Back to Home
Resources

Documentation

Last Updated: December 2025

Enterprise AI Risk Management for Sports Betting Operators

Table of Contents

  1. Getting Started
  2. Platform Overview
  3. Risk Command Center
  4. Sharp Detection Engine
  5. Integrations
  6. Configuration
  7. Best Practices
  8. FAQ

Getting Started

Quick Start Guide

Welcome to SharpShield. This guide will help you get your risk management platform operational within 24 hours.

Step 1: Account Setup

  1. Log in to your SharpShield dashboard at app.sharpshield.io
  2. Complete your organization profile under Settings > Organization
  3. Add team members and assign roles under Settings > Team
  4. Configure two-factor authentication for all admin accounts

Step 2: API Integration

  1. Navigate to Settings > API Keys
  2. Generate your production API key
  3. Integrate the SharpShield SDK into your platform (see API Reference)
  4. Begin sending bet data and player information

Step 3: Configure Thresholds

  1. Go to Settings > Risk Thresholds
  2. Set your risk tier boundaries based on your risk appetite
  3. Configure automated actions for each tier
  4. Set up alert notifications

Step 4: Enable Real-Time Monitoring

  1. Access the Risk Command Center
  2. Verify data is flowing correctly
  3. Review initial player assessments
  4. Fine-tune settings based on your player base

Platform Overview

Architecture

SharpShield operates as a cloud-based SaaS platform with real-time processing capabilities:

┌─────────────────┐     ┌──────────────────┐     ┌─────────────────┐
│  Your Platform  │────▶│  SharpShield API │────▶│  ML Processing  │
│   (Sportsbook)  │◀────│    Gateway       │◀────│    Pipeline     │
└─────────────────┘     └──────────────────┘     └─────────────────┘
                              │                         │
                              ▼                         ▼
                        ┌──────────────────┐     ┌─────────────────┐
                        │   Risk Command   │     │  Model Training │
                        │     Center       │     │    & Updates    │
                        └──────────────────┘     └─────────────────┘

Core Components

1. Risk Scoring Engine 2. Sharp Detection AI 3. Risk Command Center 4. Automation Engine

Risk Command Center

Dashboard Overview

The Risk Command Center is your central hub for monitoring and managing betting risk.

Key Metrics Panel

MetricDescription
Active Sharp PlayersPlayers currently classified as HIGH or CRITICAL risk
Today's Risk ExposureTotal liability from sharp-classified bets
Pending ReviewsPlayers requiring manual assessment
CLV LeakageEstimated value lost to sharp bettors

Live Activity Feed

The activity feed displays real-time events:

Player Management

Player Profile View

Each player profile contains:

Risk Assessment Betting Analysis Sharp Indicators Applied Limits

Bulk Actions

Perform actions on multiple players:

  1. Select players using filters or checkboxes
  2. Choose action: Adjust Limits, Add to Watchlist, Export Data
  3. Confirm and apply

Alert Configuration

Alert Types

AlertTriggerPriority
Sharp DetectionPlayer crosses HIGH thresholdHigh
Tier ChangeRisk tier upgradeMedium
Unusual PatternAnomaly detectedHigh
Syndicate LinkAccount connection foundCritical
Liability ThresholdEvent exposure limitHigh
Steam MoveSharp money detectedMedium

Notification Channels

Configure notifications via:

Sharp Detection Engine

How It Works

SharpShield's detection engine uses a multi-layered approach combining statistical analysis, machine learning, and behavioral pattern recognition.

Layer 1: Statistical Analysis

CLV (Closing Line Value) Tracking

CLV measures the difference between the odds when a bet was placed versus the closing odds. Sharp bettors consistently beat the closing line.

CLV = (Closing Odds - Bet Odds) / Bet Odds × 100

Players with average CLV > 2% over 100+ bets are flagged for review.

Win Rate Analysis

Long-term win rates above expected values indicate sharp behavior:

Market TypeExpected Win RateSharp Threshold
Point Spreads50.0%> 53%
Totals50.0%> 53%
MoneylinesVaries> Expected + 3%
Props48-50%> 52%

Layer 2: Machine Learning Models

Ensemble Model Architecture

SharpShield uses an ensemble of models:

  1. Gradient Boosting (XGBoost) - Pattern classification
  2. LSTM Networks - Temporal betting patterns
  3. Random Forest - Feature importance analysis
  4. Neural Networks - Complex pattern recognition
Feature Set

The models analyze 150+ features including:

Layer 3: Behavioral Analysis

Timing Pattern Detection

Sharp bettors often exhibit specific timing behaviors:

Syndicate Detection

Algorithms identify potential connected accounts through:

Risk Scoring

Score Calculation

The risk score (0-100) is calculated as a weighted combination:

FactorWeight
CLV Performance30%
Win Rate vs Expected25%
ML Model Prediction25%
Behavioral Patterns15%
Account Signals5%

Risk Tiers

TierScoreRecommended Action
LOW0-30No action required
MEDIUM31-60Monitor, standard limits
HIGH61-85Apply reduced limits
CRITICAL86-100Maximum restrictions

Integrations

API Integration

See API Reference for complete documentation.

Integration Modes

Mode 1: Real-Time Analysis

Submit bets for analysis before acceptance:

Your Platform → SharpShield API → Risk Analysis → Accept/Reject

Latency: < 100ms average

Mode 2: Asynchronous Processing

Submit bets after acceptance for analysis and model training:

Your Platform → SharpShield API → Background Processing → Alerts
Mode 3: Hybrid

Combine real-time analysis for flagged players with async processing for all bets.

Third-Party Integrations

Odds Providers

SharpShield integrates with major odds providers for CLV calculation:

Data Platforms

Communication Tools

Configuration

Risk Thresholds

Default Settings

{
  "risk_tiers": {
    "low": { "min": 0, "max": 30 },
    "medium": { "min": 31, "max": 60 },
    "high": { "min": 61, "max": 85 },
    "critical": { "min": 86, "max": 100 }
  },
  "auto_actions": {
    "high": {
      "reduce_limits_by": 50,
      "notify_team": true
    },
    "critical": {
      "reduce_limits_by": 90,
      "restrict_markets": ["spreads", "totals"],
      "notify_team": true,
      "require_review": true
    }
  }
}

Customization

Adjust thresholds based on your business model:

Automated Actions

Available Actions

ActionDescription
REDUCE_LIMITSDecrease stake limits by percentage
SET_MAX_LIMITSet absolute maximum stake
RESTRICT_MARKETSBlock specific market types
ADD_DELAYAdd bet acceptance delay
REQUIRE_REVIEWQueue for manual approval
SUSPEND_ACCOUNTTemporary account freeze
NOTIFY_TEAMSend alert to team

Rule Builder

Create custom automation rules:

IF player.risk_score > 75
   AND player.sport = "NFL"
   AND bet.stake > 1000
THEN
   reduce_limits(75%)
   AND notify_team("High-value NFL bet from sharp")

Best Practices

Implementation Recommendations

Phase 1: Monitoring (Week 1-2)

Phase 2: Alerts (Week 3-4)

Phase 3: Automation (Week 5+)

Operational Guidelines

Daily Tasks

  1. Review overnight alerts
  2. Check new HIGH/CRITICAL classifications
  3. Process pending reviews
  4. Monitor liability exposure

Weekly Tasks

  1. Review sharp player trends
  2. Analyze CLV leakage reports
  3. Adjust thresholds if needed
  4. Team performance review

Monthly Tasks

  1. Generate comprehensive reports
  2. Review automation effectiveness
  3. Audit false positive/negative rates
  4. Strategic planning session

Common Pitfalls to Avoid

  1. Over-Restricting: Don't limit every winning player; focus on true sharps
  2. Ignoring Context: Consider market conditions and sport specifics
  3. Delayed Response: Act on critical alerts promptly
  4. Set and Forget: Regularly review and adjust settings

FAQ

General Questions

Q: How long does it take to see results?

A: Initial classifications are available within 24-48 hours of integration. Model accuracy improves over 2-4 weeks as more data is collected.

Q: What data do you need from us?

A: At minimum: player IDs, bet details (stake, odds, market, outcome), and timestamps. Additional data like device info and session data improves accuracy.

Q: Do you store player personal information?

A: We only store the data necessary for risk analysis. We never require or store names, addresses, or payment information. See our Privacy Policy.

Technical Questions

Q: What's the API latency?

A: Average response time is < 100ms for risk queries. Real-time bet analysis averages 50-75ms.

Q: Can we use our own risk models alongside SharpShield?

A: Yes. Our API returns raw scores and features that you can incorporate into your own models.

Q: How often are models updated?

A: Models are retrained weekly with new data. Major model updates are deployed monthly with changelog notifications.

Pricing Questions

Q: How is pricing calculated?

A: Pricing is based on monthly active players analyzed. See our pricing page for current tiers.

Q: Is there a free trial?

A: We offer a 14-day free trial with full feature access. Contact sales@sharpshield.io.

Support

Contact

Resources

SLA

PlanResponse TimeUptime SLA
Growth24 hours99.5%
Professional4 hours99.9%
Enterprise1 hour99.99%

© 2025 SharpShield. All rights reserved.