Xg from WBA game

That wouldn't be an xG of 0.9 because of the position of the defenders.
In my example the position of the defenders isn't taken into account by xg. Some do some don't. But that's beside the point. Choose another variable. The point remains that you are using a very small amount of data for a model that requires a very large amount of data.
 
Must be the late chances.

I expect the blatant foul on Dieng was counted as a chance for them.
In Boro’s last 2 games we should and would have been playing against 10 men from early first half if the referees concerned had any bottle.
 
As someone mentioned earlier - the use of statistics is so much more complex than - this number means that...

So I am an intensive care doctor so will try and use my area of knowledge to explain..

In 2000 a review of deaths in the UK found that there was probably points in the preceeding 24 hours that this may have been noticed. So a move was made for improvement in management of deteriorating patients.

Early warning score systems are now used in hospital. They take a set of observations, give it a score, and if you trigger a certain number you get a review.

They are imperfect because they use a limited number of variables and do not take into account prior events.

But... the concept of - this is out of the ordinary let's look at it is similar to xG.

If patients kept dying on a ward but the warning scores were never raised - it may be statistical variance but you probably want to review it.

Long term the more data you can put into these things the better predictions you can make. What you can then do is analysis on the output (goals, xG, deaths...) by changing variables within context.

Most sports now use stats and full teams analysing them. Football is a hugely complex system and is atypical because it often chooses to follow "those that know it" rather than a mix of that and external eyes.

It is why I think England should appoint Serena Wiegman - her lack of knowledge of men's football in my eyes is unimportant because she has proven she can take a set of observable data points and create a successful tactic specifically in the role of a tournament manager.
 
TL:DR -
Stats are imperfect but useful tools to monitor trends.

They should be analysed in context.

They are getting better all the time because analysis can be more complex with modern computer algorithms.
 
Back
Top