Var debate 2019/20

Marvin said:


You have to take into account the distortion introduced on a 2D image when the camera is not in line with play. I don't think anyone can just look at that and say whether it is onside or not.

You need some kind of algorithm to adjust for the angle of view. I am sure that this is possible to do, but there's no indication that they are doing this. You can't just guess.

It's ridiculous that we are now into October and the science of the offside measurement is still unknown.

And even if they can measure it with a very small uncertainty budget, it's still a problem unless it's real time.
Click to expand...


This is it for me: excuse me being a wordy twat, but we are now talking about epistemology and ontology and statistical significance and camera frame rates for a law designed to stop people goal hanging.
 
Brightwell Bros. said:
This is it for me: excuse me being a wordy twat, but we are now talking about epistemology and ontology and statistical significance and camera frame rates for a law designed to stop people goal hanging.
Click to expand...
You have to, because they are making measurements to determine whether a player is offside.

All measurements are made up to a certain level of precision. What is the precision of VAR? This is a very legitimate question to ask. The rule is that you are offside or onside, and that you can not be level. OK, that's clear, but it only works if you can make that measurement very precisely.

They need to tell the football world what that precision is in order for us to accept it.

When you make a measurement is made, you report its maximum likelihood and a confidence interval within which you can say, with say 95% probability, that the true value lies within those boundaries. From what I understand those boundaries maybe +/- several cm, and not the mm which are really what you need to make VAR work within the existing rules.

This is just a general comment on what is required to assess VAR in relation to offside. There is still the secondary issue of whether football really wants to stop the flow of the game to do this.

The ideal solution is something that is

a) precise and,
b) fast

I believe VAR is neither.

The graph below is an image I took from the Internet to demonstrate a maximum likelihood estimation and a confidence interval which enables someone to say with a confidence level of X percent (typically 95%) that my estimate of the true value of Y is Z +/- a.

a is your uncertainty. What is VAR's uncertainty? I suspect it is unworkable.

Where is the independent study of VAR technology? They've been trialing this for years. but have they analysed it properly? Maybe it has been detailed but its significance has been lost on football people?

Source: https://datumorphism.com/wiki/statistical-estimation/confidence-interval/
gaussian-alpha.png
 
Last edited:
Marvin said:
You have to, because they are making measurements to determine whether a player is offside.

All measurements are made up to a certain level of precision. What is the precision of VAR? This is a very legitimate question to ask. The rule is that you are offside or onside, and that you can not be level. OK, that's clear, but it only works if you can make that measurement very precisely.

They need to tell the football world what that precision is in order for us to accept it.

When you make a measurement is made, you report its maximum likelihood and a confidence interval within which you can say, with say 95% probability, that the true value lies within those boundaries. From what I understand those boundaries maybe +/- several cm, and not the mm which are really what you need to make VAR work within the existing rules.

This is just a general comment on what is required to assess VAR in relation to offside. There is still the secondary issue of whether football really wants to stop the flow of the game to do this.

The ideal solution is something that is

a) precise and,
b) fast

I believe VAR is neither.

The graph below is an image I took from the Internet to demonstrate a maximum likelihood estimation and a confidence interval which enables someone to say with a confidence level of X percent (typically 95%) that my estimate of the true value of Y is Z +/- a.

a is your uncertainty. What is VAR's uncertainty? I suspect it is unworkable.

Source: https://datumorphism.com/wiki/statistical-estimation/confidence-interval/
gaussian-alpha.png
Click to expand...

So what your saying is it’s not worth shite.
 
You must log in or register to reply here.
Unread Posts

Don't have an account? Register now and see fewer ads!

SIGN UP
Back
Top
✖
  AdBlock Detected
Bluemoon relies on advertising to pay our hosting fees. Please support the site by disabling your ad blocking software to help keep the forum sustainable. Thanks.