Our Methodology

Under the Hood

Transparent. Deterministic. Data-Driven.
No black boxes. Every factor explained.

10 Factors
In projection model
MIP Solver
Not LLM guessing
Live Data
Updated daily

The Pipeline

From raw data to optimized squad

Data IngestionStep 1
Projection EngineStep 2
Squad OptimizationStep 3
Data Collection

Automated & Always Fresh

Our ETL orchestrator automatically collects and processes data from multiple sources, ensuring projections are always up-to-date.

Official FPL API integration
Scheduled refreshes throughout the day
Post-deadline data capture
Multi-season historical aggregation
Automated gameweek reconciliation
ETL Orchestrator
State-driven scheduler
Daily Refresh
03:00 UK
Post-Deadline
Hourly
GW Reconciliation
Weekly
Status
Live
Projection Engine

10 Factors That Power Our Projections

Each factor is carefully tuned and contributes to the final expected points calculation.

Factor 1

Historical Performance

Multi-season data aggregation
Recency-weighted scoring
Minutes threshold filtering
Factor 2

Position Context

Position-specific baselines
Role-aware adjustments
Scoring profile normalization
Factor 3

Attacking Threat

xG and xA integration
Per-90 normalization
Position-aware goal weighting
Factor 4

Fixture Difficulty

Opponent strength analysis
Home/away adjustments
Clean sheet probability modeling
Factor 5

Price-Based Adjustment

Premium player resilience
Budget player variance
Price-tier calibration
Factor 6

Form Analysis

Rolling form window
Baseline comparison
Gradual adjustment blending
Factor 7

Confidence Adjustment

Sample size awareness
Positional mean regression
Confidence-weighted blending
Factor 8

New Player Handling

Recency scoring
Conservative defaults
Track record requirements
Factor 9

Availability & Minutes

Nailed/rotation/fringe classification
Expected minutes prediction
Injury return adjustment
Factor 10

Planning Horizon

Multi-week planning
Uncertainty discounting
Regular starter protection
Process

How We Calculate Expected Points

Each player's projection flows through a multi-stage process that combines historical data, context, and real-time factors.

1

Historical Data

Multi-season performance aggregation

2

Position & Context

Role-aware baseline adjustment

3

Fixture Analysis

Opponent strength & home/away

4

Form & Confidence

Recent trends & sample size

5

Final Projection

Status-adjusted expected points

Output
6.2 xPts

Ready for squad optimization

Optimization

Math, Not AI Guessing

We use a Mixed-Integer Programming (MIP) solver to find the mathematically optimal squad. No hallucinations. No randomness. Smart warm-starting means faster results, and advanced symmetry-breaking finds better solutions.

Your current squad and selling prices
Available budget in the bank
Transfer limits and point hit costs
Planning horizon (1-5 gameweeks)
FPL constraints (3 per team, position limits)
Fast solving with intelligent starting points
MIP Solver
Deterministic Optimization
Budget Constraint≤ £100.0m
Squad Size= 15 players
Max Per Team≤ 3 players
Smart SolvingWarm Start
ObjectiveMax xPts

See It In Action

Connect your FPL team and get your first optimized squad recommendation.