Advanced Player Statistics
En resumen
Advanced Player Statistics extends the basic profile with throw-role classification, opponent-level scouting that bundles streaks and format splits, court win-rate heatmaps, cochonnet-distance accuracy buckets, regression-based season trends with R-squared, ELO timelines annotated with peak, low, and delta-from-peak, real-time milestone detection on personal records, and percentile peer-group comparison within the player's competition class.
Cómo funciona
This subsystem layers on top of the throw-level fact table from F10.02 and the rating snapshots feeding F10.01. Throw-type distribution (F10.07.01) classifies each player as predominantly pointeur, tireur, or milieu by computing the share of point vs tir attempts over a rolling window and writing the classification back to the player's analytics document; the dashboard shows the underlying ratio so it is transparent rather than opaque. Win-rate per opponent (F10.07.02) reads pair-keyed rollups and adds a streak tracker (current streak, longest win/loss streak) plus a format breakdown (singles/doubles/triples).
Court performance heatmap (F10.07.03) groups match outcomes by court_id and renders the venue's court layout with each cell shaded by win-rate; clubs use this for opponent scouting on home terrain. Cochonnet distance analysis (F10.07.04) buckets points into 6-7 m, 7-8 m, 8-9 m, and 9-10 m and computes pointing accuracy per bucket from the throw fact table. Season trends with linear regression (F10.07.05) fit a least-squares line to the per-season win rate and report slope, intercept, and R-squared so the player sees not just direction but also fit quality.
The ELO development timeline (F10.07.06) plots the rating snapshot series and annotates peak, low, and delta-from-peak. Personal records and milestones (F10.07.07) watch the event stream for thresholds (longest winning streak, highest single-match score, biggest comeback) and emit a milestone the moment it is broken. Peer-group comparison (F10.07.08) computes the player's percentile against the rollup-population for their competition class so a 1450 ELO is contextualized as 78th percentile in regional class.
All advanced reads share the same idempotent rollup substrate and the live-delta layer, so updates land within seconds of a match closing.
Capacidades clave
- Throw-role classification (pointeur/tireur/milieu) with transparent ratio
- Opponent scouting: pair rollups, streaks, and per-format splits
- Court-level win-rate heatmap rendered on the venue's court layout
- Cochonnet distance buckets (6-7, 7-8, 8-9, 9-10 m) for pointing accuracy
- Linear-regression season trend with slope, intercept, and R-squared
- ELO timeline annotated with peak, low, and delta-from-peak
- Percentile ranking within the player's competition class
En la práctica
Before a regional final, captain Théo opens opponent scouting on his upcoming adversary. The win-rate-per-opponent panel shows their head-to-head is 4-3, current streak is 2 wins, and his win rate jumps to 71% in doublettes. The throw-type distribution confirms the opponent is a near-pure tireur (78% tir attempts), which Théo cross-references against his own court heatmap for the venue: court 7 favours pointing duels, where the opponent's accuracy in the 8-9 m bucket drops to 41%.
He requests court 7 for the draw. Post-final, his ELO timeline shows a fresh peak of 1612, his season trend regression nudges up, and the platform fires a milestone notification: longest winning streak in 2026 now stands at 9.
Funcionalidades de este subsistema
8| ID | Status | Funcionalidades |
|---|---|---|
| F10.07.01 | Entregado | Throw-type distribution (pointeur/tireur/milieu classification) ✅ PL-F1007a |
| F10.07.02 | Entregado | Win-rate per opponent (with format breakdown and streak tracking) ✅ PL-F1007a |
| F10.07.03 | Entregado | Court performance heatmap (win rate per court_id) ✅ PL-F1007a |
| F10.07.04 | Entregado | Cochonnet distance analysis (6-10m buckets) ✅ PL-F1007a |
| F10.07.05 | Entregado | Season trends with linear regression (win-rate trend over seasons) ✅ PL-F1007b |
| F10.07.06 | Entregado | ELO development timeline (rating progression with peak/low/delta) ✅ PL-F1007b |
| F10.07.07 | Entregado | Personal records and milestones (streaks, highest score, comebacks) ✅ PL-F1007b |
| F10.07.08 | Entregado | Peer-group comparison (percentile ranking within competition class) ✅ PL-F1007b |
Partes interesadas que necesitan este subsistema
Aparece en 6 análisis de partes interesadas