FantaScreener: fantasy football for investing (and why the first draft told it wrong)

Let me start with a mistake, just to get off on the right foot. The first draft of this post, from the pen that writes with me, described FantaScreener as a tool to not blow the fantasy football draft. Wrong — and stopped before it went out. FantaScreener has nothing to do with fantasy football: it's fantasy football applied to something else. It's fantasy football for investing.

And how I got there is the whole point.

The idea: steal the engine, not the game

Fantasy football works in Italy for a reason that has nothing to do with football: it makes you study. For weeks people read probable line-ups, lineup battles, injuries — stuff nobody would read if there weren't a team, a league of friends, a standings table, a reason to come back every week.

The question behind FantaScreener was: what if I took that engine — teams, roles, standings, fixed sessions — and put learning to invest inside it instead of football? Not a tool to give tips (never — you won't find a recommended stock or a signal here, by choice and by hygiene). A community that makes financial education as engaging as a fantasy league among friends.

The build: a whole league, built alone

Here's the usual vice — the obsessive naming, the disproportionate care. FantaScreener isn't a page: it's a whole league with its own vocabulary. There are Teams and Fanta Traders, there's the FantaBroker (the simulator where you practice without risking a euro), there's FantaConoscenza with monthly SMAC sessions (Monthly Analysis and Comparison Sessions), the Operational Insights (reflections written after closing a simulated position: what happened, what I learned), the Market Analyses. There's even a Transparency section where I publish the project's aggregated expenses, because if you preach financial discipline you have to show it first.

I didn't build a product and then the jargon. I built the jargon because naming is half the design: the moment a user is a "Fanta Trader" inside a "Team", they already understand how to behave, without reading a manual.

The errors and attempts (the real part)

An engine like this banged its head plenty. Two scars I keep gladly:

• The 4.5 MB limit. There's a feature that analyzes documents uploaded by users. On certain files it broke, and my first instinct was to hunt for the bug in my own code. The real root cause: Vercel caps request bodies at 4.5 MB. It wasn't my bug, it was a wall of the platform. Fix: upload the file straight to storage with a signed upload, bypassing the route. The lesson, the usual one: the symptom always looks like your bug, but sometimes it's a limit of the field you're playing on.
• The key that arrives only when needed. The AI features (document analysis, feedback triage) run on a key that isn't mandatory: if it's missing, the feature degrades gracefully — the manual form stays identical, everything else works. Provisioning on demand. Because a piece of AI must never be able to switch off the whole product.

Why I'm writing this

Because FantaScreener is a snapshot of how my head works. The kickabout among friends twenty years ago, the commentated tournament highlights, the radio for thirty people: they were all exercises in getting people to come back and do something together, with relish. FantaScreener takes that craft — building a reason to take part — and points it at the driest subject there is.

The back-of-the-napkin numbers, honest edition: users who'll get rich thanks to me, zero (that's not the game). People who might learn something while having fun, all those who take part.

It's always the same craft, from the Belvedere pitch onward: build a team, keep a standings table, give someone a reason to come back the following week.

Except this time the ball is a chart.