[ENGINE.ARCHITECTURE.v3.4]

The Math is the Moat.

We reveal the math because the math is the moat. Human handicappers have bias, fatigue, and emotion. PIPER doesn't. We don't guess. We calculate. PIPER runs a deterministic 15-path signal engine that polls 45 sports every 15 minutes. A signal only fires when multiple independent mathematical paths align to breach the 2.0 confidence threshold.

SIGNAL PATH COMPUTATION ENGINE

PATH 01 ELO_DIFFERENTIAL base_score � 8.0x
PATH 02 INJURY_IMPACT star=-0.80 / starter=-0.40
PATH 03 LINE_VALUE_DISC threshold: >12% deviation
PATH 04 B2B_FATIGUE NHL back-to-back: -0.50
PATH 05 HOME_COURT contextual modifier
PATH 06 FORM_STREAK last 5 games weighted
PATH 07 SHARP_MONEY line movement direction
PATH 08 TOURNAMENT_CTX underdog quality = 60
PATH 09 WEATHER_FACTOR outdoor sports only

THRESHOLD confidence = 2.0
REQUIRED minimum 2 signal paths must align
OUTPUT deterministic � same inputs = same output
CYCLE 15-minute polling across 45 sports

PATH_01: ELO Differential

Quantifying absolute team quality beyond public perception.

FORMULA: (TEAM_A_ELO - TEAM_B_ELO) * 8.0x_MULTIPLIER

OUTPUT: +2.4 CONFIDENCE per 30pt ELO gap

PATH_02: Injury Impact Scoring

Real-time ESPN API polling cross-referenced with roster depth velocity.

IF STAR_PLAYER == 'OUT' THEN CONFIDENCE -= 0.80

IF STARTER == 'OUT' THEN CONFIDENCE -= 0.40

IF ROLE_PLAYER == 'OUT' THEN CONFIDENCE -= 0.15

PATH_04: Schedule Fatigue

BACK_TO_BACK: -0.50 CONF

REST_ADVANTAGE: +0.40 CONF

PATH_03: Line Value (CLV)

DEVIATION > 12%: +0.50 CONF

Detects DraftKings/ESPN spread inefficiencies vs internal Elo models.

THE EXECUTION GATE

Signals do not fire on a single metric. A pick is only pushed to the dashboard when Σ CONFIDENCE > 2.0 AND ALIGNMENT ≥ 2 PATHS.