FraudGuard Pro

Enterprise Fraud Management Suite
All Systems Operational

Threshold Configuration Monitoring

Real-time tracking of risk threshold performance, alert volume optimization, and calibration insights | Environment: Production

Active Thresholds

24

Configured across all rule sets

Threshold Breaches (24h)

1,247

Alerts generated

Optimal Range Compliance

91.7%

Thresholds within target precision

Pending Adjustments

3

Requiring review/approval

Threshold Set 1: Transaction Amount Limits

Monitors single-transaction and cumulative amount thresholds for fraud detection triggers
● Active
Current Threshold Values
₱50,000
Min: ₱10,000 Recommended: ₱45K-₱60K Max: ₱100,000
₱200,000
Min: ₱50,000 Recommended: ₱180K-₱250K Max: ₱500,000
0.5x
Applies reduced thresholds for electronics, gaming, crypto categories
Last Adjusted: 2026-02-10 by Risk Ops Team
Business Justification: Calibrated to balance fraud capture vs. customer friction per Q4 2025 review
Regulatory Reference: BSP Circular 1112, Section 4.2 (transaction monitoring thresholds)
Threshold Performance: Precision vs. Alert Volume
Current Optimal Zone Threshold Value → Metric Value
Precision Alert Volume Current Setting
Performance Metrics (Last 30 Days)
Precision at Current Threshold 93.4%
False Positive Rate 6.6%
Alerts Generated (Daily Avg) 142
Within target range: 120-180/day
Optimization Insight: Lowering threshold to ₱45,000 could increase fraud capture by ~8% but would raise alert volume by ~22%. Recommended for high-risk customer segments only.

Threshold Set 2: Transaction Velocity Limits

Controls frequency-based triggers: transactions per time window, rapid sequential activity detection
● Active
Current Threshold Values
5 txns
Min: 3 Recommended: 4-6 Max: 15
20 txns
Min: 10 Recommended: 18-25 Max: 50
90 seconds
Triggers when consecutive transactions occur faster than threshold
Last Adjusted: 2026-02-12 by Data Science Team
Calibration Method: ROC analysis on 6-month holdout dataset
Segmentation: Thresholds vary by customer tier (Retail/Premium/Corporate)
Sensitivity Analysis: Threshold vs. Detection Rate
Current: 5/10min Detection Rate False Positives Velocity Threshold (txns/10min) → Rate (%)
Fraud Detection Rate False Positive Rate
Alert Volume Trend (7-Day)
Feb 10 Today
Avg. Daily Alerts 187
Peak Day (Feb 14) 312
Target Range 150-220
Observation: Alert volume increased 67% on Feb 14 (Valentine's Day). Consider implementing holiday/seasonal threshold adjustments to reduce false positives during predictable high-volume periods.

Threshold Set 3: Composite Risk Score Thresholds

Defines action triggers based on aggregated risk scores from multiple models and rules
● Active
Current Risk Score Thresholds
30/100
Triggers enhanced monitoring, no customer impact
65/100
Triggers step-up authentication, transaction review queue
85/100
Triggers automatic hold, immediate analyst alert, potential account restriction
Scoring Model: Ensemble weighted average (Velocity 40%, Behavioral 35%, Geographic 25%)
Last Calibrated: 2026-02-15 using precision-recall optimization
Business Impact: Thresholds balance fraud loss prevention vs. customer experience friction
Risk Score Distribution with Threshold Markers
30 65 85 Risk Score (0-100) → Population %
Low Risk Medium Risk High/Critical
Threshold Trigger Metrics (24h)
Score ≥30 (Enhanced Monitoring) 847
No customer impact; internal flagging only
Score ≥65 (Step-Up Auth) 142
Customer challenged; 78% passed verification
Score ≥85 (Auto-Hold) 23
Transactions held; 19 confirmed fraud (82.6% precision)
Performance Note: Critical threshold (85) achieving 82.6% precision, exceeding target of 80%. Consider maintaining current setting; lowering to 80 would increase volume by ~40% with marginal fraud capture gain.
Threshold Governance & Optimization Framework
Threshold Management Protocol:
  • Change control: All threshold adjustments require Risk Committee approval
  • A/B testing: New thresholds validated on 5% traffic before full deployment
  • Performance monitoring: Real-time precision/recall tracking with alerting on degradation
  • Seasonal adjustments: Pre-approved threshold variations for holidays/events
  • Documentation: Business justification and regulatory alignment recorded for each change
Regulatory & Compliance Alignment:
  • BSP Circular 1112: Threshold calibration documented per transaction monitoring requirements
  • AMLC Guidelines: Risk-based threshold approach aligned with suspicious activity detection standards
  • Data Privacy Act: Threshold logic operates on pseudonymized data; customer impact assessments conducted
  • Model Risk Management: Thresholds treated as model hyperparameters with independent validation
  • Audit readiness: All threshold changes, rationales, and performance impacts logged for 7 years
Professional Risk Disclosure: Threshold configuration involves tradeoffs between fraud detection efficacy, operational cost, and customer experience. Metrics shown reflect historical performance and may vary with portfolio composition, fraud landscape evolution, and external factors. All threshold adjustments must follow institutional governance policies, regulatory guidance, and documented business justification. Illustrative data shown for system design purposes; production thresholds subject to change based on ongoing validation and risk appetite reviews.
Threshold metrics refreshed: 2026-02-16 17:10 PHT | Next calibration review: 2026-03-15