VAR's Real Flaw Isn't Bad Calls — It's Selective Review

Football has spent seven years arguing about whether VAR gets decisions right. That's the wrong argument. The Germany–Paraguay knockout tie exposed the real problem, and it isn't accuracy — it's attention.

In one match, the video officials spent several minutes forensically combing a crowded six-yard box to find a reason to disallow a German goal — then showed no interest at all in two clearer holding offences on German attackers at the other end. Same match. Same type of set-piece contact. Opposite outcomes.

That is not an error-correction system. That is discretion — and discretion applied unevenly is worse than no review at all.

First, the part VAR gets right

Credit where it's due. Semi-Automated Offside Technology (SAOT) is VAR at its best: objective, factual, fast, and consistent. Offside is binary — a player is beyond the second-last defender or he isn't — and SAOT checks every goal against one standard and returns an answer in seconds. Nobody seriously accuses SAOT of bias, because it has no discretion to abuse. Hold that thought; it's the whole argument.

The distinction that actually matters

You'll hear it said that "if every touch in the box were a foul, you'd have ten penalties a match." That's true, and the Laws agree: contact alone is not an offence. The line refereeing has always drawn is between two things:

• Ordinary mutual jostling — both players tussling for position, minimal, two-way. Play on.
• Clear, one-sided holding that stops an opponent playing — a bear-hug, a man-marking grab-and-drag when the ball isn't even playable, pulling a player to ground. A foul, and always has been.

Drawing that line is the job. It is inevitably a judgement. That is fine. The problem is never that a line exists.

What actually went wrong

The scandal in Germany–Paraguay was that VAR drew that line in opposite places depending on who benefited:

• A German goal was disallowed for minimal, mutual — arguably even involuntary — contact on the goalkeeper: squarely the "ordinary jostling" category.
• Two clear, one-sided holds on German attackers — a bear-hug at one corner, and a goalkeeper man-marking and hauling an attacker down with the ball not playable at another — were waved away: squarely the "clear holding" category.

By any single consistent standard, that combination is impossible. Apply the strict standard and the goal stands and Germany get two penalties. Apply the lenient standard and the holds go unpunished and the goal stands. The one outcome no coherent reading of the Laws produces is the exact one that happened: strict against one team, lenient for the other, minutes apart.

Collina's defence — and why it backfires

FIFA's refereeing chief, Pierluigi Collina, defended the disallowance, writing that attackers who "show no interest in the ball and move deliberately with the clear intention of obstructing" an opponent should be penalised, "even with minimal contact," especially against the goalkeeper.

His principle is sound: deliberate screening is a foul. But there are two problems.

Threshold. "Minimal contact" is the wrong bar. If minimal contact on a goalkeeper is a free kick, every keeper has a licence to initiate contact, go down, and win the whistle — the precise gamesmanship the Laws are meant to suppress. The bar has to be clear impeding, not any touch.

Consistency — and this is fatal. The moment you raise the enforcement standard for deliberate impeding, you must apply it at both ends. By Collina's own criterion, the two holds on German attackers are penalties. You cannot invoke a strict standard to erase one team's goal and a lenient one to ignore the other team's fouls in the same fixture. The statement meant to justify the call instead indicts the officials who missed the rest.

(There is also a live factual question — replay angles suggest the German player may have been struck and lurched involuntarily, which would mean the "deliberate" element at the centre of Collina's own test was never met. But even setting that aside, the consistency failure alone is enough.)

The deeper problem: intervention itself is subjective

Here is the uncomfortable truth that "clear and obvious error" was supposed to solve and doesn't. The standard only works if the choice of what to scrutinise is neutral. It isn't.

A VAR who spends three minutes hunting for a foul in a goalmouth will usually find one — six-yard boxes are a mass of contact. A VAR who doesn't look finds nothing. The outcome is effectively decided before the "clear and obvious" test is ever applied, by the invisible upstream decision of which incidents get the forensic treatment in the first place.

That is why VAR so often feels random, and why supporters reach for darker explanations. It usually isn't corruption. It's unstructured discretion — and unstructured discretion produces exactly the pattern that looks like bias: the same offence treated differently depending on whom it helps.

What SAOT teaches us about the fix

The contrast with SAOT is the entire point. Offside works now because it was taken out of discretion. There is no "should we check this one?" — the system checks every goal, applies one objective standard, and returns a factual answer. Nobody accuses it of bias because it has no discretion to abuse.

The lesson is not necessarily "scrap VAR." It's that VAR is trustworthy exactly where it is objective and factual — offside (SAOT), ball out of play, mistaken identity, the clear safety red card — and untrustworthy exactly where it is subjective and selective: marginal fouls, holds, "was it a screen."

A growing number of voices in the game now argue the honest move is to pull the discretionary layer out altogether and keep only the objective tools. Whether or not you go that far, the minimum is non-negotiable: the same standard, applied to every comparable incident in a match, with the reasoning made public. Transparency and consistency — or nothing.

Calling it straight — including against the team you support

It's easy to make this argument as an aggrieved fan. So, for the record: in our full analysis of Germany–Paraguay, we judged three of the four contested calls wrong — and one correct. The call we said the officials got right was a handball appeal denied to Germany: a defender's arm, struck by the ball during a slide, in a position justified by the movement — no unnatural silhouette, no offence, correctly no penalty. (A respected former referee disagreed and saw a penalty; it is a genuinely close call. We still lean no-penalty, and we say so.)

That is the whole point of an impartial arbiter: it applies one standard whether the call helps your team or hurts it. That is the standard we are asking VAR to meet — and the one this match showed it did not.

The bottom line

VAR's problem was never that it can't get a call right. It's that the decision to intervene is itself a subjective, uneven act — and until that upstream discretion is structured, made consistent, and made transparent, every "clear and obvious error" will carry the suspicion that the booth simply chose where to look. SAOT proved football can build officiating tools that are genuinely trusted. The rest of VAR has to earn that trust the same way: by removing the discretion, not by asking us to take it on faith.

See a call you're not sure about?

Upload the clip and get a structured verdict — the IFAB clause, frame-by-frame reasoning, and a confidence rating — applied the same way every time, whoever it helps. First 20 credits free, no card required.

OURVAR.AI is an independent AI Video Assistant Referee. Frame-by-frame reasoning, IFAB Laws cited on every verdict — including when the call goes against the team you support.