S15M11: We're safe (for now)

Its's 2025. I'm a busy man. So I did what anybody would do. I asked ChatGPT to lighten my load.

My load is not lightened.

After four attempts, it’s clear that asking ChatGPT to write a football blog is like asking a goldfish to explain offside. The enthusiasm’s there — lots of words, plenty of confidence — but somewhere between the algorithmic optimism and the “irreverent tone” brief, it loses the plot. One version started musing about kettles, another reviewed imaginary teams, and by the third, we’d entered full existential crisis territory. So, rather than risk another 400 words of synthetic banter and misplaced metaphor, we’re cutting our losses. The AI’s gone back to its data cave, and we’re sticking to what it can’t ruin — the results. Cold, clean, factual results. No tangents, no tortured analogies, no leaf-clearing philosophy. Just the scores, as they happened, free from poetic interpretation or robotic rambling. Sometimes you have to admit defeat, whistle for full-time, and take the point. Now, on to the football.

This weekend saw an interesting series of events where all the models (av 7.1) outperformed the normal, regular joes (av 5.5) by quite a bit. This isn't quite an AI thing, but made me think about how expertise has been reviewed in isolation, and what a collated model might be, One for the off season perhaps...

Let's get statty:

This week, 22 people played
Most popular predicted result: Arsenal WIN (22/22)
Most disputed predicted result: Spurs vs Chelsea (6-9-7 split)

Highest odds: Josh Daniels (13,815/1)
Lowest odds: AFM (835/1)
Average odds: 4,854/1

Best predictor: Joe Abbott (9/10 - bloody west ham!)
Worst predictor: Daniel Wigman (3/10)
Average score: 6.00/10

Best predicted result: Arsenal WIN (22/22)
Worst predicted result: West Ham WIN (0/22)

Everyone's score:

Leaderboard (>2/3; /)

---

To this week's predos:

Good luck all