England vs Argentina Prediction: Our Model and the AIs

You should know right out the gate... the creator of this site is a big England fan. Hi there, in case it’s the first time we’re meeting, my name is Ash Anderson, and I’m that England fan. I also happen to love sports, and that’s why we’re here, to talk about them.

At the start of the World Cup I made a prediction model to simulate the entire World Cup. I ran 10,000 sims, and then ran 10,000 more after every single game. One thing has remained consistent, and it’s something I don’t like: Argentina will win the World Cup.

Don’t get me wrong, I’m a big fan of Messi. Messi though, he wasn’t there when I was a kid watching England fail time and time again to bring football home. So, here I am in 2026 with a dilemma. Believe my model, or believe my heart.

What The Data Says

As I mentioned, we’ve consistently had Argentina as the winner of the World Cup. Their opponent has gone back and forth between whether it be France or Spain. So, while the model settled on the wrong side yesterday (we had Spain at just 49.6% to advance, and they snuck through), it was insanely close.

Overall, our model has predicted at 81% throughout the knockout rounds. It’s been identical to Claude Sonnet 5, but slightly behind on the brier score (calibration). That 81% is the track record though, not tomorrow’s odds. Our model gives England a 38% chance tomorrow, and that’s the one I’m holding onto.

Claude Sonnet has Argentina at 68%. Our model has Argentina at 62%. Every other AI is picking Argentina at between 59 and 68%. We were in a similar spot yesterday with the France game too. Every AI picked France to advance, albeit with lower confidence.

Regardless, we grade it

As with everything on PredictedSports, whether we want the outcome or not, it will be graded and stored. You can see the real time grades on the website for all the models we’re tracking.

A flat bet on the model’s picks throughout the World Cup would have netted just north of a 20% return, so I’d consider that a decent model.

We’re doing the same for MLB, UFC, and soon NFL and College Football too, so I’d encourage signing up for the newsletter if you’re not on here already. And, if you want more depth, consider a free or pro account to make the most of one of the deepest sporting datasets around.