Game1 AI - Frequently Asked Questions

Source: PDF uploaded by Shane (2026-02-12)

Outcomes & Validation

• 95% hit rate = tied to pro contracts in top 10 European leagues
• Means the algorithm will rarely overlook top talent. If you never let go of flagged players, you won't lose top talent.
• "Pro contract" = signing and being active in a top 10 European league
• Methodology detailed in Game1 whitepaper

Data Scope

• Nearly 10 years of historic data
• Model validated retrospectively on unseen training data
• Dataset collected in Europe; players tracked worldwide
• Follows players in top 4 tiers of the country they play in (goes deep in the pyramid)

Model Design - Biological Maturation

• Model accounts for biological age using: height, seated height, weight, chronological age
• Can identify late developers earlier
• Makes early developers and late bloomers comparable in trajectory
• Measures future potential, not current performance, eliminating selection bias based on biological maturity

Methodology - Mixed Small-Sided Games

• Teams continuously rotated so every player plays with and against everyone
• 6-10 games needed for true plus/minus model
• Game1 provides schedules to ensure proper rotation
• Teammate quality isolated through rotation (evens out)
• Algorithm backtested as predictive for all positions except goalkeepers
• Large number of games in different team contexts ensures individual contribution is isolated