Alan Burke wrote:Did I say that the statistics were from every match ever played or if they might just be from one particular game ? So, without you knowing where I got my evidence to make my claim just how can you try and counter it ?
They won't be from every match ever played, or just one game. It'll be a sample, or perhaps all matches in a particular league. You're right that you didn't say that, but it was an obvious assumption that it was neither of the two extremes you mention.
Alan Burke wrote:If in one football game there are 12 corners, the statistics of that match will prove that fact.
The number of corners in a football game is not a statistic, it's data. The mean number of corners in football games in a particular league for a particular season would be an example of a statistic with two parameters (the particular league, and the particular season).
Alan Burke wrote:A few years ago many, many videos were studied and analysed to discover just how many decisions were correctly made by referees in comparison to those they got wrong - the result proved that on those occasions the referees got well over 90% of decisions correct.
How do I know ? Because I was involved with the experiment.
OK, if the study was done by viewing other matches then the result might be different, but on that occasion the statistics DID prove that wrong decisions were minimal and thus my comment, based on that evidence, was correct.
"OK, if the study was done by viewing other matches then the result might be different" - Exactly. Which is why it hasn't proven anything. You can say, with a certain level of confidence (which you haven't specified), is that referees get well over 90% of decisions correct.
This is not the same as saying that all referees always get well over 90% of decisions correct. You would need to look at every example in the population, and at that point, you're no longer dealing with statistics, and you enter the realm of population parameters.
The whole point of statistics is that you're using a sample to represent a population, because the latter is impossible to consider on practical grounds. Which is why statistics can never prove anything; you can never say anything with 100% confidence.