Football (English-style) has a new companion – the video assisted referee (VAR).

The proliferation and tumbling price of video technology have made it possible for plays and incidents in a football match to be scrutinised from many different angles at great speed, so that a referee not on the pitch can spot something that the referee on the pitch might have missed. Was a goal offside? Was there a handball or foul as a goal was scored? Spitting or kicking away from the immediate action?

This season’s UK Premier League season has kicked off, with VAR being used for the first time. So, what are the VAR rules?

The Football Association Rules say this under Law Five (emphasis added)

The referee may be assisted by a video assistant referee (VAR) only in the event of a ‘clear and obvious error’ or ‘serious missed incident’ in relation to:

  • goal/no goal
  • penalty/no penalty
  • direct red card (not second caution)
  • mistaken identity when the referee cautions or sends off the wrong player of the offending team

The assistance from the video assistant referee (VAR) will relate to using replay(s) of the incident. The referee will make the final decision which may be based solely on the information from the VAR and/or the referee reviewing the replay footage directly (‘on-field review’).

Except for a ‘serious missed incident’ the referee (and where relevant other ‘on-field’, match officials) must always make a decision (including a decision not to penalise a potential offence); this decision does not change unless it is a ‘clear and obvious error’.

The International Football Association Board (IFAB) goes further in explaining the basic principles.

It did not take long for the new VAR rules to create controversy. A goal for Manchester City was disallowed because the shoulder (or armpit?) of the attacker receiving a key pass was assessed to have been fractionally offside:

And today a goal for Wolves was disallowed because the ball was assessed to struck an attacker’s arm/hand en route to goal. (This year there is a also new substantive ‘strict liability’ rule that even if such ball-hand/arm contact by an attacker as a goal is scored is accidental, the goal is disallowed.)

That would mean that the epic goal scored by Tottenham against Manchester City in last season’s Champions League match would (now) be ruled out – from one angle (but not all angles) there is a slight glance as the ball brushes his arm before hitting his side and flying into the net.

In different cases this week the VR system used the best available technology to see what exactly happened, and (it seems) got the right answer under the rules as then applying. This has not stopped pundits moaning that VAR has ‘not got it right’. The same people seem not to grasp that for long years and decades before VAR they were moaning about poor refereeing.

So the correct way to look at an innovation like this is NOT to compare the disadvantages of VAR with the advantages of the previous 100% human referee on-pitch decisions, but rather to look at the disadvantages AND advantages of VAR and then compare those with the advantages AND disadvantages of 100% human referee on-pitch decisions

What about wider considerations?

First and foremost, life is not precise. However clever the technology and the snazzy graphics and computer-generated angles, there are always limits on what can be detected, and so all decisions are in important sense arbitrary.

Take the handball issue. How microscopically small does ball-hand contact have to be before it is deemed irrelevant? If someone’s shoulder is a mere atom offside (assuming that that can be measured), should that count?

Note that it makes no sense to say there should be an agreed ‘margin of error’, as the controversy simply shifts to whether an action is within that margin or not. Margins too have margins.

One reasonable way to deal with such things is to lay down a principle rather than a rule: if (say) a ‘factual’ situation such as offside is not clear, the advantage is deemed to lie with the attacking side (or the defending side), where the former option is likely to lead to rather more goals being allowed.

This is how VAR now works. There is a principle that (a) a referee on the pitch has to make a decision (eg goal or no-goal) and then VAR overturns it only if there is a clear and obvious (sic) error.

It nonetheless seems to me that the new system risks being annoyingly over-intrusive. I’d do it differently.

I’d keep the option of the off-pitch VAR alerting to the referee to a significant missed incident that would probably have led to a booking had the referee spotted it (eg where X kicks Y well away from the action) at any point in the game.

But otherwise I’d make VAR checks available only on appeal by one side or the other, where each team has (say) three VAR appeals per match. If the appeal is upheld by VAR the appeal does not count against the appealing team’s appeal tally (as in tennis Hawkeye appeals).

The same principle would apply as now, namely that the decision of the on-pitch referee would be overturned only in the case of clear/obvious error. Plus any VAR review would need to be completed within 30 seconds or else the referee’s decision stands.

My scheme gives the players themselves more responsibility – do they want to risk losing an appeal in doubtful cases? If the defending side does not see a slight glancing handball or cannot detect a teensy offside disadvantage, so be it. They don’t appeal and the goal stands. But it no doubt would lead to at least six VAR stoppages per match.

Is that too many? Maybe. But the current VAR regime is no less open-ended and has one odd feature that also can disrupt a game badly, namely that there is no time pressure to review the decision quickly, as accuracy is more important than speed (VAR Handbook, Section 2).

Why is accuracy more ‘important’ than speed? If it’s not clear to the VAR screens within (say) 20 seconds of the review being launched that something egregious has occurred, it’s self-evident that any error by the referee was NOT clear and obvious? And what if there is ultimately no ‘accuracy’ available anyway, even in theory?


Leicester City manager Brendan Rodgers got it wrong today:

“[VAR] has to always be consistent and if there’s a handball and that leads to a goal, it gets ruled out.

The officials are trying to work it the best they can. It’s not perfect yet but as long as you get to the right decision, that’s what we want.”


The aim of VAR is NOT to get every decision right. It is to reduce the number of obviously wrong decisions, and in as timely a manner as the technology allows. As the IFAB handbook says:

Match officials make hundreds of decisions in every match, including decisions that an offence has not occurred. It would be impossible, without completely changing football, to review every decision.

So, again, people griping about VAR ‘got something wrong’ in any given incident miss the point. It’s all about tweaking the inevitable margin of error.

The right test is to take a bloc of decisions that previously looked wrong, and compare them with what would happen now with VAR. If there is a significant reduction (say 50% or more), the system is working well as intended even if a goodly number of decisions remain unhappy or (on even closer scrutiny) ‘wrong’.