10cm "wang buffer"

[Peter Grove]

I think that's Tr*** administration-level layered misreporting

I was also surprised to see a claim that the error could be so large so I looked for the source - it seems to be the following article:

Sterling 2cm offside allowed for 13 cm error

Yes, I know - it's the Daily Mail but that doesn't necessarily mean it's incorrect.

The article goes on to explain that while this particular decision had a 13cm margin of error, based on a time lapse of 0.02 secs between frames and Sterling's running speed of 23.4 km/h, if you took the maximum running speed ever recorded in the Premier League, the error could theoretically be as much as 38.8cm. The average obviously would be much lower.

I'm no mathematician though, so if the technical arguments presented there are wrong, perhaps someone more gifted than me could point out why and how.

 

[one]

I have presented that same argument here a couple times now, the first time in 2018. The maximum margin of error can actually be much bigger if you consider a defender running in the opposite direction to put the attacker offside (relative speed).

The only technology that came close to being fit for purpose was the one used in the world cup where they had multiple high speed, high resolution cameras in each game and very high speed ones dedicated the offside line for VAR.

 

[Sheffields Finest]

Getting wanged off and paying £9k for the pleasure... there’s a headline!

 

[socal lurker]

What do you think are the changes since then that make it so much harder to understand?

I agree that most of the complications are on the less common flavors of OS, particularly interference with an opponent, which has been shrunk down in scope. Where I see the biggest increased complication is more in the realm of interpretation than the Law itself. The line between a defender playing the ball vs a deflection moved, I think for the worse.

In the US we had been taught the test was "possess and control"--a high level of defense action to wipe out OS; we are very far from that now. It's sort a like it used to be a presumption of a deflection and now its a presumption of play. (Yes, I know that's imprecise, but I think it is conceptually where we went.)

 

[one]

Going on a tangent. In the past the intent to play the ball was considerd but so were other considerations like the speed of the ball, distance, possibility of being able to play it. The words action vs reaction were often used. Nowadays the interpretations I see are pretty much aligned with intent only. If you have intent to play the ball and the ball touches you, you have played it and offside is reset.

 

[Eddie]

At first I shared the majority view on this...what’s the point as everyone will still argue about it being a toe over the line, only the line is now 10cm further back/forward.

Upon consideration I’m all for it.

In it’s current guise players are being ruled offside down to mm’s, when VAR can’t even be certain which frame the ball was played in. Not to mention the even more ridiculous armpit start/end point.

If a players toe is over the 10cm line they’re beyond the margin of error, they can’t argue they’re only a toe offside because they’ll be a toe +10cm offside- in ‘reality’.

If they’re persisting with VAR then they’ve got to change it. The proposed format is better than what we have now IMO. I’d also suggest they change to feet only to stop the guesswork on where a non-visible armpit begins.

 

[one]

If a players toe is over the 10cm line they’re beyond the margin of error, they can’t argue they’re only a toe offside because they’ll be a toe +10cm offside- in ‘reality’.

This will not resolve the problem. It's like putting a band aid on a wound which needs 40 stitches. Where did they get 10cm from? Looks like an arbitrary number to shut everyone up that we are taking margin of error into account so we are now accurate. So in your example saying they are 10cm+ a toe offside can be completely wrong and the actual fact (the reality, not what TV screen shows) could very well be they are 10cm on side because the actual margin of error was -20cm. What the actual margin of error is dependent on many things (see below) and just picking 10cm is as bad as not taking a margin of error into account in the first place.

1. Camera speed (frames per second) - this impacts if we are freeze framing at the exact moment of first contact with the ball
2. Resolution of the camera - This will impact determining the 'end points' of attacker and defenders
3. Bandwidth of image/video transfer - camera to software then to VAR display - same as the two above
4. Resolution of VAR viewing screens - same as 2
5. Accuracy of line drawing - both vertical and parallel lines as well as parallel-ness to goal line (angle)
6. Distance of SLD defender and offending attacker to camera - This impacts resolution as well as accuracy of line drowning. the closer to camera, the more accurate
7. The speed of movement of both attacker rand defender - the faster the move the bigger the margin of error
8. Angle of camera to location of offence

I may have missed a few here. But you see just saying give a margin of error of 10cm is not good enough.

For me, either fork out some money and get high tech equipment and software with negligible proven margin of error (a few mm) with hawk eye type images for 'factual' decisions, or remove all the technology stuff and make it a subjective decision. If it is not obvious by just looking at replay that the AR has made a mistake, stick with the AR call.

 

[Mr Dean]

1. Camera speed (frames per second) - this impacts if we are freeze framing at the exact moment of first contact with the ball
2. Resolution of the camera - This will impact determining the 'end points' of attacker and defenders
3. Bandwidth of image/video transfer - camera to software then to VAR display - same as the two above
4. Resolution of VAR viewing screens - same as 2
5. Accuracy of line drawing - both vertical and parallel lines as well as parallel-ness to goal line (angle)
6. Distance of SLD defender and offending attacker to camera - This impacts resolution as well as accuracy of line drowning. the closer to camera, the more accurate
7. The speed of movement of both attacker rand defender - the faster the move the bigger the margin of error
8. Angle of camera to location of offence

9. How many martinis the VAR has downed.

 

[Eddie]

This will not resolve the problem. It's like putting a band aid on a wound which needs 40 stitches. Where did they get 10cm from? Looks like an arbitrary number to shut everyone up that we are taking margin of error into account so we are now accurate. So in your example saying they are 10cm+ a toe offside can be completely wrong and the actual fact (the reality, not what TV screen shows) could very well be they are 10cm on side because the actual margin of error was -20cm. What the actual margin of error is dependent on many things (see below) and just picking 10cm is as bad as not taking a margin of error into account in the first place.

1. Camera speed (frames per second) - this impacts if we are freeze framing at the exact moment of first contact with the ball
2. Resolution of the camera - This will impact determining the 'end points' of attacker and defenders
3. Bandwidth of image/video transfer - camera to software then to VAR display - same as the two above
4. Resolution of VAR viewing screens - same as 2
5. Accuracy of line drawing - both vertical and parallel lines as well as parallel-ness to goal line (angle)
6. Distance of SLD defender and offending attacker to camera - This impacts resolution as well as accuracy of line drowning. the closer to camera, the more accurate
7. The speed of movement of both attacker rand defender - the faster the move the bigger the margin of error
8. Angle of camera to location of offence

I may have missed a few here. But you see just saying give a margin of error of 10cm is not good enough.

For me, either fork out some money and get high tech equipment and software with negligible proven margin of error (a few mm) with hawk eye type images for 'factual' decisions, or remove all the technology stuff and make it a subjective decision. If it is not obvious by just looking at replay that the AR has made a mistake, stick with the AR call.

No idea where the 10cm came from but it appeared to be mentioned in the PL meeting so it’s safe to say we’re running with that.

I agree they should throw money at getting the best kit they can, IF they’re to persist with calling OS by the mm. They might, but certainly not by next season after everything’s been scrutinised and had the stamp of approval.

As you’ve highlighted they don’t currently have the technology implemented to be calling it by the mm. So they need to do something about it, introducing the margin of error isn’t perfect, my point is it’s better than what we have now.

 

[Peter Grove]

One idea I've had on this, from a while ago now, is that they could maybe use the inherent qualities of the lines they show on the final image. As I understand it, based on the images I've seen as the VAR is doing the analysis, they draw very thin, black lines to make the actual determination then convert those to red and blue lines (one for the defender, one for the attacker) that are displayed to the viewing public. These red and blue lines being thicker, often touch or overlap. My thought was that if the coloured lines touch or overlap at all, the players should be considered essentially level and so there would be no offside.

Given the width of the coloured lines, this would give in effect, something like the 10cm that the PL is calling for without having to change any of the technology currently in use, just how it's applied to the determination of the offside position.

As far as I recall, in all the cases where people have complained about it being too close to call, this method would have resulted in a "not offside" decision.

 

[Peter Grove]

Here's an example I just found in a quick Google search.


In this case, the white forward (Roberto Firmino) was adjudged offside because the front edge of the red line representing the position of his armpit is ahead of the front edge of the blue line for the defender's knee. Whereas using the method I mentioned above, because the lines are touching, this would not be offside.

 

[one]

Good idea in principal but again it shifts the problem to a new area. How wide do you choose the thick line to be? How do you define 'touching lines'? Are we talking pixels here? What is the quality of the image you check if the lines are touching and how far are you zoomed in or out to check it? The edges of the red line are not actually red but shades between red and it's nearby colours.

In your image the lines are overlapping in the inner side. See the outer edges of the lines. If the lines were a bit further apart determining if they are touching becomes fuzzy. We are back to technology fail.

 

[Peter Grove]

Just use the lines as they currently are. If they touch or overlap, the players are judged to be level for the purposes of offside. The image I posted is an image capture of an image capture, which makes it more fuzzy. The original images used in the VAR room would be clearer.

The same issues of pixel size, image quality etc, already apply to the images they use and yet they are still able to make a decision using them. I'm just changing the way that decision is made, while using the same technology, pixel size and quality that they already have.

 

[Peter Grove]

My other thought (and it's already been mentioned) is that they could use the method employed by the MLS. Don't actually draw lines, just let the VAR eyeball it. If he can't see that it's offside, it isn't.

In an interview, Howard Webb, general manager of the Professional Referees Organisation in the US told the Philadelphia Inquirer that the MLS will not use a VAR line drawing technology for offside decisions. Instead they will use the following process:

“The on-field officials make an onside/offside judgment and that will be considered correct unless the VAR identifies, in their opinion from looking at the footage, a clear and obvious error.”

The IFAB have already hinted that they are moving away from the idea that offside should be a totally factual decision and thinking of introducing some sort of "clear and obvious" criteria to it - which I have to say, surprised me but there we are. One way to do that might be to admit that the search for a technically perfect solution hasn't provided "what football wants or expects" and use a deliberately less precise system relying on human judgment.

I know that rather goes against the whole idea of using all this wonderfully advanced modern technology to come to an incontrovertibly correct decision but I think we've realised by now that that may not be possible anyway, due to the limitations already discussed.

 

[santa sangria]

My other thought (and it's already been mentioned) is that they could use the method employed by the MLS. Don't actually draw lines, just let the VAR eyeball it. If he can't see that it's offside, it isn't.

In an interview, Howard Webb, general manager of the Professional Referees Organisation in the US told the Philadelphia Inquirer that the MLS will not use a VAR line drawing technology for offside decisions. Instead they will use the following process:


The IFAB have already hinted that they are moving away from the idea that offside should be a totally factual decision and thinking of introducing some sort of "clear and obvious" criteria to it - which I have to say, surprised me but there we are. One way to do that might be to admit that the search for a technically perfect solution hasn't provided "what football wants or expects" and use a deliberately less precise system relying on human judgment.

I know that rather goes against the whole idea of using all this wonderfully advanced modern technology to come to an incontrovertibly correct decision but I think we've realised by now that that may not be possible anyway, due to the limitations already discussed.

I like it. Ditch the lines.

TBH I would go further and ditch VAR in-game all together.

I still firmly believe a much better solution would be for referees to review match footage after the game and sanction offences or correct sanctions missed in the game. Individual refs or the head ref would explain decisions, explain extra sanctions/decisions. This would humanize refereeing and provide content for the news cycle on off days while preserving the fan experience in the stadium.

It wouldn’t reverse goals/pens/cards in-game but I think it would be much better for football.

 

[cdsman]

Based on last night's 'non-offside', I bet Fulham can't wait to get back into the Premier League!

 

[Big Cat]

My other thought (and it's already been mentioned) is that they could use the method employed by the MLS. Don't actually draw lines, just let the VAR eyeball it. If he can't see that it's offside, it isn't.

In an interview, Howard Webb, general manager of the Professional Referees Organisation in the US told the Philadelphia Inquirer that the MLS will not use a VAR line drawing technology for offside decisions. Instead they will use the following process:


The IFAB have already hinted that they are moving away from the idea that offside should be a totally factual decision and thinking of introducing some sort of "clear and obvious" criteria to it - which I have to say, surprised me but there we are. One way to do that might be to admit that the search for a technically perfect solution hasn't provided "what football wants or expects" and use a deliberately less precise system relying on human judgment.

I know that rather goes against the whole idea of using all this wonderfully advanced modern technology to come to an incontrovertibly correct decision but I think we've realised by now that that may not be possible anyway, due to the limitations already discussed.

I like the idea of the AR's call counting for something

 

[GraemeS]

My other thought (and it's already been mentioned) is that they could use the method employed by the MLS. Don't actually draw lines, just let the VAR eyeball it. If he can't see that it's offside, it isn't.

In an interview, Howard Webb, general manager of the Professional Referees Organisation in the US told the Philadelphia Inquirer that the MLS will not use a VAR line drawing technology for offside decisions. Instead they will use the following process:


The IFAB have already hinted that they are moving away from the idea that offside should be a totally factual decision and thinking of introducing some sort of "clear and obvious" criteria to it - which I have to say, surprised me but there we are. One way to do that might be to admit that the search for a technically perfect solution hasn't provided "what football wants or expects" and use a deliberately less precise system relying on human judgment.

I know that rather goes against the whole idea of using all this wonderfully advanced modern technology to come to an incontrovertibly correct decision but I think we've realised by now that that may not be possible anyway, due to the limitations already discussed.

The concern with that is always going to be camera angles. I regularly watch the MLS Instant Replay show, and it's fairly common in that league for the VAR to go "I haven't got a particularly good angle, so I guess we'll just go with the onfield decision". And in a world where people can't accept lines, I'm far from convinced they'll accept the VAR just shrugging.

 

[Eddie]

The concern with that is always going to be camera angles. I regularly watch the MLS Instant Replay show, and it's fairly common in that league for the VAR to go "I haven't got a particularly good angle, so I guess we'll just go with the onfield decision". And in a world where people can't accept lines, I'm far from convinced they'll accept the VAR just shrugging.

Yeah you can imagine the uproar if a title winning goal is ruled out on the above basis, only for the post-match analysis to show if we had the lines, he would’ve ruled him on.

There’s never going to be a system that suits everyone. They’ll complain no matter what.

I’m not a fan of how the PGMOL have implemented VAR this season but can we really go back to before?

 

[JamesL]

The concern with that is always going to be camera angles. I regularly watch the MLS Instant Replay show, and it's fairly common in that league for the VAR to go "I haven't got a particularly good angle, so I guess we'll just go with the onfield decision". And in a world where people can't accept lines, I'm far from convinced they'll accept the VAR just shrugging.

Yes, but it would emphasise the clear and obvious philosophy.
I think the precision of lines gives the clear (less margin of error) but not obvious.
Take the lundstrom offside for example, his toes was adjudged to be offside, this was clear, but it wasn't obvious, without the precision applied by the lines.
I also think that if it can't be seen with a naked eye then on field call. It emphasises the final decision being the OFRs and the VAR moving back to the advosry role the protocol says it should take.