The FairFootball correction scheme
a better guess than the official winner
Our goal is to correct
obvious errors of the referee in enforcing the laws of the game, as defined in the
“Laws of the game 2005” that can be found for instance here:
http://www.fifa.com/en/regulations/regulation/0,1584,3,00.html
However, these
laws abound in phrasings like “considered by the referee to be”, “in the
referee’s view / opinion” and other similar ones. Read strictly, it would mean
that anything the referee decides is right, by the fact itself that he has thus
decided, even if the teams played handball over the pitch without him having anything
to say. Or, for something to be unlawful, it would have to be proven that the
ref in fact thought it was wrong / unfair, though he didn’t say it (in other
words intentional wrongdoing; unintentional mistakes would be lawful). Of course,
we consider this approach to be totally unacceptable, else we wouldn’t be
creating this site. We also think that the laws are badly written in this
respect; any such mention should in our view be removed from them, and if
necessary a single phrase should be added stating that the laws are applied by
the referee as he considers appropriate, that his duty is to enforce them to
the best of his ability but that he is the only one to make all decisions.
Therefore, our
goal is to correct plays according to the “objective” laws, that is with all such
phrasings removed.
We will judge every
action to our best ability, based on video recordings. And, though we welcome
feedback, on our site we will be the referee, that is have the final say over
everything.
We will only
reverse the referee’s decision when we think it was definitely wrong based on
the video material. Whenever there is reasonable doubt, we will give credit to
the referee. This includes the case when he has made other errors in the same
game or in others, regardless of their number and gravity.
We will correct
based on any video material at our disposal, not just the official TV covering.
We will use a video material only when we will be reasonably sure that it is
authentic and accurate.
Rules for correcting a game’s score:
These first 2
are pretty obvious.
These points can
be thought of as being too simplistic. Indeed, one could argue that the chance
of a penalty being scored depends on the player who takes it and on the goalkeeper
who defends it; that the chance of a free kick being scored varies smoothly with
the distance and the angle, as well as with many other factors (2 important
ones being again the player who delivers it and the keeper who defends it);
that the un-granted penalties, free kicks and stopped actions should be
combined and not treated separately, and possibly other objections. The problem
is that it would be enormously difficult to take all these points appropriately
into consideration. Therefore, we have chosen to simplify things in a uniform
manner.
Establishing the final ranking based on the corrected scores:
This wouldn’t
be an issue in complete tournaments like national leagues, where all teams face
each other. But in eliminatory competitions like the Champions League or the
World Championship, a basic problem appears: if the corrected scores of a round
result in different teams to qualify than the official ones, these teams will
not be allowed to play in the next rounds instead of the wrongly qualified ones.
Of course, there is no 100% right way to correct this; one might think there is no way that is reasonably OK. However, we think there is a way that is more correct than ranking based on obvious referee blunders. Therefore:
Thus, if
officially teams A and B qualify in this order, but upon correction the order
is B first then A, the order will be considered unchanged. If officially teams
A and B qualify in this order but upon correction the order is C first then A, the
considered order will be A first then C (instead of B), that is B will
represent C as defined above.
Thus, take
the case of the World Cup 2002, when Korea first beat Italy, then Spain, both incorrectly,
then correctly lost to Germany. Korea then represents both Italy and Spain and
should trade places with them. However, first it trades with Spain; thus, after
the first trade, Spain gets a semi-final place and Korea a quarter-final place.
This one it then trades with Italy, thus Italy gets the quarter-final place and
Korea gets the eighth-final place.
The model is
far from perfect. In the above example, Italy could rightly claim that it may
have beaten Spain given the chance; both Italy and Spain can also claim they
would have beaten Germany, which is not only possible but very likely.
Unfortunately, those games were in fact never played; this is the most that can
be done based only on games that have been played and on their corrected
results. It’s the limitation of the model. But we believe a final ranking with
Spain semi-finalist, Italy quarter-finalist and Korea eighth-finalist is definitely
more correct than the official one.
Some heated debates about the validity of our approach and about our style of refereeing
can be found here:
Google groups: rec.sport.soccer: World Cup 2006 refereeing
Google groups: rec.sport.soccer: Is Italy win tainted ?