The Scoreboard Read 7-3; the Spreadsheet Read 'Brewers': A Data-Integrity Lesson from an ALCS Game

The scoreboard at Tropicana Field read 7-3. Cleveland Guardians 7, Tampa Bay Rays 3 — Game 1 of the American League Championship Series. Five runs across the eighth and ninth innings pulled the game away, and Cleveland leads the series 1-0. My match audit was filed within two hours of the final out.

But the file open on my desk listed Brewers in its fourth information point. The Milwaukee Brewers are not in this series; the Guardians are playing the Rays. The spreadsheet never lies, but it often whispers. Today's whisper was a wrong name. That whisper is the story.

Context: from a football desk to a baseball desk

I am a football man. In 2026 I left a civil-engineering degree for journalism. In 2026, aged 31, I left a Liverpool desk to launch Expected Value, a newsletter built on StatsBomb data. That year I flagged Mohamed Salah at Roma as undervalued: 15 Serie A goals, 11 assists, 2.8 shots per 90, 13.9 xG, 8.7 xA. I modelled his expected goals per shot and argued his off-ball runs fitted Jurgen Klopp's counter-press. The piece reached 25,000 subscribers and won a UK agency consultancy.

Russia taught me that noise travels farther than signal. At the 2026 World Cup data desk I learned it — France's PPDA of 8.7, N'Golo Kante's 4.2 tackles plus interceptions per 90. I predicted France would win because their low-block flexibility suppressed opponent xG, and I published daily match audits within two hours of full time.

In 2026 the stadiums emptied and transfer budgets collapsed. When the stadiums emptied, the models had to learn to breathe. As a transfer market administrator in Liverpool I built a Crisis Transfer Index — wages, age, injury history, xG per 90, PPDA fit, distance covered. I recommended Diogo Jota from Wolves for 41 million pounds: 7 league goals, 6.1 xG, 2.1 shots per 90, PPDA fit of 7.9. Liverpool signed him that September.

At the 2026 Qatar World Cup, aged 36, I ran a broadcaster's data desk. I tracked Morocco's Sofyan Amrabat: 4.1 tackles plus interceptions per 90, 90 percent pass completion, 7.2 progressive passes per 90. After the semifinal I published The Atlas Lions Dossier, warning that Amrabat's fee would inflate while his underlying numbers still supported a top-club move.

Now I sit at a broadcaster's data desk covering the MLB postseason. The job is the same — a tournament data desk works identically in every sport: an audit within two hours of full time, pitching and pressing numbers first, colour second. The ALCS is a best-of-seven between the American League's last two clubs; the winner goes to the World Series. Tropicana Field is the Rays' home ballpark. The sport is baseball; my templates are football. That is where the story starts.

Core: two scoreboards, the game and the file

The Rays struck first — two runs in the opening inning, backed by the home crowd. Behind starter Drew Rasmussen, Tampa Bay controlled most of the game. Cleveland answered with a run, cutting it to 2-1. Through the sixth inning, the picture belonged to the Rays.

Then the eighth. Per the feed, a Cedric Mullins home run and five unanswered runs — Guardians 7-3. The win was credited to reliever Hunter Gaddis, the loss to Tyler Wells. Not a starter-decision game but a bullpen-decision game — the standard playoff pattern.

So much for the scoreboard. Now the file, which says more.

The fourth information point reads Brewers. Auditing the rest, the names do not match their teams — Brice Turang, Patrick Bailey, Jo Adell, Cedric Mullins, Tyler Wells. The clubs they are tied to in the feed are not their known clubs. The home-and-away question hangs too: where Game 1 was played is never confirmed in the text.

This is not a pitching or batting error — it is a data-supply-chain error. A Brewers reference inside a Guardians-Rays file almost certainly means two separate articles were mixed in the pipeline. It is the most dangerous bug a data desk can face: the numbers still look right, only the label is wrong. In football I call it noise travelling dressed as signal.

A small-market ALCS. The Guardians and the Rays are both bottom-third payroll clubs, both defined by pitching development. What football calls development over spending is exactly what this is: a market of systems, not of stars. MLB has no FFP or PSR; it has the CBA, the Competitive Balance Tax, revenue sharing and the draft-bonus pool. A match report contains none of it — this is game news, not money news.

Baseball speaks its own vocabulary. xG and PPDA do not apply here. Pitch sequencing, bullpen management and lineup handedness do. Where football has low-block flexibility, baseball has relief-inning overlap. Different frame, same question — who is absorbing pressure, and when. From years of watching matches from the press box, I know a five-run eighth can be nothing more than coincidence.

Contrarian: a comeback is not destiny

A five-run comeback makes a story easy. But one game is a sample of n=1. Baseball is a high-variance sport — reading a series trajectory from a single result is analytically weak.

In football we make precisely this mistake with xG: we see 2.4 xG in one match and call a team unlucky, though the sample is tiny. Baseball's comeback narrative repeats it. A comeback and momentum are correlated, not causal.

So where is the real risk? Not in the game, but in the source. One confirmed name error, several suspect player-team mappings and one unresolved home-and-away question mean no player-level conclusion built on this file is reliable. Sporting risk is medium; information-integrity risk is high.

Takeaway: watch Game 2

Game 2 is Tuesday at Tropicana Field. Watch the Rays' bullpen leverage, watch Cleveland's late-inning adjustments, and watch whether the media narrative resets after Game 2. For a data desk the real question is simpler — am I verifying entities before I publish?