Football analytics is solved. GPT-6 Astra + Higgsfield built us a tool that tracks players, maps passes and follows possession while you watch. Watching sports is more fun with AI.

Great job though, but a very horrible idea. This is so much distraction. No football ⚽️ lover will like this 😂.

So cool man great product

brilliant

This could be incredibly useful for TV broadcasts and interactive streams very soon. One thing I’d love to see is a video-game-style option: show only the name of the player in possession and the nearest opponent by default, with the option to turn on labels for every player. Much cleaner and easier to follow.

Can somebody move the analytics? There’s a football match behind it.

Pep’s Guardiola data engineer anyway

No it’s not. Fuck out of here

Soon they won’t even need to play the game anymore, AI can just simulate the game & give us the result. What fun!

Watching sports is more fun when it’s not over-analysed. The sheer amount of data available now is half the problem - football always used to be about scoring more goals than the opposition, and that’s been lost somewhere along the way in the name of getting KPIs up.

clearly you’re not a football person, this implementation already exists for a very long time, slapping a bunch of numbers on a screen is not solving football analytics, there are APIs way richer than this display with the real metrics people wanna see in apps and on TV

Franek Smuda rozpoznawał dobrego piłkarza po tym, jak wchodzi po schodach.

Our model is way better 😎 @Goal_metrices

GPT-6 Astra + Higgsfield is insane.

Higgsfield solved nothing lol

Why do we need GPT 6 Astra for this? This can easily be built using a small vision transformer

you could do this on AWS for years now with computer vision, hence Prime on NFL

Horrible, its impossible to focus on the game, the ball, the players and the spaces - at the same time.

Analytics in any sport are never fully solved. Come on.

Is not more fun, too many distractions

I’m all for AI, but art and sport should remain sacred.

Moi mon cerveau le fait déjà naturellement en regardant un match. Par contre ça peut être très utile dans les centres de formation / écoles pour apprendre le placement, les lignes de passe et la lecture des espaces. Ou pour les paris sportifs en temps réel.

Astra: 'I can see inside brains now' Also Astra: 'and I can tell you why your favorite team is mid' 🧠

i made similar tool enabling any llms to auto labeling any image totally local supporting or external providers 🔗

I especially like how 20 Bernardo is taking the corner and also is in the box at the same time.

Different 🕶️

@sportsballpm stunt on these hoes

None of this solves soccer it doesn't even answer any of the questions a good analyst would want to ask

The person who created this doesn’t know that we don’t want black boxes spoiling our match. You can add this as a feature for coaches not us. There’s a freaking match going on. The only advantage for users is betting.

No, thanks. If I wanted to watch a crappy footie game, I’d play FIFA 93.

It’s only a step away from using robots instead of players, managers and referees. If we have this much real-time information, humas are no longer able to process and execute them. My guess people advocating this have rarely done any sport.

When doing analysis? Fine, when watching football? Heck no

Solved’ is doing a lot of work for a demo clip. Overlay tracking on broadcast footage isn’t new — clubs have run this internally for a decade. The actual unlock would be if this runs live, per-frame, on any match without a bespoke pipeline per broadcast angle.

you can barely see what is happening with all that in the screen

What is higgfield ai role in this @grok