 Stephen Siklos 149

Solution to problem 65
Call the two players P1 and P2 .
(i) Once P1 has been given a slot, there are 2n − 1 slots for P2 , in only one of which will he or she play P1 .
The probability of P1 playing P2 is therefore
1
.
2n −1
Note that this works for n = 1 and n = 2.
(ii) Long way. To meet in the final, P1 and P2 must each win every round before the final, and must also
not meet before the final. The probability that P1 and P2 do not meet in the first round and that they
both win their first round matches, is
" # " #
1 1 1 2n−1 − 1
1− n i.e. .
2 − 1 22 2 2n − 1
The probability that they win each round and do not meet before the final (i.e. for n − 1 rounds) is
" # " # " #
1 2n−1 − 1 1 2n−2 − 1 1 21 − 1 1 1
× × · · · × i.e. .
2 2n − 1 2 2n−1 − 1 2 22 − 1 2n−1 2n − 1

(ii) Short way. Since all processes are random here, the probability that any one pair contests the final is
the same as that for any other pair. There are a total of 12 × 2n (2n − 1) different pairs, so the probability
for any given pair is 1/[2n−1 (2n − 1)].
(iii) Long way. We need to add the probabilities that P1 and P2 meet in each round. The probability that
they meet in the kth round is the probability that they reach the kth round times the probability that they
meet in the kth round given that they reach it, the latter (conditional) probability being 1/(2n−k+1 − 1),
as can be inferred from part (i). As in part (ii), the probability that they reach the kth round is

1 2n−k+1 − 1
,
2k−1 2n − 1
so the probability that they meet in the kth round is

1 2n−k+1 − 1 1 1 1
× n−k+1 = k−1 n .
2k−1 2n − 1 2 −1 2 2 −1
Summing this as a geometric progression from k = 1 to n gives 1/2n−1 .
(iii) Short way. By the same short argument as in part (ii), the probability of a given pair meeting in any
given match (not necessarily the final) is 1/[2n−1 (2n − 1)]. Since the total number of matches is 2n − 1
(because one match is needed to knock out each player, and all players except one get knocked out), the
probability of a given pair playing is
2n − 1 1
= n−1 .
2n−1 (2n − 1) 2

Post-mortem
If (like me) you plodded through this question the long way, you might be wondering how you were
supposed to think of the short way. Instead of working out what happens to individual players as they
progress through the tournament, you think about the space of all possible outcomes (the sample space),
and attach a probability to each. That way, you can use the symmetry between all the players to help
you.
150 Advanced Problems in Mathematics

Problem 66: Harry the calculating horse (!!)

Harry the Calculating Horse will do any mathematical problem I set him, providing the answer
is 1, 2, 3 or 4. When I set him a problem, he places a hoof on a large grid consisting of unit
squares and his answer is the number of squares partly covered by his hoof. Harry has circular
hoofs, of radius 14 unit.
After many years of collaboration, I suspect that Harry no longer bothers to do the calculations,
instead merely placing his hoof on the grid completely at random. I often ask him to divide 4 by
4, but only about 14 of his answers are right; I often ask him to add 2 and 2, but disappointingly
only about 16 1
π of his answers are right. Is this consistent with my suspicions?
I decide to investigate further by setting Harry many problems, the answers to which are 1, 2, 3,
or 4 with equal frequency. If Harry is placing his hoof at random, find the expected value of his
answers. The average of Harry’s answers turns out to be 2. Should I get a new horse?

1997 Paper II

Comments
Hans von Osten, a horse, lived in Berlin around the turn of the last century. He was known far and
wide for his ability to solve complex arithmetical problems. Distinguished scientists travelled to Berlin
to examine Hans and test his marvellous ability. They would write an equation on a chalkboard and
Hans would respond by pawing the ground with his hoof. When Hans reached the answer he would
stop. Though he sometimes made errors, his success rate was far higher than would be expected if his
answers were random. The accepted verdict was that Hans could do arithmetic.
Hans’s reputation as a calculating horse nosedived when an astute scientist simply made sure that nei-
ther the person asking the questions nor the audience knew the answers. Hans became an instant failure.
His success was based on his ability to sense any change in the audience: a lifted eyebrow, a sigh, a nod-
ding head or the tensing of muscles was enough to stop him from pawing the ground. Anyone who
knew the answer was likely to give almost imperceptible clues to the horse. But we shouldn’t overlook
Hans’s talents: at least he had terrific examination technique.
We are investigating the situation when Harry places his hoof at random, so that the probability of the
centre of his hoof lands in any given region is proportional to the area of the region. We therefore move
swiftly from a question about probability to a question about areas. You just have to divide a given
square into regions, each determined by the number of squares that will be partially covered by Harry’s
hoof if its centre lands in the region under consideration. Remember that the areas must add to one, so
the most difficult area calculation can be left until last and deduced from the others.
To answer the last part properly, you really need to set out a hypothesis testing argument: you will accept
the null hypothesis (random hoof placing) if, using the distribution implied by the null hypothesis, the
probability of obtaining the given result is greater than some pre-decided figure. Obviously, nothing so
elaborate was intended here, since it is the last demand of an already long question: just one line would
do.
 Stephen Siklos 151

Solution to problem 66
The diagram shows one square of the grid.
Harry’s hoof will land completely within exactly one square if he places the centre of his hoof in a square
of side 12 centred on the centre of any square, shown with horizontal and vertical lines in the diagram.
The area of any such square is 14 square units.
His hoof will partially cover exactly four squares if he places the centre in a circle of radius 14 centred on
any intersection of grid lines. The total area of any one such circle is 161
π square units, which may be
thought of as four quarter circles, one in each corner of any given square, shown lightly shaded in the
diagram.
His hoof will partially cover exactly two squares if he places the centre in any one of four 1
2 by 1
4 rectan-
gles, of total area 4 × 18 in any given square, shown darkly shaded in the diagram.
Otherwise, his hoof will partially cover three squares; the area of this remaining region, shown with no
shading in the diagram, is
1 π 1
1− − − .
4 16 2

If Harry placed his hoof at random, the probabilities of the different outcomes would be equal to the
corresponding area calculated above (divided by the total area of the square, which is 1). Thus the data
given in the question are exactly consistent with random placement.
The expected value given random placements is
" #
1 1 1 1 π π π
1× +2× +3× 1− − − +4× =2+
4 2 4 2 16 16 16

The expected value (given that Harry gets all questions right) is (1 + 2 + 3 + 4)/4 = 5/2 . Harry has a less
accurate expected value even than the random expected value. He is clearly hopeless and should go.

Post-mortem
The explanations in the first paragraph of the solution are very concise. To understand them, you should
draw yourself lots of diagrams.
The proportion of Harry’s answers that are correct, in the case of the 2 + 2 calculation is given as ‘about
16 π. That should have given you pause for thought. Why not 5 ? Clearly, it was supposed to be help-
1 1

ful, so it should have been no surprise that the probability of getting the answer 4 under the random
assumption turned out to be exactly 16 1
π. Of course, these questions shouldn’t turn into a game of
spotting hidden clues, but it is always worth looking for unusual features of a question — or of any
mathematical problem.
152 Advanced Problems in Mathematics

Problem 67: PIN guessing (!)

In order to get money from a cash dispenser I have to punch in a Personal Identification Number.
I have forgotten my PIN, but I do know that it is equally likely to be any one of the integers 1,
2, . . . , n. I plan to punch in integers in ascending order until I get the right one. I can do this at
the rate of r integers per minute. As soon as I punch in the first wrong number, the police will
be alerted. The probability that they will arrive within a time t minutes is 1 − e−λt , where λ is
a positive constant. If I follow my plan, show that the probability of the police arriving before I
get my money is
-n
1 − e−λ(k−1)/r
.
n
k=1

Simplify the sum.

On past experience, I know that I will be so flustered that I will just punch in possible integers
at random, without noticing which I have already tried. Show that the probability of the police
arriving before I get my money is

1
1− .
n − (n − 1)e−λ/r

2002 Paper I

Comments
This was originally about getting money from a cash dispenser using a stolen card, but it was decided
that STEP questions should not be immoral. Hence the rather more improbable scenario.
The exponential distribution, here governing the police arrival time, is often used for failure of equipment
(light-bulbs, etc). It has the useful property (not used here) that the reliability for a light bulb (here
the probability of the police not coming within a certain time period of duration t) doesn’t depend on
which time period you choose; i.e. given that the light bulb has survived to time T , the probability of it
surviving until time T + t is independent of T (which doesn’t seem very suitable for light bulbs). This
is referred to as the memoryless property.
To simplify the sums, you have to recognise a geometric series where the common ratio of terms is an
exponential; not difficult, but easy to miss the first time you see it.
 Stephen Siklos 153

Solution to problem 67
The probability that I get the right number on the kth go is 1/n . The time taken to key in k integers33 is
(k − 1)/r . The probability that the police arrive before this is 1 − e−λ(k−1)/r so the total probability that
the police arrive before I get my money is

1 / /
n
-
× 1 − e−λ(k−1)/r .
n
k=1

We have
1 / /
-n -n
−λ(k−1)/r e−λ(k−1)/r 1 − e−λn/r
× 1−e =1− =1− ,
k=1
n
k=1
n n(1 − e−λ/r )

since the last part of the sum is a geometric progression with common ratio e−λ/r .
This time, the probability of getting my money on the kth go is 1/n times the probability of not having
punched in the correct integer in the preceding k − 1 turns:
" #k−1
1 n−1
P(money on kth go) = × .
n n

Thus the probability of the police arriving before I get my money is

-∞ " #k−1 / /
1 n−1 1 1 1 1
× × 1 − e−λ(k−1)/r = × − × −λ/r
,
n n n 1 − (n − 1)/n n 1 − e (n − 1)/n
k=1

which is easily seen to be the given answer.

Post-mortem
For each part, you have to calculate the probability of the first success occurring on the kth go. You
have to remember to multiply the probability of success on this go by the probability of failure on the
preceding k − 1 goes — a very typical idea in this sort of probability question. After that, the question is
really just algebra. STEP questions on this sort of material nearly always involve significant algebra or
calculus.
Car batteries as well as light bulbs often crop up as examples of objects to which exponential reliability
can be applied. Suppose your car battery has a guarantee of three years, and your car is completely
destroyed after two years. Your insurance company offers to give you one third of the price of the
battery. What do you think they would say to your counterclaim for the full price, based on the fact that,
given it had lasted two years, the exponential model says that it would have had an expected further
three years life in it?

33
Think of fences and posts: k posts here but only k − 1 fences.
154 Advanced Problems in Mathematics

Problem 68: Breaking plates (!)

Four students, one of whom is a mathematician, take turns at washing up over a long period
of time. The number of plates broken by any student in this time obeys a Poisson distribution,
the probability of any given student breaking n plates being e−λ λn /n! for some fixed constant
λ, independent of the number of breakages by other students. Given that five plates are broken,
find the probability that three or more were broken by the mathematician.

2001 Paper I

Comments
The way this is set up, it is largely a counting exercise (but see postmortem). To start with, you work
out the probability of five breakages, then follow that with the probability that the mathematician broke
three of more plates. You need to calculate the number of ways that 5 plates can be shared amongst four
students, for which you have to consider each partition of the number 5 and the number of different
ways it can arise.
 Stephen Siklos 155

Solution to problem 68
First we work out the probability of five breakages.
Let P(5, 0, 0, 0) denote the probability that student A breaks 5 plates and students B, C and D break no
plates. Then

e−λ λ5 + −λ ,3 λ5 e−4λ
P(5, 0, 0, 0) = (Prob. of breaking 5) × (Prob of breaking none)3 = e = .
5! 5!
The probability that one student breaks all the plates is 4 × λ5 e−4λ /5!, the factor 4 because it could be
any one of the four students.
Let P(4, 1, 0, 0) denote the probability that student A breaks 4 plates, student B breaks one plate and
students C and D break no plate. Then

P(4, 1, 0, 0) = (Prob. of breaking 4) × (Prob. of breaking one) × (Prob of breaking none)2

e−λ λ4 e−λ λ + −λ ,2 λ5 e−4λ
= e = .
4! 1! 4!
The probability that any one student breaks four plates and any other student breaks one plate is there-
fore 4 × 3 × λ5 e−4λ /4! .
Considering P(3, 2, 0, 0), P(3, 1, 1, 0), P(2, 2, 1, 0) and P(2, 1, 1, 1) in turn, together with the number of
different ways these probabilities can occur, shows that the probability of five breakages is
" # " #
5 −4λ 4 12 12 12 12 4 5 −4λ 1024
λ e + + + + + =λ e .
5! 4! 1! 3! 2! 3! 1! 1! 2! 2! 1! 2! 1! 1! 1! 5!
The probability that the mathematician (student A, say) breaks three or more is P(5, 0, 0, 0)+3P(4, 1, 0, 0)+
3P(3, 2, 0, 0) + 3P(3, 1, 1, 0), i.e.
" # " #
4 3 3 4 106
λ5 e−4λ + + + = λ5 e−4λ .
5! 4! 1! 3! 2! 3! 1! 1! 5!
The probability that the mathematician breaks three or more, given that five are broken is therefore
106/1024 .

Post-mortem
If you did the question by the method suggested above, you will probably be wondering about the
answer: why is it independent of λ; and why is the denominator 45 . You will quickly decide that the
Poisson distribution was a red herring (though I promise you that it was not an intentional herring),
since there is no trace of the distribution in the answer.
The denominator is pretty suggestive. A completely different approach is as follows, It is clear that the
probability that the mathematician breaks any given plate is 14 . The probability that he or she breaks k
plates is therefore binomial:
" # " #k " #5−k
5 1 3
k 4 4

and adding the cases k = 3, k = 4 and k = 5 gives rather rapidly

90 + 15 + 1
.
1024
For the method given in the solution, we could substitute Pk for e−λ λk /k! (as the probability of breaking
k plates) and the calculation would work just the same, with all the Pk s disappearing; try it.
156 Advanced Problems in Mathematics

Problem 69: Lottery (!!)

The national lottery of Ruritania is based on the positive integers from 1 to N , where N is very
large and fixed. Tickets cost £1 each. For each ticket purchased, the punter (i.e. the purchaser)
chooses a number from 1 to N . The winning number is chosen at random, and the jackpot is
shared equally amongst those punters who chose the winning number.
A syndicate decides to buy N tickets, choosing every number once to be sure of winning a share
of the jackpot. The total number of tickets purchased in this draw is 3.8N and the jackpot is £W .
Assuming that the non-syndicate punters choose their numbers independently and at random,
find the most probable number of winning tickets and show that the expected net loss of the
syndicate is approximately + ,
5 1 − e−2.8
N − W .
14
2001 Paper II

Comments
This is a binomial distribution problem: the probability that n out of m punters choose the winning
ticket is " #
m n m−n
p q
n
where here m = 2.8N , p = 1/N and q = 1 − p. It is clear that an approximation to the binomial
distribution is expected (for example, the question uses the word ‘approximately’ and you have to think
about how an approximation might arise); and the presence of the exponential in the given result gives a
pretty broad hint that it should be a Poisson distribution — the use of which has to be justified. One can
expect the Poisson approximation to work when the number of trials (call it m) is large (e.g. m > 150)
and when np ≈ npq, i.e. the mean is roughly equal to the variance (since these are equal for the Poisson
distribution) — so q ≈ 1.
You can’t find the most probable number of winning tickets by differentiation (unless you fancy differ-
entiating the factorial x!); instead, you must look at the ratios of consecutive terms and see when these
turn from being greater than one to less than one.
 Stephen Siklos 157

Solution to problem 69
Let X be the random variable whose value is the number of winning tickets out of the 2.8N tickets
purchased by the non-syndicate punters. Then X ∼ Poisson(2.8) .
Let pj = P(X = j) . Then
(2.8)j e−2.8 pj+1 2.8
pj = ⇒ = .
j! pj j+1
This fraction is greater than 1 if j < 1.8, so the most probable number of winning tickets by non-
syndicate punters is 2. Overall (including the ticket bought by the syndicate), the most probable number
of winning tickets is 3, which is very plausible.
The expected winnings of the syndicate is
" # " #
1 1 −2.8 1 1 1 1 2
W p0 + W p1 + W p 2 + · · · = e W 1 + × (2.8) + × (2.8) + · · ·
2 3 2 1! 3 2!
2.8
e − 1
= e−2.8 W ,
2.8
so the expected loss is as given (note that 2.8 = 5 ).
14

Post-mortem
This is rather interesting. You might perhaps have considered whether it would be worth borrowing
money to buy every single lottery combination, in order to win a share of the jackpot. Clearly, the
people who run the lottery have to think about this sort of thing.
Suppose the number of tickets sold, excluding the N that we plan to buy, is expected to be kN . Suppose
also that a fraction α of the total is paid out in the jackpot. Then, setting 2.8 = k and W = α(k + 1)N ,
the expected loss formula given in the question becomes
" #
1 − e−k
N 1− α(k + 1) .
k

If k is very small (take k = 0), we lose N (1 − α) (obviously). If k is large, k = 3 say, the exponential can
be ignored and we lose N (1 − α(k + 1)/k). If k - 1, this becomes N (1 − α) again. In between, there is a
value of k that, for each fixed α, gives a minimum loss (which may be a gain if α is close to 1).
Note how informative it is to have k rather than 2.8; the numerical value was chosen in the question to
model roughly that lottery system in the UK.
Having gone back to this solution after a break, I am now wondering about the use of the Poisson
approximation. Of course, the set-up (large m, small p) begs us to approximate, but did we need to?
Certainly not for the first result, since we can just as well look at the ratio of two terms of the Binomial
distribution as at two terms of the Poisson distribution. Try it; the result is of course the same.
The second part is more difficult. What we want is the expectation of 1/(n + 1), and this turns out
to be a difficult sum using the Binomial distribution (in fact, it can only be expressed in terms of a
hypergeometric function, which would then have to be approximated to get a less obscure answer).
It seems to me that the solution above is therefore a bit unsatisfactory. It would surely have been better
to work with the exact distribution until it was necessary to approximate, even though one knows that
the approximation is so good that the answers would be the same. We are, after all, mathematicians and
not engineers.
158 Advanced Problems in Mathematics

Problem 70: Bodies in the fridge (!!)

My two friends, who shall remain nameless, but whom I shall refer to as P and Q, both told me
this afternoon that there is a body in my fridge. I’m not sure what to make of this, because P
tells the truth with a probability of p, while Q (independently) tells the truth with a probability
of only q. I haven’t looked in the fridge for some time, so if you had asked me this morning, I
would have said that there was just as likely to be a body in the fridge as not. Clearly, in view
of what my friends have told me, I must revise this estimate. Explain carefully why my new
estimate of the probability of there being a body in the fridge should be
pq
.
1 − p − q + 2pq

I have now been to look in the fridge and there is indeed a body in it; perhaps more than one. It
seems to me that only my enemy E1 or my other enemy E2 or (with a bit of luck) both E1 and
E2 could be in my fridge, and this evening I would have judged these three possibilities equally
likely. But tonight I asked P and Q separately whether E1 was in the fridge, and they each said
that she was. What should be my new estimate of the probability that both E1 and E2 are in my
fridge?
Of course, I always tell the truth.

1987 Paper II

Comments
The most difficult part of this problem is unravelling the narrative! The first paragraph says essentially
‘what is the probability that there is body in the fridge, given that P and Q both say there is?’. It
can therefore be tackled by the usual methods of conditional probability: tree diagrams, for example,
or Bayes’ theorem. All the other words in the first paragraph are there to tell you about the a priori
probabilities of the events, without knowledge of which the question above would be meaningless.
In the second paragraph, the situation becomes more complicated, but the method used for the first
paragraph will still work.
In case you want to use it, here is the statement of Bayes’ theorem, in its simplest form:

P(B) × P(A|B)
P(B|A) = .
P(A)
 Stephen Siklos 159

Solution to problem 70
This problem can be solved using tree diagrams. A more sophisticated, but not necessarily better,
method is to use Bayes’ theorem.
Here, we take the events A and B to be
A = P and Q both say that there is a body in the fridge
B = there is a body in the fridge

From the information given in the question, P (B) = 12 , so

1
2× pq
P(B|A) = .
P(A)
Now
P(A) = P(there is a body) × P(P and Q both say there is)
+P(there is not a body) × P(they both say there is)
1 1
= 2 × pq + 2 × (1 − p)(1 − q)
which gives the required answer.
For the second paragraph, let
X = P and Q both say that E1 is in the fridge
Y = E1 and E2 are in the fridge
There are three possibilities (since we know that there is at least one body in the fridge): only E1 is in the
fridge; only E2 is in the fridge; and both E1 and E2 are in the fridge. These are given as equally likely,
so the a priori probabilities are each 13 .
We want P(Y |X), which by Bayes’ theorem is
1
P(Y ) × P(X|Y ) × pq
= 3 .
P(X) P(X)
Now
P(X) = P(only E1 is in the fridge) × P(P and Q told the truth)
+P(both E1 and E2 are in the fridge) × P(P and Q both told the truth )
+P(only E2 is in the fridge) × P(P and Q both lied )
1 1 1
= 3 × pq + 3 × pq + 3 × (1 − p)(1 − q)

pq
so the answer is .
1 − p − q + 3pq

Post-mortem
Note how much more difficult it is when the answer is not given; when the question was originally set,
most candidates arrived at the given answer to the first part but were not sufficiently confident to extend
their method to the second paragraph: they received 20 8
for their efforts.
The last line of the question is not entirely frivolous; if I may have lied about what my friends answered
when I asked them if there is a body in the fridge, the problem becomes difficult. However, my claim to
be truthful is vacuous (it tells you nothing) because I may be lying. Contrast with the statement ‘I am
lying’, which is inconsistent.
160 Advanced Problems in Mathematics

Problem 71: Choosing keys (!)

I have k different keys on my key ring. When I come home at night I try one key after another
until I find the key that fits my front door. What is the probability that I find the correct key, for
the first time, on the nth attempt in each of the following three cases?

(i) At each attempt, I choose a key that I have not tried before, each choice being equally likely.

(ii) At each attempt, I choose a key from all my keys, each of the k choices being equally likely.

(iii) At the first attempt, I choose from all my keys, each of the k choices being equally likely. At
each subsequent attempt, I choose from the keys that I did not try at the previous attempt,
each of the k − 1 choices being equally likely.

2000 Paper II

Comments
This is very easy, and would really be far too easy if the answers were given.
You should set out your argument clearly and concisely, because if you come up with the wrong answer
and an inadequate explanation you will not get many marks. Even if you write down the correct answer
you may not get the marks if your explanation is inadequate.
You should of course run the usual checks on your answers. Do they lie in the range 0 " p " 1? If you
sum over all outcomes, do you get 1? (The latter is a good and not difficult exercise; I insist that you try
it.)
You will probably be struck by the simplicity of the answer to part (i), after simplification. Is there an
easy way of thinking about it (or, if you did it the easy way, is there a hard way)?
 Stephen Siklos 161

Solution to problem 71

(i)
P(finding correct key, for the first time, on nth attempt)

= P(fail first, fail second, ... , fail (n − 1)th, succeed nth)


 1

 for n = 1;

 k


= k−1 k−2 k − (n − 1) 1 1
 × × ··· × × = for 2 " n " k;

 k k − 1 k − (n − 2) k − (n − 1) k




0 for n > k.

(ii)
P(find correct key, for the first time, on nth attempt)

" #n−1
k−1 1
= P(fail first, fail second, ... , succeed nth) = .
k k

(iii)
P(find correct key, for the first time, on nth attempt)

= P(fail first, fail second, ... , succeed nth)


 1

 for n = 1;
 k
= " #" #n−2

 k−1 k−2 1

 for n # 2.
k k−1 k−1

Post-mortem
The above solution for part (i) is the ‘hard way’. The easy way is to consider the keys all laid out in a
row, instead of being picked sequentially. Exactly one of the keys is the correct one, which is equally
likely to be any of the keys in the row, so has a probability of 1/k of being in any given position. Clearly,
this corresponds exactly to picking them one by one.
I gave the hard solution above, because I think most people will be drawn into doing it this way by the
phrasing of the question: keys are picked one by one and tried before going on to the next.
Did you check that your probabilities sum to 1? It is obvious for part (i), but you have to sum geometric
progressions for parts (ii) and (iii).
162 Advanced Problems in Mathematics

Problem 72: Commuting by train (!!)

Tabulated values of Φ(·), the cumulative distribution function of a standard normal variable, should not
be used in this question.
Henry the commuter lives in Cambridge and his working day starts at his office in London at
0900. He catches the 0715 train to King’s Cross with probability p, or the 0720 to Liverpool Street
with probability 1 − p. Measured in minutes, journey times for the first train are N (55, 25) and
for the second are N (65, 16). Journey times from King’s Cross and Liverpool Street to his office
are N (30, 144) and N (25, 9), respectively. Show that Henry is more likely to be late for work if
he catches the first train.
Henry makes M journeys, where M is large. Writing A for 1 − Φ( 20 13 ) and B for 1 − Φ(2), find, in
terms of A, B, M and p, the expected number, L, of times that Henry will be late and show that,
for all possible values of p,
BM " L " AM.

Henry noted that in 35 of the occasions when he was late, he had caught the King’s Cross train.
Obtain an estimate of p in terms of A and B.
+ ,
Note: A random variable is said to be N µ, σ 2 if it has a normal distribution with mean µ and
variance σ 2 .

2000 Paper II

Comments
This is impossible unless you know the following result:

If X1 and X2 are independent and normally distributed according to X1 ∼ N (µ1 , σ12 ) and X2 ∼ N (µ2 , σ22 ),
then X1 + X2 is also normally distributed and X1 + X2 ∼ N (µ1 + µ2 , σ12 + σ22 ).

Even if you didn’t know this result, you do now and it should not be difficult to complete the first parts
question.
For the last part, you need to know something about conditional probability, namely that

4 P(B ∩ A)
P(A4B) =
P(B)

which makes sense intuitively and can easily be understood in terms of Venn diagrams. The denomi-
nator is essentially a normalising constant. The formula may be taken as the definition of conditional
probability on the left hand side.
 Stephen Siklos 163

Solution to problem 72
Let T1 be the random variable representing the total journey time via Kings Cross, so that

T1 ∼ N (55 + 30, 25 + 144) = N (85, 169),

using the result mentioned on the previous page, and and let T2 be the random variable representing
the total journey time via Liverpool Street, so that

T2 ∼ N (65 + 25, 16 + 9) = N (90, 25).

Then the probabilities of being late are, respectively, P(T1 > 105) and P(T2 > 100), i.e. 1 − Φ( 20
13 ) and
1 − Φ(2). Note that Φ( 20
13 ) < Φ(2).
We have
L = [pA + (1 − p)B]M = [B + (A − B)p]M.

L increases as p increases, since A > B, hence the given inequalities corresponding to p = 0 and p = 1.
We have
P(Late and Kings Cross)
P(Kings Cross given late) =
P(Late)
An estimate for p (all it p̃) is therefore given by

3 Ap̃
= ,
5 Ap̃ + B(1 − p̃)

3B
so p̃ = .
2A + 3B

Post-mortem
The result mentioned in the comment on the previous page is just the sort of thing you should try to
prove yourself rather than take on trust. Unfortunately, such results in probability tend to be pretty hard
to prove. You can also prove it fairly easily using generating functions (which are not in the syllabus for
STEP I and II). You can also prove it from first principles. That would be difficult for you because of the
awkward integrals involved, but not impossible.
164 Advanced Problems in Mathematics

Problem 73: Collecting voles (!)

A group of biologists attempts to estimate the magnitude, N , of an island population of voles

(Microtus agrestis). Accordingly, the biologists capture a random sample of 200 voles, mark them
and release them. A second random sample of 200 voles is then taken of which 11 are found to
be marked. Show that the probability, pN , of this occurrence is given by
+ ,2
(N − 200)!
pN =k ,
N !(N − 389)!

where k is independent of N .
The biologists then estimate N by calculating the value of N for which pN is a maximum. Find
this estimate.
All unmarked voles in the second sample are marked and then the entire sample is released.
Subsequently a third random sample of 200 voles is taken. Using your estimate for N , write
down the probability that this sample contains exactly j marked voles, leaving your answer in
terms of binomial coefficients.
Deduce that 5 65 6 5 6
200
- 389 3247 3636
= .
j=0
j 200 − j 200

2000 Paper II

Comments
This is really just an exercise in combinations. (Recall that a permutation is a reordering of a set of
objects, and a combination is a selection of a subset from a set.) You assume that you are equally likely
to choose any given subset of the same size, so that the probability of a set of specific composition is the
number of ways of choosing a set of that composition divided by the total number of ways of choosing
any set of the same size. Of course, you are assuming that the voles are indistinguishable, except for the
marks made by the biologists.
Maximising a discrete (not a continuous) function of N came up on one of the previous questions: you
have to compare adjacent terms.
The numbers look rather bad, though they turn out OK. My instinct would be to do it algebraically
first: replace 200 by a and 11 by b, then substitute back at the end. I am sure it will lead to a better
understanding of what is going on.
 Stephen Siklos 165

Solution to problem 73
For the second sample, 200 out of N voles are already marked, so pN is just the number of ways of
choosing 11 from 200 and 189 from N − 200 divided by the number of ways of choosing 200 from N :
" #" #
200 N − 200 200! (N − 200)!
11 189 11! 189! 189! (N − 389)!
pN = " # = i,
N N!
200 200! (N − 200)!
so
(200!)2
k= .
11! (189!)2
At the maximum value, pN ≈ pN −1 , i.e.

(N − 200)2
≈ 1,
N (N − 389)

which gives N ≈ 2002 /11 ≈ 3636 (just divide 40000 by 11).

At the third sample, there are 389 marked voles and an estimated 3636 − 389 = 3247 unmarked voles.
Hence 5 65 6
389 3247
j 200 − j
P(exactly j marked voles) = 5 6 .
3636
200
The final part follows immediately, using
200
-
P(exactly j marked voles) = 1 .
j=0

Post-mortem
Adding the Latin name of the species was a nice touch, I thought (not my idea); it adds an air of verisimil-
itude to the problem.
I suppose that this might be the basis of a method of estimating the population of voles — rather clever
really. I don’t know how well it works in practice though. The assumption mentioned earlier, that
picking one set of voles of size 200 is just as likely as picking any other set, surely relies on perfect
mixing of the marked voles, which would be rather difficult to achieve (especially as female voles can
be highly territorial).
You might be asking yourself why it was OK to find the maximum by setting pN ≈ pN −1 . This is
a standard method, but of course it only works if the distribution is one-humped, like a normal dis-
tribution. An alternative approach would have been to approximate the distributions using Stirling’s
approximation, which at its most basic is

ln N ! ≈ N ln N − N .

This gives exactly the same equation as the pN ≈ pN −1 method, and shows that the distribution is
indeed one-humped.
166 Advanced Problems in Mathematics

Problem 74: Breaking a stick (!!)

A stick is broken at a point, chosen at random, along its length. Find the probability that the
ratio, R, of the length of the shorter piece to the length of the longer piece is less than r, where r
is a given positive number.
Find the probability density function for R, and calculate the mean and variance of R.

1999 Paper II

Comments
Continuous probability distributions often seem harder than discrete distributions, for no good reason:
the concepts are the same, and in fact integrals are normally easier than sums.
Here, part of the difficulty is that you have to set up the problem yourself. You have a random vari-
able with a known distribution (corresponding to the point at which the stick is broken), but you are
interested in another random variable (the ratio of lengths) derived from the first. As always, it is best
to work with the cumulative distribution functions rather than the probability density functions when
deriving the distribution of the second random variable.
 Stephen Siklos 167

Solution to problem 74
Let the length of the stick be 2' and let X be the length of the shorter piece of stick, so that X " '.
The random variable X is uniformly distributed on the interval 0 " x " ', so
x
P(0 " X " x) = .
'
X
Now R = , by definition, so
2' − X
2'R
X= .
1+R
The cumulative distribution function for R is given by
" # " #
X 2'r 2'r 2r
P(R " r) = P "r =P X" = = .
2' − X 1+r (1 + r)' 1+r

Let’s check that this satisfies the conditions for a cumulative distribution function: it should increase
from 0 to 1 as r goes from its smallest value, which is 0 to its greatest value, which is 1. And it does, so
that’s OK.
The probability density function is the derivative of the cumulative distribution function:
" #
d 2r 2
= .
dr 1 + r (r + 1)2

Integrating to find E(R) and E(R2 ) gives

) 1 ) 1
2r 2 2r2
E(R) = dr = 2 ln 2 − 1 ; E(R ) = dr = 3 − 4 ln 2 ,
0 (1 + r)2 0 (1 + r)2

and Var (R) = (3 − 4 ln 2) − (2 ln 2 − 1)2 = 2 − 4(ln 2)2 .

Post-mortem
Note the check we made to verify that the function we found for the cumulative distribution could actu-
ally be a cumulative distribution function, which is equivalent to checking that the probability density
function integrates to 1.
We should also check that the final answers make sense. Putting numbers into a calculator gives 2 ln 2 −
1 = 0.23 for the expected value, which seems reasonable; at least it is less than 1. If you sketch the
density function, you see that this could easily be its average value (do it!).
For the variance, I got 0.08, which means that the ratio is likely to be in the plausible range 0.15 to 0.31.
There are a variety of other stick-breaking problems, including a rather pleasing one about breaking the
stick in two places and finding the probability that the pieces form a triangle. In this case, it matters
how you break the stick: you could break it once and then break one of the pieces; or you could choose
two points on the stick randomly at which to break it. For an interesting discussion, see the excellent
cut-the-knot web site: [Link]
168 Advanced Problems in Mathematics

Problem 75: Random quadratics (!!!)

The random variable B is normally distributed with mean zero and unit variance. Find the
probability that the quadratic equation

X 2 + 2BX + 1 = 0

has real roots.

Given that the two roots X1 and X2 are real, find, giving your answers to three significant figures:

(i) the probability that both X1 and X2 are greater than 15 ;

(ii) the expected value of |X1 + X2 |.

1988 Paper II

Comments
It is quite difficult to find statistics questions at this level that are not too difficult and are also not
simple applications of standard methods. For example, χ2 tests are not really suitable, because the
theory is sophisticated while the applications are usually rather straightforward. Most questions in the
probability/statistics area tend therefore to concentrate on probability, and many of these have a bit of
pure mathematics thrown in.
Here, the random variable is the coefficient of a quadratic equation, which is rather pleasing. But you
have to handle the inequalities carefully. The difficulty is increased by the conditional element: for parts
(i) and (ii) you are only interested in the case of real roots.
If you don’t have statistical tables handy, don’t bother to find some: just leave the answers in terms of
the probability, Φ(z), that a standard (µ = 0, σ = 1) normally distributed random variable Z satisfies
Z " z.
 Stephen Siklos 169

Solution to problem 75
The solution of the quadratic is *
X = −B ± B2 − 1

which has real roots if |B| # 1. Let Φ(z) be the probability that a standard normally distributed vari-
able Z satisfies Z " z. Then the probability that |B| # 1 is (taking the two tails of the normal distribution)
+ ,
Φ(−1) + 1 − Φ(1) = 2 − 2 Φ(1) = 0.3174.
√
(i) We need the smaller root to be greater than 15 . The smaller root is −B − B 2 − 1. Now provided
√
B 2 − 1 is real, we have
* *
−B − B 2 − 1 > 15 ⇔ B + 15 < − B 2 − 1
⇔ (B + 15 )2 > B 2 − 1 and (B + 15 ) < 0
⇔ 2
5B + 1
25 > −1 and B < − 15
⇔ − 15 > B > − 13
5 .
√
However, if B < − 15 , then the condition that B 2 − 1 is real, i.e. |B| # 1, implies the stronger condition
B " −1. The condition that both roots are real and greater than 15 is therefore

5 < B " −1
− 13

and the probability that both roots are real and greater than 1
5 is

Φ(−1) − Φ(−2.6) = Φ(2.6) − Φ(1) = 0.9953 − 0.8413 = 0.1540.

The conditional probability that both roots are greater than 15 given that they are real is
+ ,
+ 1 4
4 , P both roots > 15 and both roots real
P both roots > 5 both real = + ,
P both roots real
0.1540
= = 0.485.
0.3174

(ii) The sum of the roots is −2B, so we want the expectation of |2B| given that |B| # 1, which is
) −1 ) ∞
1 1 2 1 1 2 1 1
√ (−2x)e− 2 x dx + √ 2xe− 2 x dx 2 × √ × 2e− 2
2π −∞ 2π 1 2π
) −1 ) ∞ = + , = 3.05.
1 1 2 1 1 2
√ e− 2 x dx + √ e− 2 x dx 2 1 − Φ(1)
2π −∞ 2π 1
 Syllabus

The mathematical requirements for this book are based on the syllabus for STEP Mathematics I and II.
The syllabus listed below serves as a rough guide. Some of the questions in the book require no knowl-
edge of the syllabus; and some cover material that is not included in the syllabus, but is introduced
in the question itself. You can find more information about the STEP examinations from the web site
[Link] .

PURE MATHEMATICS

Specification Notes

General
Mathematical vocabulary and notation including: equivalent to; necessary and sufficient; if
and only if; ⇒ ; ⇔ ; ≡ .
Methods of proof including proof by contradiction and disproof by
counterexample; proof by induction.
Algebra
Indices and surds including rationalising denominators.
Quadratics including proving positivity by completing a square.
The expansion for (a + b)n including knowledge of the general term;
/n/ n!
notation: = .
r r! (n − r)!
Algebraic operations on polynomials and including factorisation, the factor theorem, the
rational functions remainder theorem;
including understanding that, for example, if

x3 + bx2 + cx + d ≡ (x − α)(x − β)(x − γ),

then d = −αβγ.
Partial fractions including denominators with a repeated or quadratic
factor.
Sequences and series including use of, for example, an+1 = f (an ) or
an+1 = f (an , an−1 );
including understanding of the terms convergent,
divergent and periodic in simple cases;
7n
including use of k=1 k to obtain related sums.
The binomial series for (1 + x)k , where k is a including understanding of the condition |x| < 1 .
rational number
Arithmetic and geometric series including sums to infinity and conditions for
convergence, where appropriate.
 Stephen Siklos 171

Inequalities including solution of, eg, a−x

1 x
> x−b ;
including simple inequalities involving the modulus
function;
including the solution of simultaneous inequalities
by graphical means.
Functions
Domain, range, composition, inverse including use of functional notation such as
+ ,
y = f (ax + b) , x = f −1 (y) and z = f g(x) .
Increasing and decreasing functions the precise definition of these terms will be included
in the question.
Exponentials and logarithms including x = ay ⇔ y = loga x , x = ey ⇔ y = ln x ;
including the exponential series
The effect of simple transformations such as y = af (bx + c) + d.
The modulus function.
Location of roots of f (x) = 0 by considering
changes of sign of f (x)
Approximate solution of equations using
simple iterative methods
Curve sketching
General curve sketching including use of symmetry, transformations,
behaviour as x → ±∞, points or regions where the
function is undefined, turning points, asymptotes
parallel to the axes.
Trigonometry
Radian measure, arc length of a circle, area of
a segment
Trigonometric functions including knowledge of standard values, such as
tan( 14 π) , sin 30◦ ;
including identities such as sec2 φ − tan2 φ = 1 ;
including application to geometric problems in two
and three dimensions.
Double angle formulae including their use in calculating, eg, tan( 18 π) .
Formulae for sin(A ± B) and cos(A ± B) including their use in solving equations such as
a cos θ + b sin θ = c.
Inverse trigonometric functions definitions including domains and ranges;
notation: either arctan θ or tan−1 θ, etc.
Coordinate geometry
Straight lines in two-dimensions including the equation of a line through two given
points, or through a given point and parallel to a
given line or through a given point and
perpendicular to a given line;
including finding a point which divides a segment in
a given ratio.
Circles using the general form (x − a)2 + (y − b)2 = R2 ;
including points of intersection of circles and lines.
172 Advanced Problems in Mathematics

Cartesian and parametric equations of curves

and conversion between the two forms.
Calculus
Interpretation of a derivative as a limit and as including knowledge of both notations f ! (x) and dy
dx .
a rate of change
Differentiation of standard functions including algebraic expressions, trigonometric and
inverse trigonometric functions, exponential and log
functions.
Differentiation of composite functions,
products and quotients and functions defined
implicitly.
Higher derivatives including knowledge of both notations f !! (x) and
d2 y
dx2 ;
dn y
including knowledge of the notation dx n .

Applications of differentiation to gradients, including finding maxima and minima which are not
tangents and normals, stationary points, stationary points;
increasing and decreasing functions including classification of stationary points using the
second derivative.
Integration as reverse of differentiation
Integral as area under a curve including area between two curves;
including approximation of integral by the rectangle
and trapezium rules.
Volume within a surface of revolution rotation about either z or y axes.
8
Knowledge and use of standard integrals including the forms f ! (g(x))g! (x) dx and
8 !
f (x)/f (x) dx ;
including transformation of an integrand into
standard (or some given) form;
including use of partial fractions;
not including knowledge of integrals involving
inverse trigonometric functions.
Definite integrals including calculation, without justification, of simple
8∞ 81 1
improper integrals such as 0 e−x dx and 0 x− 2 dx
(if required, information such as the behaviour of
xe−x as x → ∞ or of x ln x as x → 0 will be given).
Integration by parts and by substitution including understanding their relationship with
differentiation of product and of a composite
function;
8
including application to (e.g.) ln x dx .
Formulation and solution of differential formulation of first order equations;
equations solution in the case of a separable equation or by
some other method given in the question.
Vectors
Vectors in two and three dimensions including use of column vector and i, j, k notation.
Magnitude of a vector including the idea of a unit vector.
Vector addition and multiplication by scalars including geometrical interpretations.
Position vectors including application to geometrical problems.
 Stephen Siklos 173

The distance between two points

Vector equations of lines including the finding the intersection of two lines;
understanding the notion of skew lines (knowledge
of shortest distance between skew lines is not
required).
The scalar product including its use for calculating the angle between
two vectors.

MECHANICS

Questions on mechanics may involve any of the material in the Pure Mathematics syllabus.

Specification Notes
Force as a vector including resultant of several forces acting at a point
and the triangle or polygon of forces;
including equilibrium of a particle;
forces include weight, reaction, tension and friction.
Centre of mass including obtaining the centre of mass of a system of
particles, of a simple uniform rigid body (possibly
composite) and, in simple cases, of non-uniform
body by integration.
Equilibrium of a rigid body or several rigid including use of moment of a force;
bodies in contact for example, a ladder leaning against a wall or on a
movable cylinder;
including investigation of whether equilibrium is
broken by sliding, toppling or rolling;
including use of Newton’s third law;
excluding questions involving frameworks.
Kinematics of a particle in a plane including the case when velocity or acceleration
depends on time (but excluding knowledge of
acceleration in the form v dx
dv
);
questions may involve the distance between two
moving particles, but detailed knowledge of relative
velocity is not required.
Energy (kinetic and potential), work and including application of the principle of conservation
power of energy.
Collisions of particles including conservation of momentum, conservation
of energy (when appropriate);
coefficient of restitution, Newton’s experimental law;
including simple cases of oblique impact (on a plane,
for example);
including knowledge of the terms perfectly elastic
(e = 1) and inelastic (e = 0);
questions involving successive impacts may be set.
Newton’s first and second laws of motion including motion of a particle in two and three
dimensions and motion of connected particles, such
as trains, or particles connected by means of pulleys.
174 Advanced Problems in Mathematics

Motion of a projectile under gravity including manipulation of the equation

gx2
y = x tan α − ,
2V cos2 α
2

viewed, possibly, as a quadratic in tan α ;

not including projectiles on inclined planes.

PROBABILITY AND STATISTICS

The emphasis is towards probability and formal proofs, and away from data analysis and use of stan-
dard statistical tests. Questions may involve use of any of the material in the Pure Mathematics syllabus.

Specification Notes

Probability
Permutations, combinations and including sampling with and without replacement.
arrangements
Exclusive and complementary events including understanding of
P(A ∪ B) = P(A) + P(B) − P(A ∩ B) , though not
necessarily in this form.
Conditional probability informal applications, such as tree diagrams.

Distributions
Discrete and continuous probability density including calculation of mean, variance, median,
functions and cumulative distribution mode and expectations by explicit summation or
functions integration for a given (possibly unfamiliar)
distribution (eg exponential or geometric or
something similarly straightforward);
notation: f (x) = F! (x).
Binomial distribution including explicit calculation of mean.
Poisson distribution including explicit calculation of mean;
including use as approximation to binomial
distribution where appropriate.
Normal distribution including conversion to the standard normal
distribution by translation and scaling;
including use as approximation to the binomial or
Poisson distributions where appropriate;
notation: X ∼ N(µ, σ 2 ) .
Hypothesis testing
Basic concepts in the case of a simple null including knowledge of the terminology null
hypothesis and simple or compound hypothesis and alternative hypothesis, one and two tailed
alternative tests.
 This book need not end here...

At Open Book Publishers, we are changing the nature of the traditional academic book.
The title you have just read will not be left on a library shelf, but will be accessed online
by hundreds of readers each month across the globe. We make all our books free to read
online so that students, researchers and members of the public who cant afford a printed
edition can still have access to the same ideas as you.
Our digital publishing model also allows us to produce online supplementary material, in-
cluding extra chapters, reviews, links and other digital resources. Find Advanced Problems
in Mathematics on our website to access its online extras. Please check this page regularly
for ongoing updates, and join the conversation by leaving your own comments:

[Link]

If you enjoyed this book, and feel that research like this should be available to all readers,
regardless of their income, please think about donating to us. Our company is run entirely
by academics, and our publishing decisions are based on intellectual merit and public value
rather than on commercial viability. We do not operate for profit and all donations, as with
all other revenue we generate, will be used to finance new Open Access publications.
For further information about what we do, how to donate to OBP, additional digital mate-
rial related to our titles or to order our books, please visit our website:
[Link]
 You may also be interested in...

[Link]

Teaching Mathematics is nothing less than a mathematical manifesto. Arising in response to

a limited National Curriculum, and engaged with secondary schooling for those aged 11-14
(Key Stage 3) in particular, this handbook for teachers will help them broaden and enrich
their students’ mathematical education. It avoids specifying how to teach, and focuses
instead on the central principles and concepts that need to be borne in mind by all teachers
and textbook authors but which are little appreciated in the UK at present.
This study is aimed at anyone who would like to think more deeply about the discipline
of ‘elementary mathematics’, in England and Wales and anywhere else. By analysing and
supplementing the current curriculum, Teaching Mathematics provides food for thought for
all those involved in school mathematics, whether as aspiring teachers or as experienced
professionals. It challenges us all to reflect upon what it is that makes secondary school
mathematics educationally, culturally, and socially important.

Tony Gardiner, former Reader in Mathematics and Mathematics Education at the University of
Birmingham, was responsible for the foundation of the United Kingdom Mathematics Trust in 1996,
one of the UK’s largest mathematics enrichment programs. In 1997 Gardiner served as President of
the Mathematical Association, and in 2011 was elected Education Secretary of the London Mathe-
matical Society.