How does an FPL optimizer actually work?

Gathering the data

Before anything else, the tool needs to know who's available and how much each player costs. It pulls official player data directly from the Fantasy Premier League API — every player's price, position, recent form, and injury status, updated before each gameweek.

But FPL data alone doesn't tell you who's likely to score. So we layer in live odds from betting markets, which reflect the collective wisdom of millions of bets placed on each match. These odds tell us which teams are heavy favourites to win, which games are likely to be low-scoring, and which goalkeepers are most likely to keep a clean sheet. That fixture context is what separates a smart pick from a lucky one.

Calculating expected points

Once we have both the player stats and the fixture context from betting odds, every single player gets assigned a predicted score for the upcoming gameweek. This is called their expected points (xP).

The xP calculation takes into account a player's position, their recent scoring and assist form, and how favourable their next fixture looks according to the markets. For example, a striker playing for a team that's a heavy favourite at home against a side struggling to keep clean sheets will get a high xP. The same striker facing a tough away fixture against the league's best defence gets a much lower one. This means every pick reflects the actual difficulty of each gameweek's fixtures — not just who scored well last week.

AI-powered squad selection

With an xP score for every available player, the optimizer's job is to find the best possible combination of 15 players. The catch: there are hundreds of millions of valid squads that fit FPL's rules — so you can't just check them all by hand.

Instead, we use a technique called linear programming — an optimization algorithm that mathematically solves which squad maximizes total expected points while respecting every constraint simultaneously: your budget (to the nearest £0.1m), a maximum of 3 players from any single club, the correct positional split (2 GKPs, 5 DEFs, 5 MIDs, 3 FWDs), and any players you've chosen to include or exclude. This is the same class of algorithm used in airline scheduling, logistics route planning, and financial portfolio optimization — applied here to fantasy football.