The Problem
The Old Way
VALUATION BASED ONLY ON CURRENT FORM
SUBJECTIVE, BIASED DECISIONS
NO FUTURE VALUE ESTIMATION
SCATTERED USE OF HISTORIC DATA
RELIANCE ON SMALL EXPERT GROUPS
VALUE JUDGED ONLY IN THE PRESENT
The New Way
PREDICTIVE MODELS USING MACHINE LEARNING & TIME SERIES
DATA-DRIVEN, UNBIASED PROJECTIONS
LONG-TERM FORECASTS WITH CONFIDENCE INTERVALS
SYSTEMATIC INTEGRATION OF PLAYER HISTORY AND CONTEXT
SCALABLE, AUTOMATED INTELLIGENCE
PLAYER AS A QUANTIFIABLE FUTURE ASSET
Technical Description
Feature Vector Construction
Each player is modeled as a high-dimensional feature vector:
This vector includes historical performance, contractual status, injuries, transfers, and contextual variables.
Cluster-Based
Model Segmentation
Players are grouped into clusters based on similar evolution patterns. A specialized regression model is trained for each cluster using supervised learning and multivariate time series.
Value Forecast Computation
The future market value is predicted as:
Where Δ is the projected value change over time horizon T, calculated based on trends, volatility, and performance dynamics.
Confidence & Risk Output
Each prediction includes confidence intervals and full prediction ranges — quantifying uncertainty due to age, injuries, transfers, and context — enabling clubs to assess upside, downside, and volatility.
Details:
This chart shows the predicted evolution of a player’s market value over time, using Lionel Messi as an example:
-
Current Market value: The green dot on the left marks the current market value (€55.0M).
-
Forecasting: The blue dashed line represents the median forecast of the player's market value at multiple future horizons (from Sep 2025 to May 2027).
-
Confidence: Each gray box shows the confidence interval (IQR) — the tighter the box, the higher the certainty of the forecast.
-
Trend: The curve shows a downward trend, reflecting aging or decline in performance.
This type of visualization helps clubs assess how a player’s value is expected to evolve, supporting smarter investment decisions with quantified risk and timeline clarity.
When Winning was creating the first GTO Solver for football — turning uncertainty into strategy, they realized this was impossible as we did not have the proper ways of measuring this.
Other Objetives
Reduce Risk
Reduce investment risk through reliable projections and quantified uncertainty.
Precision
Reliability
Confidence
Optimize
Optimize academy strategies (development, loans) based on data-driven growth paths.
Scouting
Growth
Opportunity
HighTalented
Identify high-potential talent despite current underperformance.
Development
Data
Planning
Predictive
Strengthen negotiations using predictive arguments rather than subjective opinions.
Leverage
Accuracy
Fairness
Focus
Professionalize football investment by shifting focus to long-term value and profitability.