Video review is restricted to a short list of incident types and a high threshold for intervention. Both limits exist because the alternative degrades the match faster than it improves the decisions.
The threshold is a workload control
Almost every match contains dozens of contacts that could plausibly be fouls, and a meaningful share of them would repay a second look. Reviewing all of them would mean stopping play so often that the match stopped resembling a match.
Limiting review to errors that are clear on replay reduces the caseload to a handful of incidents rather than a continuous audit. The filter has to be strict, because the pool of reviewable moments is effectively unbounded.
The threshold is therefore less about respecting the referee and more about keeping the review system inside a workable volume.
Marginal calls do not get more correct on replay
For genuinely close incidents, additional angles rarely produce agreement. Two reviewers can watch the same footage and reach opposite conclusions, because the footage supplies detail but not a rule for weighing it.
That means intervening on marginal decisions swaps one contestable judgement for another while adding a stoppage to the match.
Restricting intervention to clear errors targets the cases where replay genuinely resolves something the naked eye could not.
Scope creep changes what fans expect
Once a system reviews one category of incident, pressure builds to extend it to adjacent ones that feel similarly important.
Each extension is defensible in isolation and cumulatively turns a safety net into a general appeals process. Nobody chooses that outcome; it arrives one reasonable decision at a time.
Competitions that have widened scope have generally seen review counts rise faster than the number of decisions actually changed.
Every stoppage has a knock-on cost
A review interrupts the rhythm of play, cools the crowd and extends the effective duration of the match. Players stand still in cold weather, which carries its own physical cost.
Added time compensates for the clock but not for the momentum, and players report that restarts after long checks feel different.
Those costs are diffuse and hard to measure, which is why they are easy to discount when arguing for wider review.
Where the line sits is a policy choice
There is no technically correct threshold. The setting reflects how a competition weighs accuracy against continuity.
Leagues that prize flow tend to intervene rarely and accept more uncorrected errors. Leagues that prize accuracy accept more interruptions.
Both positions are internally consistent, and most of the public argument about video review is really an argument about which trade-off a competition should make.