22
Apr 2002
What
is the most common digit?
Daft as it may sound, there is indeed such a thing, and the digit
is 1. It doesn't matter whether you're talking about the
lengths of rivers (in miles, kilometres, cubits or any other measurement),
the populations of cities, the item prices on your shopping receipts or
whatever: in each random set of numbers 1 is the most common of
all digits. This is Benford's Law, which is now understood
to a sophisticated extent, and is used by the United States Inland Revenue
Service and others to check for non-standard (i.e. doctored) distributions
on corporate returns etc.
How
many ways can you arrange things?
For 3 things, you can put them in 1 × 2 × 3 different orders, like this:
- 1-2-3
- 1-3-2
- 2-1-3
- 2-3-1
- 3-1-2
- 3-2-1
The formula is 1 × 2
× 3 × 4 for four objects, and so on. These are factorials, denoted
by an exclamation mark e.g. factorial 4 = 4!
If you're putting individual
letters into addressed envelopes, and you wickedly want to know how many
different ways you can put all the letters into the wrong envelopes, this
is far more involved. Just one example: with 5 letters, there are
44 ways of getting every single one wrong, leaving 75 ways of getting
some of them wrong, and only one way of getting the job right.
How
do you calculate Pi?
Here is a routine in pseudo-code:
Initial Values of
Variables
a = 1; d = 1; y = 1; z = 1;
b = ( 1 / sqrt(2) )
c = 0.25
Loop indefinitely
y = a
a = ( a + b ) / 2
b = sqrt(b × y)
c = c - (z × (a - y)2 )
z = z × 2
display (a + b)2 / (c × 4)
How
many colours are needed for a map?
Four. This has been known since 1878, but was not proved until
1976. The final proof was hundreds of pages long, took 1200
computer-hours of calculations, and is beyond the understanding of all
but a handful of specialist mathematicians.
How
do you construct a magic square?
| 12 |
13 |
1 |
8 |
| 6 |
3 |
15 |
10 |
| 7 |
2 |
14 |
11 |
| 9 |
16 |
4 |
5 |
The rows, columns and diagonals
all add up to 34.
What
is the largest number?
A googol is 10100 which is pretty big, and a googolplex
is 10googol, which is even larger. But the winner
by billions of universes is Graham's Number, which dwarfs both
of these, and for that matter any other quantity ever conceived.
Graham's number is the answer to a problem in Combinatorics (an abstruse
branch of Maths dealing with sets) and although the algorithm for calculating
it can be described, the number itself is so large that it cannot be written
down in any existing notation at all - even using towers of exponents.
Why not? Because there would not be enough material in the universe
to do so. This is a monster of a number.
|
| PREVIOUS
PUZZLES |
The
Nov-Dec puzzle was:
| Find
the two numbers, the sum of which is 10, and the product
of which is 40. Or put another way, what are the values
of x and y when (x+y)=10 and
xy=40? There is no trickery
with bases: 10 = Ten and 40 = Forty. |
The solution is that
one number =
5 + sqrt(-15) and the other number =
5 - sqrt(-15). This should be enough for you, provided
you remember that:
| (a
+ b) * ( a - b ) = (a * a) - (b * b) |
If anyone's still lost, then please email
me, and I'll take you through it.
The Jan-Feb puzzle
asked you to give the next item in this series:
1, 4, 7, 11, 15, 18, 21, 24, 27...
A bit of lateral thinking required here, as it was not really a Maths
puzzle at all. The solution lies in the number of letters in
the English names for these numbers. 73 (SEVENTYTHREE) is the
answer |
| BOTTLE
IMP PARADOX |

I had quite a few emails from people
who did not understand this problem. Here it is.
There is a short story by Robert Louis Stevenson called The
Bottle Imp, which centres on an imp who will grant
its owner their every wish - no strings attached. Ask
for £1bn and a lovestruck Meg Ryan, and it's all yours for keeps.
But if you're still in possession of the imp when you
die, you'll spend an eternity in Hell. Other thing is that the
only way to renounce ownership of the imp is to sell it on to someone
for less than you paid for it, so the question is:
What is the
minimum price that it is sensible to pay for the imp?
Right, so you're not
going to buy it for 1p, since you can never get rid of it. Nor
will you buy it for 2p, since no-one is going to buy it off you for
1p. But if it's illogcal to buy it for 2p, then it must
be illogical to buy it for 3p, since no-one will buy it for 2p.
See
where we're going? In a world of logicians there would
simply be no price at all that it would be sensible to pay for the
imp. Now in practice there will indeed be a price at which
you might reckon people will not follow their logic through to the
end. But do it at your own risk!! |
| YOU
MISSED YOUR CHANCE! |
| Until
15 Mar 2002 Faber & Faber were offering a nice little earner ($1million)
to anyone who could prove Goldbach's Conjecture:
In 1742, Christian
Goldbach speculated that every even number greater than 2 could
be expressed as the sum of two primes. This has been verified
for all numbers up to about 400 trillion, but it has never actually
been proved. All you needed to do is send Faber & Faber
the proof before 15 Mar 2002, and you'd have been home and dry.
Hard luck.
|
| PI
IN THE SKY |
The
history of Pi is littered with delusions and false starts. In
1853 William Shanks published his calculations of Pi to 707
decimal places. How long that took him by hand you wouldn't
want to think about, but they had determination in those days.
In 1945, it was discovered that Shanks had made an error in his calculation
of the 528th place, thus making his last 180 digits wrong. Mercifully,
he'd been dead for 63 years.
More fun and
games in 1857, when Edwin J Goodwin inspired the introduction
of a bill in the House of Representatives, to enshrine in American
law a supposedly exact value of Pi. This optimistic bill got
through its first reading, but was then held up on the intervention
of a mathematician who by chance was visiting the House. Its
second reading is still to be held. |
| PRIMES
AGAIN |
| Not
all conjectures have lasted as long as Goldbach's. Pierre
de Fermat (1601-1665) thought he had discovered a formula for
generating primes with:
[2^(2^n)]
+1
where ^ indicates
exponentiation. In 1742, however, Euler discovered this
to be false where n=5. To be fair to Fermat (one of the greatest
mathematicians of all time) this is the only one of his many conjectures
ever to have been proved wrong.
|
| A
BAD CONJECTURE |
| No
such plaudits for de Polignac, who conjectured that every odd
number was the sum of a power of 2 and a prime. De Polignac
claimed to have verified this up to 3 million, and if he did indeed
go to these lengths it was a monumental waste of effort; for
if he had only been more careful when examining the number 127,
for example... |
|