The Honesty of an Empty Cell: In Football's Data Economy, 'I Don't Know' Is the Rarest Skill

One in the morning, Brisbane. A live match dashboard is open on the laptop — event feed, shot map, possession curve, and an xG counter ticking away in the corner like a clock. In the 67th minute there should be a corner from the right. The cell is empty. Still empty in the 68th. By the 70th it has filled itself with a wrong entry — a shot that never happened, filed under a name.

Nobody in the commentary box stopped. The graphic floated up, an arrow got drawn on the shot map, the commentator said “according to the latest data.” At the time I put the empty cell down to technical noise. Later I understood it was the most honest document of that night.

Based on my nine years of watching football and sitting behind a microphone, here is what I know: the rarest skill in football’s data economy is not inventing new numbers, it is having the nerve to publish nothing. An empty cell never lies; the system that fills it with a guess is the one doing the lying.

I went looking for the highlight reel and found a spreadsheet instead — and that spreadsheet taught me that respecting an empty cell is really about respecting your own receipts.

To see why, you have to picture the pipeline. A top-league match generates somewhere between fifteen hundred and two thousand event entries — passes, duels, clearances, throw-ins, set pieces. Tracking systems record 10 to 25 frames per second, and on top of that sit the models: xG, xGA, PPDA, progressive passes. At the far end of all of it sit three kinds of buyers: broadcast, club analytics, and the in-play market.

The first two can survive a bad number for a few minutes. The third cannot, because there the number is no longer information. It is a price.

Two dates from my own file. May 7, 2026: the A-League Grand Final, Sydney FC 1-1 Melbourne Victory, 4-2 on penalties. That regular season Sydney had taken 66 points from 27 games, a league record. Plenty of people wrote them up as the boring champions. The 66-point season taught me that volume is not the same as voltage — banking points week after week and deciding a final across 120 minutes are two different skills.

Second date: June 20, 2026. Three days earlier Germany had lost 1-0 to Mexico at the Russia World Cup. I wrote that Germany would not get out of the group. On June 27 they lost 2-0 to South Korea and finished bottom of Group F. I had also filed Croatia reaching the final during the group stage.

Then I published a scorecard: 11 predictions, 9 correct, 2 wrong, every one timestamped. Every hot take starts as a hunch; the receipts decide if it survives. The habit has a name — writing the prediction down before the event and checking it afterwards. In a science lab it is called pre-registration. In football punditry it has no name, because nobody does it.

So what actually happens when a feed comes back empty? I recognise three layers of failure. One is collection — cameras, operators, event mapping. Another is the model: the input is fine, but the model does not know the context of the match. The last is distribution: the data is right but arrives two seconds late, and two seconds is all the market needs to move.

What do you really learn by staring at an empty cell? You learn the data is telling you something about its own health. The best way to judge how reliable a feed is, is not its most spectacular number but how many empty cells it leaves behind. If a provider misses the 67th-minute corner, its 92 per cent pass accuracy also joins the suspect list.

Then comes the question of context. The same shot means two things. A shot from outside the box at 0-0 and the identical shot in the 88th minute while two goals down register the same value to a model and two different events to a match. Outside the building, people read them as one. Inside the building, they know how high the defensive line was, who was running on fumes, who had to take the risk.

Which is where I hang an old coat on an old hook: the goalkeeping market. A keeper with 92 per cent pass completion looks superb on a highlight reel and turns a pleasant shade of green on a dashboard. The post-shot data sitting next to it says how many goals he actually prevented that season. The market still pays more for the first number than the second. I call that a failure of pricing, not a measure of quality.

This is where volume and voltage separate. What does two thousand passes in a season mean? A midfielder who plays 70 passes a game but never plays the one line-breaking pass that decides it has high volume and low impact. Weight the actions by match state or the spreadsheet will turn a space-occupier into a star. Accumulating numbers and turning a match are not the same thing — that idea keeps returning to almost every argument I have.

The other end of the pipeline is more uncomfortable, and it is discussed far less. Live event data does not only travel to a television screen; within seconds it lands in the in-play market. Corners, cards, shots, injury time — everything gets an instant price. Which means that when a feed arrives two seconds late or carries a bad entry, the people building a market on top of it are standing on a false foundation with real money.

And those who are not at the market? They hold a phone, watch a delayed stream, and start filling the empty cell with a guess. Latency stops being data and becomes an edge. This is the darkest corner of football’s datafication — information built to keep competition fair turns into a price in the hands of whoever has the fastest connection.

That third lesson worked its way into how I operate. Every bet, every headline, every “I’m seeing this” gets written down first, with a date. When the match ends, the scorecard goes public. Right gets an argument, wrong gets a name. There is one clear benefit: there is no room left to lie to myself.

Still, data scepticism gets misread. “If the cell is empty, keep it empty” is not a call to stop watching football. It is the opposite: lining up what the eye sees against what the feed records becomes more necessary, not less. Walking away from data is not honesty; it is another kind of laziness wearing a better outfit.

And one thing needs saying plainly. The most common reason a live feed comes back empty is production, not analysis. A camera started late, an operator took a break, an event-tagging map had the wrong set-piece code. An empty cell is not always a seal of truth. Sometimes it is just the signature of a broken pipeline.

Now let me turn my own argument over, otherwise it becomes another half-truth that circulates forever.

The simplest objection is that I am treating empty cells as proof of integrity when they may simply be failed collection. Feeds go blank for technical reasons far more often than moral ones. “No data” and “bad data” are not the same thing, and if I hand every empty cell a certificate of honesty, I have shut the door on verification.

Another objection comes from my own street. If scepticism has been my identity for years, then “the data is broken” becomes a product in my hands. And when scepticism becomes a brand, it becomes merchandise; merchandise needs an audience, and an audience needs numbers. The circle closes.

A final objection is aimed at my own memory. The eye test is data too, just badly logged. My memory is a low-sample, high-noise recorder. A feed that misses one corner in the 67th minute misses fewer corners than my memory does — nobody is just keeping count.

Even with those three objections standing, the core claim holds: pretending to know what you do not know, and selling a thing as known when it is not, are different offences — and football rewards the second one more.

So where should we look? I would not be surprised if, within the next two seasons, at least one major league or broadcaster publishes a data-integrity audit, spelling out in black and white how many entries each provider left blank. If it does not happen, I will take it that nerve is still not a purchasable commodity in this market.

The next time a number floats up in the corner of your screen, with a small credit line underneath, take one second to ask: who is making this number, and who are they selling it to? Brisbane gave me the rhythm; the internet gave me the megaphone. The rest is receipts.

I file every prediction under “verifiable” — the question is whether football actually wants that folder.