Players, priced honestly.

How it prices a player

A hedonic fair-value model (R² = 0.74 on live data) prices every player from performance, age, minutes, league difficulty, contract and club context. The gap between fair value and market price is the scouting signal — validated per league, with the backtest published including the leagues where it is weak.

• Calibrated probabilities, not vibes — P(value rises), P(falls), P(rises 50%+) per player from a gradient-boosted classifier with isotonic calibration. Rolling out-of-sample design: AUC 0.87–0.90, Brier skill +0.33 to +0.48, clean reliability curves.
• Forward-tested — players flagged as undervalued were signed 1.7× more often than comparable peers in the following window (19.7% vs 10.7%), ranked before the window opened. An earlier version claimed ~3×; my own audit found look-ahead bias in the dating of prices, so I rebuilt it. The honest number is narrower — and defensible.
• Knows its own limits — it declines to price players above €60M: in 98,000 player-seasons there is no comparable case to calibrate on, so it says so instead of guessing.
• Built for the buyer — club view with squad gaps and realistic budgets, "similar players under €1M from your region", contract-expiry radar, one-click club reports.

Explorer · scout ranking & probabilities Player profile · fair value & form Club view · Boca Juniors sample

Alpha Scout's scouting signal is built on a market inefficiency I identified and validated on 52,213 player-window observations across 69 leagues — statistically strong (t = 34.8), economically meaningful, and still systematically underpriced by the transfer market. What the signal is stays in-house: the edge is the product.

• Discipline over drama — a candidate data provider was dropped after forward testing showed no predictive gain. Data that doesn't earn its place stays out of the model.
• Every claim audited — the numbers on this page include the ones that got worse when I checked them properly. That is the point.