Future plans

This is a prototype. I intend to run it for years.

Ninety was built in thirteen days for a hackathon. What exists today is a working prototype — a real distributed system with a real data feed and a real on-chain component, but a prototype nonetheless.

I am not building this to win a prize and move on. This is a long-term project, and the World Cup is a starting point rather than the destination. There is a football match somewhere every single day, and the loop Ninety is built on works for all of them.

#Where the money goes

Everything currently runs on free tiers — Vercel, Aiven, Fly, Solana devnet. That was a deliberate constraint and it produced a better architecture, but it is also the ceiling. The first investments:

Infrastructure

Proper compute — dedicated EC2 instances sized for the engine and the quant worker, rather than shared 512MB machines. Managed Postgres with real connection pooling instead of a 20-connection free tier. A production Redis. Proper observability.

TxLINE credits

The free World Cup tier is generous, but paid TxLINE access unlocks more competitions, deeper markets, and higher rate limits. That is the single highest-leverage purchase — it turns one tournament into every league.

Sportmonks subscription

For the general football data the current free tiers ration — squads, deep player statistics, historical results, and the wider competition coverage that makes a consumer experience feel complete rather than tournament-shaped.

#The product roadmap

1
Mobile
A native application is the obvious next surface. Football is watched with a phone in hand, and a scroll-and-tap experience built for that context will always beat a responsive website.
2
A more seamless interface
The current UI is dense and functional. The next pass is about flow — fewer decisions, faster paths, better transitions, and an experience that feels effortless rather than merely complete.
3
Deeper data
Lineups, expected goals, head-to-head history, heatmaps, and per-player market impact across every competition.
4
The AI Pundit Bot
EarlyWhistle with text-to-speech, so the match can talk to you while you are doing something else.
5
Social and creator layers
Shareable Moments, rivalries, group leagues, and a public record of who actually predicts football best.

#Why this compounds

The interesting long-term asset is not the interface — it is the data. A play-money exchange produces a record of how people actually price football when they have no financial incentive to distort it. No bookmaker can farm that, because nobody plays honestly when their own money is at risk.

That record is what makes everything else possible: a better prediction engine, a genuinely useful consumer product, and a clean, legal top-of-funnel that a sportsbook could never build for itself.

#Where the money goes

Infrastructure

TxLINE credits

Sportmonks subscription

#The product roadmap

Mobile

A native application is the obvious next surface. Football is watched with a phone in hand, and a scroll-and-tap experience built for that context will always beat a responsive website.

A more seamless interface

The current UI is dense and functional. The next pass is about flow — fewer decisions, faster paths, better transitions, and an experience that feels effortless rather than merely complete.

Deeper data

Lineups, expected goals, head-to-head history, heatmaps, and per-player market impact across every competition.

The AI Pundit Bot

EarlyWhistle with text-to-speech, so the match can talk to you while you are doing something else.

Social and creator layers

Shareable Moments, rivalries, group leagues, and a public record of who actually predicts football best.

#Why this compounds

That record is what makes everything else possible: a better prediction engine, a genuinely useful consumer product, and a clean, legal top-of-funnel that a sportsbook could never build for itself.