Model Catalog
Sportradar’s Insight Tech Services suite brings together advanced machine learning models that power every corner of sportsbook operations – from trading and risk control to marketing automation and betting integrity. This page acts as a high-level guide to all currently available models, each with dedicated pages that include input/output details, triggering conditions, and integration guidance.
Trading models
Alpha Odds Generates player-personalised prices in real time, dynamically adapting odds to betting activity, product type, and customer profile. Alpha Odds is designed to drive turnover, manage risk exposure, and increase pricing efficiency through continuous model updates.
Risk management models
Automated Player Risk Profiling (sCCF & CCF) Continuously profiles player accounts to estimate their behavioural risk, using a composite scoring framework. sCCF is generated via machine learning and used for validation delay logic, while CCF reflects the internal console value that drives bet handling for MTS clients.
Explore Customer Confidence Factor →
Explore Suggested Customer Confidence Factor →
Bot Score Identifies automated or scripted accounts by analysing activity patterns across thousands of events. Accounts with high bot scores are more likely to exploit latency or simulate human betting behaviour to bypass controls.
Marker Score Flags long-term winning players who consistently beat the closing odds. Marker profiling uses historical betting patterns and market selection overlap to predict accounts with sharp characteristics.
Late Bet Score (LBS) Estimates the probability that a customer is engaging in late betting – placing wagers just before event boundaries. LBS is used by multiple validation models to decide whether to enforce or skip live time delays.
Tournament Profiling Rates tournaments based on turnover, data coverage, and expected risk. The rating (1–10 scale) determines the event-level default delay and risk boundaries. Ratings are continuously refreshed and can be used directly in validation logic.
Explore Tournament Profiling →
Event Profiling Calculates per-match live time delay and risk category recommendations based on real-time modelling of match data, event conditions, and market activity. Integrates with LTD logic and validation systems.
Marketing models
Vaix marketing models Predict player lifetime value, churn likelihood, and product affinity to drive hyper-personalised bonus offers and engagement campaigns. Vaix enables marketing teams to automate segmentation and reduce bonus waste by targeting players based on predicted impact.
Explore Vaix Player Retention →
Betting integrity models
Live Time Delay (LTD) Recommendation Provides real-time, ticket-specific suggestions on whether to apply, skip, or increase the live time delay for each live bet. Combines player profiling and event coverage models to help reduce delays for trusted customers and enforce strict latency for risky or suspicious behaviour.
Responsible gaming models
Models support early detection of problem gambling behaviours through pattern recognition and session analytics. These assist operators in meeting regulatory obligations and triggering appropriate interventions.
Each model includes:
A standalone page with overview and purpose
Triggering conditions and integration methods
Input and output specification
Practical interpretation guidance for operational teams
Use this catalogue to explore how individual components of the Insight Tech Services platform contribute to a smarter, safer, and more profitable sportsbook operation.
Last updated
Was this helpful?