English Football

[folkestoneger 8,233]

1 minute ago, graeme_4 said:

You’ll be drawing lines on screenshots using Microsoft paint in a minute.

The technology is there to get it right, and it does. 

Except it doesn’t . It’s been shown for a couple of years there is a small margin for error. A few inches maybe but it’s not 100% 

[folkestoneger 8,233]

1 minute ago, rossco87 said:

100% with you however I am not entirely convinced the technology is up to being as accurate as it needs to be to make such tight decisions.

 

Image below was discussing a call from August 2019 but is basically highlighting that VAR uses 50Hz (50 frames per second) and picks the image that they can be certain the ball has been played. In the one they were using Sterling was running at 23.4kph and would have moved up to 13cm between frames so there is a possibility that he could have been onside if the ball was actually played earlier.

 

Easiest solution for me would be to build in a tolerance factor to the really tight calls. Should be pretty easy to get an algorithm to figure out the relative speeds that the players involved were travelling and therefor maximum distances covered between frames - i.e. attacker could have moved 20cm forward and defender 10cm back between the frames, therefore there is a 30cm range where the technology is not accurate enough to definitively say where both players were.
 

Carry out the offside check as normal but if the distance between the two lines is within the range then it is too close to call and goes back to onfield decision (or attackers get the advantage even).

 

The main issue at the moment is that we are pretending VAR for offsides is 100% accurate. For the vast majority of cases it is, however for really tight calls the technology being used can’t be accurate enough to give a definitive answer.

Exactly my point. What happened to the solution of drawing thicker lines and if they touch then giving the benefit to the attacker?

I can’t say for sure it was onside just as we can’t be certain it was off. Cases like that should go with the onfield decision

[born a blue nose 90,907]

5 minutes ago, rossco87 said:

100% with you however I am not entirely convinced the technology is up to being as accurate as it needs to be to make such tight decisions.

 

Image below was discussing a call from August 2019 but is basically highlighting that VAR uses 50Hz (50 frames per second) and picks the image that they can be certain the ball has been played. In the one they were using Sterling was running at 23.4kph and would have moved up to 13cm between frames so there is a possibility that he could have been onside if the ball was actually played earlier.

 

Easiest solution for me would be to build in a tolerance factor to the really tight calls. Should be pretty easy to get an algorithm to figure out the relative speeds that the players involved were travelling and therefor maximum distances covered between frames - i.e. attacker could have moved 20cm forward and defender 10cm back between the frames, therefore there is a 30cm range where the technology is not accurate enough to definitively say where both players were.
 

Carry out the offside check as normal but if the distance between the two lines is within the range then it is too close to call and goes back to onfield decision (or attackers get the advantage even).

 

The main issue at the moment is that we are pretending VAR for offsides is 100% accurate. For the vast majority of cases it is, however for really tight calls the technology being used can’t be accurate enough to give a definitive answer.

Good post. 

Changed my mind, definitely off 

[wully 47,850]

 

[Somemightsay7 2,464]

58 minutes ago, rossco87 said:

100% with you however I am not entirely convinced the technology is up to being as accurate as it needs to be to make such tight decisions.

 

Image below was discussing a call from August 2019 but is basically highlighting that VAR uses 50Hz (50 frames per second) and picks the image that they can be certain the ball has been played. In the one they were using Sterling was running at 23.4kph and would have moved up to 13cm between frames so there is a possibility that he could have been onside if the ball was actually played earlier.

 

Easiest solution for me would be to build in a tolerance factor to the really tight calls. Should be pretty easy to get an algorithm to figure out the relative speeds that the players involved were travelling and therefor maximum distances covered between frames - i.e. attacker could have moved 20cm forward and defender 10cm back between the frames, therefore there is a 30cm range where the technology is not accurate enough to definitively say where both players were.
 

Carry out the offside check as normal but if the distance between the two lines is within the range then it is too close to call and goes back to onfield decision (or attackers get the advantage even).

 

The main issue at the moment is that we are pretending VAR for offsides is 100% accurate. For the vast majority of cases it is, however for really tight calls the technology being used can’t be accurate enough to give a definitive answer.

This is all valid, but the issue with tolerances is there will always be calls that require millimetre decisions. 

You could have a 20cm tolerance, but you're still having to make a decision on whether it is 19.99cm or 20.01cm. It just moves where the decision has to be made from line A to line B.

It's the same issue with some discussions around changing offside to be 'Clear line of sight' or the furthest back part of the body rather than the furthest forward. At the end of the day you are still comparing 2 lines, you're just putting the lines in different places. 

When it comes to offside, the most important thing is consistency and of all the things VAR gets wrong, offside is the one they are most consistent with. 

Also - to clear a lot of this up, semi-automated offside uses a 500hz equivalent polling, so this will avoid most of the tolerance issues.

There's a good article from The Athletic today that goes over it here.

If you don't have a subscription here is a snippet 

 

Munich-based Kinexon is the company that worked in conjunction with FIFA and Adidas for five years to produce 1,500 of these high-tech footballs for the men’s 2022 World Cup and 1,500 for the 2023 Women’s World Cup.

After its success, it will do the same for the men’s Euro 2024 and women’s Euro 2025 match ball called ‘Fussballliebe’ — German for ‘love of football’.

“Our in-ball chip is an advanced add-on for SAOT,” says Daniel Linke, product marketing and strategy lead at Kinexon.

“Using only an optical-based system has a greater error margin due to the shutter time of the camera, motion blur, picture angle and resolution.”

The standard broadcast cameras (used by VAR) record at 50 frames per second but Kinexon’s in-ball chip can judge when the ball has been touched 500 times per second.

With VAR in its current form, there is contention about whether the picture was frozen on the correct frame for offside calls — eg, working out the exact moment the ball had left Callum O’Hare’s boot for Coventry’s disallowed goal on Sunday — but this is where Kinexon’s in-ball chip can offer a more precise judgement.

“We’ve all watched games with VAR where play is interrupted and they spend five minutes looking, only to draw the wrong conclusion,” says Linke.

“It’s so tricky to see from the video images exactly when the ball is touched as the picture might be blurred or occluded. With our connected ball, you get that information instantly. It is in perfect synchronisation with the video signal; we have it down to one or two milliseconds.

“Then the camera system’s algorithms are constantly working to assess offside every time a touch signal is sent by the ball chip.”

Linke talks to The Athletic through one of FIFA’s example videos, which shows how the chip complements the cameras to provide virtually real-time offside calls.

 

[graeme_4 34,866]

1 hour ago, folkestoneger said:

Except it doesn’t . It’s been shown for a couple of years there is a small margin for error. A few inches maybe but it’s not 100% 

It’s consistently applied, and consistently correct.

The Coventry comeback would have been great, but if it had stood Man U would have been the ones robbed. 

[folkestoneger 8,233]

Just now, graeme_4 said:

It’s consistently applied, and consistently correct.

The Coventry comeback would have been great, but if it had stood Man U would have been the ones robbed. 

It’s not though. Everybody knows ( except you) that there is a small margin for error. It may be consistently applied but that doesn’t mean it’s consistently correct.

It should be scrapped or amended to change only decisions that are blatantly incorrect. Drawing lines looking for millimetres of difference in positions of moving players should not even come into it when it is known the technology is not perfect.

In all honesty the mancs would have been more robbed by yet another shit handball decision that the offside call

[Somemightsay7 2,464]

3 minutes ago, folkestoneger said:

It’s not though. Everybody knows ( except you) that there is a small margin for error. It may be consistently applied but that doesn’t mean it’s consistently correct.

It should be scrapped or amended to change only decisions that are blatantly incorrect. Drawing lines looking for millimetres of difference in positions of moving players should not even come into it when it is known the technology is not perfect.

In all honesty the mancs would have been more robbed by yet another shit handball decision that the offside call

How do you define what is 'blatantly incorrect'?

[graeme_4 34,866]

1 minute ago, folkestoneger said:

It’s not though. Everybody knows ( except you) that there is a small margin for error. It may be consistently applied but that doesn’t mean it’s consistently correct.

It should be scrapped or amended to change only decisions that are blatantly incorrect. Drawing lines looking for millimetres of difference in positions of moving players should not even come into it when it is known the technology is not perfect.

In all honesty the mancs would have been more robbed by yet another shit handball decision that the offside call

There is, but the system and rules are applied consistently.

How do you define a ‘blatantly incorrect’ offside? Because regardless if your tolerance is a mm or 3 metres you’ll still measure it using the existing technology and approach.

[folkestoneger 8,233]

2 minutes ago, Somemightsay7 said:

How do you define what is 'blatantly incorrect'?

 

2 minutes ago, graeme_4 said:

There is, but the system and rules are applied consistently.

How do you define a ‘blatantly incorrect’ offside? Because regardless if your tolerance is a mm or 3 metres you’ll still measure it using the existing technology and approach.

You take what has been reckoned by scientists to be the margin of error and expand the lines by that much each way. Attacker back, defender forward. If they overlap at all then it should be classed as level and onside.If they don’t then offside with the error rate already accounted for. Yes you would still be using technology but the margin for error would already be built in so less chance of very close calls being wrong. The previous principle was the attacker should get the benefit of any doubt in close calls after all. 

I would still scrap var as it’s killing the game though. 

[rossco87 692]

34 minutes ago, Somemightsay7 said:

This is all valid, but the issue with tolerances is there will always be calls that require millimetre decisions. 

You could have a 20cm tolerance, but you're still having to make a decision on whether it is 19.99cm or 20.01cm. It just moves where the decision has to be made from line A to line B.

It's the same issue with some discussions around changing offside to be 'Clear line of sight' or the furthest back part of the body rather than the furthest forward. At the end of the day you are still comparing 2 lines, you're just putting the lines in different places. 

When it comes to offside, the most important thing is consistency and of all the things VAR gets wrong, offside is the one they are most consistent with. 

Also - to clear a lot of this up, semi-automated offside uses a 500hz equivalent polling, so this will avoid most of the tolerance issues.

There's a good article from The Athletic today that goes over it here.

If you don't have a subscription here is a snippet 

 

Munich-based Kinexon is the company that worked in conjunction with FIFA and Adidas for five years to produce 1,500 of these high-tech footballs for the men’s 2022 World Cup and 1,500 for the 2023 Women’s World Cup.

After its success, it will do the same for the men’s Euro 2024 and women’s Euro 2025 match ball called ‘Fussballliebe’ — German for ‘love of football’.

“Our in-ball chip is an advanced add-on for SAOT,” says Daniel Linke, product marketing and strategy lead at Kinexon.

“Using only an optical-based system has a greater error margin due to the shutter time of the camera, motion blur, picture angle and resolution.”

The standard broadcast cameras (used by VAR) record at 50 frames per second but Kinexon’s in-ball chip can judge when the ball has been touched 500 times per second.

With VAR in its current form, there is contention about whether the picture was frozen on the correct frame for offside calls — eg, working out the exact moment the ball had left Callum O’Hare’s boot for Coventry’s disallowed goal on Sunday — but this is where Kinexon’s in-ball chip can offer a more precise judgement.

“We’ve all watched games with VAR where play is interrupted and they spend five minutes looking, only to draw the wrong conclusion,” says Linke.

“It’s so tricky to see from the video images exactly when the ball is touched as the picture might be blurred or occluded. With our connected ball, you get that information instantly. It is in perfect synchronisation with the video signal; we have it down to one or two milliseconds.

“Then the camera system’s algorithms are constantly working to assess offside every time a touch signal is sent by the ball chip.”

Linke talks to The Athletic through one of FIFA’s example videos, which shows how the chip complements the cameras to provide virtually real-time offside calls.

 

Never been for the line of sight idea as, as you say, just moves the issue on to another aspect.

 

If the new automated system can be that accurate and quick then that changes the dynamic completely and we are in the realms that the technology can deliver a definitive answer.

 

Until we get to that point though (and let’s be honest we aren’t affording that technology up here and other leagues are probably going to struggle - La Liga found out last night they don’t have goal line technology anymore ffs!) I think the system needs to factor in a degree of tolerance that is relative to the limitations of the technology.  Using technology to figure out what the error range between the frames (and then maybe putting on a +\- 10% safety factor) takes away some of the argument - i.e. if player x is a judged a fraction outside the zone even after an allowance for potential movement has been made then there really isn’t an argument as VAR has given every variable possible within the limitations of what it can do.

 

The instances where it is so close that it is even going to matter are probably minimal, but the implications of getting them wrong can be huge.

[graeme_4 34,866]

Just now, folkestoneger said:

 

You take what has been reckoned by scientists to be the margin of error and expand the lines by that much each way. Attacker back, defender forward. If they overlap at all then it should be classed as level and onside.If they don’t then offside with the error rate already accounted for. Yes you would still be using technology but the margin for error would already be built in so less chance of very close calls being wrong. The previous principle was the attacker should get the benefit of any doubt in close calls after all. 

I would still scrap var as it’s killing the game though. 

No it wouldn’t.

You’d just be using thicker lines so people would be classed as onside when they could be marginally off.

If what you’re saying is you want to give an advantage to the attacker, then that’s fine I can understand that, but what you’re saying above is just reducing the level of accuracy ultimately. 

[rossco87 692]

24 minutes ago, graeme_4 said:

It’s consistently applied, and consistently correct.

The Coventry comeback would have been great, but if it had stood Man U would have been the ones robbed. 

That’s a little misleading. It is consistently correct (leaving aside squinty lines / lines drawn in the wrong places arguments) in that the lines on the freeze frame shown are correct.

 

At present we have no way of 100% accurately getting an image of when the ball leaves a players foot. The margin for error is 0.02s which, as above, can be up to 20 - 30cm depending on the speed and direction players are moving at.

 

In most instances that variance isn’t going to make a difference, however in very tight calls there is no way of definitively saying that the image the lines are being drawn on represents where every one was at the exact moment the ball left the players foot.

[folkestoneger 8,233]

1 minute ago, graeme_4 said:

No it wouldn’t.

You’d just be using thicker lines so people would be classed as onside when they could be marginally off.

If what you’re saying is you want to give an advantage to the attacker, then that’s fine I can understand that, but what you’re saying above is just reducing the level of accuracy ultimately. 

I’m saying precisely that. Before VAR we had the benefit of doubt to the attacker (supposedly). They have already changed parts of the offside and handball rules to benefit attackers so why not.

The offside rule was brought in to prevent goal hanging after all not to look for excuses to stop teams scoring goals which is the object of the game.

Or we could acknowledge it’s a sport played by and refereed by humans and just go back to that. I’ve been watching football since the 1960s and this shit is killing my love for it. 

[Somemightsay7 2,464]

23 minutes ago, folkestoneger said:

 

You take what has been reckoned by scientists to be the margin of error and expand the lines by that much each way. Attacker back, defender forward. If they overlap at all then it should be classed as level and onside.If they don’t then offside with the error rate already accounted for. Yes you would still be using technology but the margin for error would already be built in so less chance of very close calls being wrong. The previous principle was the attacker should get the benefit of any doubt in close calls after all. 

I would still scrap var as it’s killing the game though. 

That's the problem with margin of error though.

Let's say we agree the margin of error is 5cm, which sounds fair. You still need to decide if something is 5cm.

So instead of deciding if an attackers leg is 0.00cm or 0.01cm ahead of a defender, you're deciding if it is 5.00cm or 5.01cm.

In theory that makes sense as it gives the attacker a bit of an advantage and is more in line with the point of the offside rule in the first place, but the decision that needs to be made is still is a line 0.01cm one way or the other. It doesn't change the problem at all, it just moves it.

I do think people have forgotten how bad some offside decisions were before VAR. VAR offsides are significantly more accurate than just a Linesman, which isn't up for debate. It might not be 100%, but if a linesman was 80% accurate then VAR is 99%.

 

Regardless, I think the thing we will all agree on is the main issue is not the decisions, it's the speed. If it was instant I don't think anyone would have an issue with it. Semi Automated offside gets us much closer to that and hopefully in the future it's improved further so it is more instant like goal line technology. Of all the things VAR is bad at I don't think Offside is the problem. The inconsistency in handballs and going back 30 seconds to find any sort of foul to disallow a goal is what needs sorting.

[folkestoneger 8,233]

2 minutes ago, Somemightsay7 said:

That's the problem with margin of error though.

Let's say we agree the margin of error is 5cm, which sounds fair. You still need to decide if something is 5cm.

So instead of deciding if an attackers leg is 0.00cm or 0.01cm ahead of a defender, you're deciding if it is 5.00cm or 5.01cm.

In theory that makes sense as it gives the attacker a bit of an advantage and is more in line with the point of the offside rule in the first place, but the decision that needs to be made is still is a line 0.01cm one way or the other. It doesn't change the problem at all, it just moves it.

I do think people have forgotten how bad some offside decisions were before VAR. VAR offsides are significantly more accurate than just a Linesman, which isn't up for debate. It might not be 100%, but if a linesman was 80% accurate then VAR is 99%.

And with a recognised margin of error built in it could be over 99%. As you pointed out the offside rule was never brought in to look for players millimetres ahead of a defender but to stop people just standing up near the goal. 

[graeme_4 34,866]

40 minutes ago, folkestoneger said:

And with a recognised margin of error built in it could be over 99%. As you pointed out the offside rule was never brought in to look for players millimetres ahead of a defender but to stop people just standing up near the goal. 

There already is a margin for error, you’re just increasing it… 

[graeme_4 34,866]

58 minutes ago, rossco87 said:

That’s a little misleading. It is consistently correct (leaving aside squinty lines / lines drawn in the wrong places arguments) in that the lines on the freeze frame shown are correct.

 

At present we have no way of 100% accurately getting an image of when the ball leaves a players foot. The margin for error is 0.02s which, as above, can be up to 20 - 30cm depending on the speed and direction players are moving at.

 

In most instances that variance isn’t going to make a difference, however in very tight calls there is no way of definitively saying that the image the lines are being drawn on represents where every one was at the exact moment the ball left the players foot.

They are correct within that margin for error / stated tolerance though.

[folkestoneger 8,233]

17 minutes ago, graeme_4 said:

There already is a margin for error, you’re just increasing it… 

No there isn’t. They’re measuring millimetres. 

[rossco87 692]

49 minutes ago, graeme_4 said:

They are correct within that margin for error / stated tolerance though.

They aren’t. They are correct in that they take the closest frame they can to the ball leaving the players foot which could be +\- 0.02s or up to 30cm (depending on speed of players and direction of travel) and using that as an arbitrary reference for a definitive answer.

 

A margin for error would be factoring in a zone where the result would have to be “we don’t know - go with the on-field decision” or something similar.

 

Personally I don’t think just making thicker lines works as the margin for error will vary depending on the speed and direction players are travelling at and also the moment the ball is played vs the nearest frame that is available, hence I would advocate that the margin for error is varied taking into account those elements.

 

 

[graeme_4 34,866]

1 hour ago, folkestoneger said:

No there isn’t. They’re measuring millimetres. 

 

17 minutes ago, rossco87 said:

They aren’t. They are correct in that they take the closest frame they can to the ball leaving the players foot which could be +\- 0.02s or up to 30cm (depending on speed of players and direction of travel) and using that as an arbitrary reference for a definitive answer.

 

A margin for error would be factoring in a zone where the result would have to be “we don’t know - go with the on-field decision” or something similar.

 

Personally I don’t think just making thicker lines works as the margin for error will vary depending on the speed and direction players are travelling at and also the moment the ball is played vs the nearest frame that is available, hence I would advocate that the margin for error is varied taking into account those elements.

 

 

That’s literally their margin for error. 

[folkestoneger 8,233]

13 minutes ago, graeme_4 said:

 

That’s literally their margin for error. 

If they were operating a 30cm margin of error that offside call yesterday would never have been made 

[Rangers_no1 32,810]

TBH there are pictures that look like the United line was over AWB foot.

Thankfully the new system will be in place and no more line drawing by a human.

[Moody Blue Legend 61,774]

Some game.

[RealWorldRich 716]

5 hours ago, Somemightsay7 said:

That's the problem with margin of error though.

Let's say we agree the margin of error is 5cm, which sounds fair. You still need to decide if something is 5cm.

So instead of deciding if an attackers leg is 0.00cm or 0.01cm ahead of a defender, you're deciding if it is 5.00cm or 5.01cm.

In theory that makes sense as it gives the attacker a bit of an advantage and is more in line with the point of the offside rule in the first place, but the decision that needs to be made is still is a line 0.01cm one way or the other. It doesn't change the problem at all, it just moves it.

I do think people have forgotten how bad some offside decisions were before VAR. VAR offsides are significantly more accurate than just a Linesman, which isn't up for debate. It might not be 100%, but if a linesman was 80% accurate then VAR is 99%.

 

Regardless, I think the thing we will all agree on is the main issue is not the decisions, it's the speed. If it was instant I don't think anyone would have an issue with it. Semi Automated offside gets us much closer to that and hopefully in the future it's improved further so it is more instant like goal line technology. Of all the things VAR is bad at I don't think Offside is the problem. The inconsistency in handballs and going back 30 seconds to find any sort of foul to disallow a goal is what needs sorting.

A margin of error does solve the problem. A guy who is 5.01cm offside has much much less of a complaint than a guy who is 00.01cm offside.

 

The reason why there needs to be a fairly significant margin of error is because the VAR lines are trying to apply a very exact science to something that's rather inexact.... for example I highly doubt a football pitch is a perfectly flat rectangle where both sidelines and bylines are EXACTLY parallel. Most pitches will probably have some kind of camber too.

The Z axis too (the vertical line the draw doen from s shoulder or knee) placed on a 2d image is never going to be accurate either.

 

And perhaps the biggest problem is that the passing player could have his foot in contact with the ball dor a few frames, which frame they choose could be the difference between on & off If the recieving player is sprinting.