gWAR Calculation Service

[grovecj]

Overview

Implement the core gWAR (Grove WAR) calculation service that computes our transparent, simplified WAR metric.

Parent Issue

Part of #94 (WAR and Advanced Stats)

gWAR Formula

gWAR = (Batting + Baserunning + Fielding + Positional + Replacement) / 10

Batting     = wRAA = ((wOBA - lgwOBA) / wOBAScale) × PA
Baserunning = wSB  = (SB × 0.2) + (CS × -0.41)
Fielding    = OAA × 0.9 (from Baseball Savant)
Positional  = Position lookup × (games / 162)
Replacement = PA × (20.5 / 600)

Positional Adjustments (per 162 games)

Position	Adjustment
C	+12.5
SS	+7.5
2B	+3.0
CF	+2.5
3B	+2.5
LF	-7.5
RF	-7.5
1B	-12.5
DH	-17.5

Pitcher gWAR (FIP-based)

pitchingRuns = (lgFIP - FIP) / 9 × IP
replacement = IP × 0.03
gWAR = (pitchingRuns + replacement) / runsPerWin

Files to Create

domain/gwar/GwarCalculationService.java

@Service
public class GwarCalculationService {
    public void calculateAndApply(PlayerBattingStats stats, String position);
    public void calculateAndApply(PlayerPitchingStats stats);
    
    private BigDecimal calculateWraa(BigDecimal woba, int pa, LeagueConstants lc);
    private BigDecimal calculateWsb(int sb, int cs);
    private BigDecimal calculateFieldingRuns(Integer oaa);
    private BigDecimal calculatePositional(String position, int games);
    private BigDecimal calculateReplacement(int pa);
}

domain/gwar/GwarComponents.java

Record for component values.

Test Cases

• Known player produces expected gWAR (within tolerance)
• DH gets correct -17.5/162 positional adjustment
• Player without OAA gets fielding = 0
• Edge cases: 0 PA, null wOBA, null position
• Pitcher FIP-based calculation

Acceptance Criteria

• Calculation matches expected values for test players
• All component values stored on entity
• Handles missing data gracefully (nulls)
• Unit tests cover all calculation paths

Dependencies

• gWAR Schema: League Constants and Component Fields #169 (gWAR Schema)