It’s sometimes interesting to look back at old projects. Sometimes I wonder how on earth I managed to make them work in the first place.
The last time the World Cup came around, back in Qatar in 2022, I was producing predictions and analysis based on what can only be described as a gloriously overcomplicated spreadsheet. It pulled together goal timings, compared records between teams and spat out numbers that I then used to make my own assessments of each game.
It wasn’t exactly elegant, but it was mine.
Fast forward a few years and the landscape has changed completely. Gareth Southgate has gone, Thomas Tuchel is in charge of England and, perhaps more surprisingly, artificial intelligence has become a genuine part of my workflow.
I should probably confess something at this point: I’m not a software developer.
I’ve spent my career writing about sport rather than building software or web apps. Yet over the past year I’ve found myself repeatedly asking the same question whenever an idea pops into my head:
“I wonder if AI could help me build that?”
More often than not, the answer has turned out to be yes.
What started as an idea for improving my football analysis has gradually become a web application that lets me dig into team trends far more quickly than I ever could with a spreadsheet. Along the way I’ve used Python (thanks, AI) to generate scripts that automatically collect data, discovered the importance of asking AI the right questions and accepted that fixing one bug usually creates two more.
It’s addictive.
The football tool is still a work in progress. I’m continually adding features, changing things that don’t work and occasionally breaking everything before putting it back together again.
The name, however, remains a problem.
AI has suggested names like Fixture Context and Soccer Sandbox. Neither has quite clicked for me yet, although the latter is currently winning by default. If you’ve got a better suggestion, I’m all ears.
The football project is only the beginning, though.
I have the beginnings of a Test cricket simulator that I hope will eventually evolve into a management game, a field hockey management game, a horse racing analysis tool, a tennis model that’s currently refusing to cooperate and even a couple of things I built for my eldest son simply because I wondered if I could.
I’m beginning to realise that the common thread isn’t football.
It’s curiosity.
This is going to be somewhere to document that curiosity. Some weeks there’ll be football analysis. Other weeks I’ll be writing about building a cricket simulator, wrestling with AI prompts or explaining why something I thought would take twenty minutes somehow consumed an entire weekend.
Some experiments will work.
Others definitely won’t.
But I suspect there’ll be something interesting to learn either way.
If you’d like to have a play with the football tool, I’d genuinely love to hear what you think: https://soccer-sandbox.lovable.app
Even if your feedback is simply, “This bit doesn’t make any sense.”
That’s often the most useful kind.
What I’ve been working on this week
⚽ Continued refining the football analysis tool to include capacity for xG statistics and goalscorers
🏏 Expanded the Test cricket simulator to include a range of venues with their own characteristics
🐎 Experimented with racecard analysis