Basic HTML Version
54
55
the
manager
WINTER 2011
The BACK OFFICE
about Daniel
Finkelstein
Daniel Finkelstein is executive
editor of
The Times
and author
of its weekly Fink Tank column on
football statistics
but either just fail or just succeed to
score in injury time. Obviously, the
more games there are, the more this
element of luck will even itself out.
If the season went on for five years,
the upside would be that you could
have almost complete confidence that
the best side would win. Fans might
not like that, though, so 38 games is a
reasonable compromise, giving a fan-
friendly season length with a tolerable
probability that the best side will win.
But anyway… we’ve been looking at
the stats for a decade and the chance of
a favourite winning in any given season
is rarely higher than 45 per cent.
Then take the FA Cup. As that’s a
knockout, the chances of the best side
winning are reduced. There’s such a
large probability of losing one game
in a series, or of being drawn against
tough competition, that it’s unusual for
a favourite to have more than a 20 per
cent chance of lifting that trophy.
To illustrate the importance of
fortune in a knockout, consider the
chances of a single giant-killing act in
the third round. Although any one of
those may look unlikely in itself, when
you take the chances together it is very
high. In fact, the chance of at least
one side defeating a team in a higher
division in the FA Cup third round is
almost always around 98 per cent.
And what about the Champions
League? It has a large league element
and the knockout games are played
across two legs, guarding against
elimination from one shock defeat.
For both of these reasons, the chances
of the favourite winning should be
reasonably high. Save for this... at the
top end, the gap between the clubs
isn’t great. Very occasionally Barcelona
have enjoyed a 25 per cent chance
of winning the trophy as the season
begins, but no British club has gone
above 20 per cent (at the very most).
And more recently it has been
less than that. Because of their poor
defence Real Madrid have not been
serious Champions League contenders
for years. But now they are, which cuts
the chances of an English winner to
something more like 15 per cent.
I should, incidentally, remind
pundits not to overcomplicate
things. There seems to be some idea
that it makes a big difference in the
Champions League whether you play
home or away first in the knockout
round. Our data work shows that this
makes no difference whatsoever.
So let’s take them all together. The
chances of winning the treble are (at
best) 20 per cent of 20 per cent of 45
per cent. That is 1.8 per cent. So we
should expect it to happen less than
once every 50 years. Naturally, we
don’t know the distribution; it could
happen three times in my lifetime and
not at all in the life of my grandchild.
Indeed, clustering is quite likely
because you might get a very good
team or two that is on a roll. But you
get the idea… it’s rare.
And this assumes that each event is
independent. If winning the FA Cup
cut your chances of a league title, then
you might only see a treble every 60
years. Once in a lifetime.
So it’s worth celebrating when
it happens, I reckon. Whether you
support the team in question or not.
When Manchester United
won the Barclays Premier League, the
FA Cup and the Champions League
back in 1999, I was very excited. And
happy. Which is strange, because
I’m not a United fan. Far from it. It’s
just that I thought the ‘treble’ was
something I might never see.
Think of it this way. A strong
favourite to win the league might
have a 45 per cent chance. This is,
if you consider the number of sides
in with a chance, a reasonably high
figure. The reason is simple… it’s
the number of games. In each match
there is an element of luck (refereeing
decisions, for example, are often
cited). And a side might be pressing
remember manchester united’s historic treble in 1999? daniel finkelstein does... and he’s
got the numbers to prove that it was even more impressive than you may already think
Analysis
thinking about
winning three
competitions in
a single season?
you have a 1.8 per
cent chance