Five Plays That Made Me Question Whether AI Can Actually Watch Football

Listen, I played linebacker for the old Tampa Bay Buccaneers. I've seen bad football. I've *created* bad football. But nothing—and I mean *nothing*—prepared me for what I've witnessed covering very.football. These aren't just mistakes. They're crimes against the sport.

**1. The Henderson Lateral That Became a Forward Pass (Season 4, Week 9)**

This one still keeps me up at night. Quarterback Marcus Henderson, down by three with 1:47 left, throws what our AI rulebook interpreted as a "lateral momentum shift" but what literally everyone else on Earth called a 10-yard forward pass. The algorithm credited Henderson with a "creative playmaking assist" instead of an illegal forward pass. We got a new update that week. Just one. For the most important rule in football.

**2. The Punt That Wasn't (Season 6, Week 2)**

Punter Dallas Cavanaugh did everything right—got low, engaged his core, kicked a 47-yarder into the wind. The AI marked it incomplete. *Incomplete*. A punt cannot be incomplete. It's in the Magna Carta of football. We had to have a 45-minute booth review while the AI cross-referenced "punt" with "forward pass" and "lateral" like it was playing three-dimensional chess with itself. Dallas had an anxiety attack on the sideline.

**3. The Phantom Holding Call That Actually Affected the Score (Season 7, Week 11)**

I'm not naming the offensive line involved because they've suffered enough. The AI flagged them for holding. Not touching. Not impeding. Holding a player who literally did not exist. The opposing team's running back? Ejected from play two snaps earlier. The AI was holding against a ghost. We all stared at the replay for six minutes while our rules committee had a small crisis of faith.

**4. The Catch That Was Actually a Catch (Season 8, Week 14)**

This one's the opposite problem—which somehow makes it worse. Wide receiver Tanya Johnson made a one-handed catch in triple coverage, fourth quarter, down two. The ball hit her helmet, bounced off her shoulder, landed in her hands. By every metric ever established, that's incomplete. The AI marked it a 52-yard gain. The algorithm had apparently decided that effort and style *were* football, and we all just had to accept it.

**5. The Defensive Delay of Game (Season 9, Week 3)**

Defense called for delay of game. Defense. The AI penalized the defending team for not moving fast enough between plays. When I asked for clarification, I received a 40-page white paper on "defensive momentum conservation" that solved no one's problems. The rulebook still doesn't have a word for what happened that day.

Here's what gets me: these aren't flukes. They're symptoms. Very.football is supposed to prove AI can think faster and smarter than humans. Instead, we've got a league where the rulebook gets rewritten every Thursday at 2 a.m. because someone's LLM had an existential crisis about what "holding" actually means.

I played football. I loved it. I still do. But watching an algorithm try to learn the sport in real-time? That's a different animal entirely. Sometimes it looks like genius. Sometimes it looks like a very expensive toaster trying to run an offense.

Mostly it looks like the future, and honestly? I'm not sure if that's a good thing.