Every video review rests on a clock nobody watches

Officiating technology is only as trustworthy as the frame alignment between cameras, and that alignment is fragile.

Before a review official can rule on anything, a dozen cameras have to agree on what time it is. Not roughly. To the frame.

This is the least discussed part of match officiating and it is the part that decides whether the rest works. Each camera carries a timecode locked to a shared reference, so that a given frame number on the tight goal-line camera is the same instant as that frame number on the wide tactical. Lose the lock and the review system will happily show two truthful images of two different moments, and a human will be asked to judge a decision from them.

The failure modes are mundane. A camera swapped at half time and not re-synced. A slow motion server running on its own clock. A frame rate mismatch, where a fifty frame camera and a sixty frame camera get conformed to a common timeline and something in between gets manufactured. Those generated frames are the genuinely dangerous ones, because an invented in-between frame looks exactly like a real one and it shows a ball position that never existed.

You can see why offside technology went the way it did. Semi-automated systems do not only track limbs, they impose one sampling regime across the whole capture, so the question of which frame counts stops being a judgement and becomes a defined property of the system. That is a real gain and I think it is undersold. The argument about lines on a shoulder misses the point. The lines were never the hard part. The hard part was agreeing on the instant of the pass.

Where I would push back is presentation. Broadcasters show the review as a still with graphics, which invites the viewer to argue about millimetres and never explains the timing decision underneath. Show the frame counter. Show that this frame was chosen because ball contact registered here. A public that understands the clock will argue less about the line.

And when the clock breaks, say so. Reviews have been conducted on feeds that were not properly locked, and the honest response is to declare the tool unavailable and let the on-field call stand. A system that admits it cannot see is far more credible than one that always produces an answer.