About Sibyls Edge
Bringing institutional-grade sports analytics to everyday bettors using machine learning.
Our Story
Sibyls Edge was founded in 2024 by a team of data scientists and sports analysts who were frustrated by a simple problem: the tools used by professional bettors and quant trading firms to model sports outcomes were simply out of reach for everyday bettors. Sharp money had access to sophisticated models, historical databases, and real-time line movement analytics. Everyone else was guessing.
The idea was straightforward. Take the methodology used in institutional quantitative analysis, apply it to sports data at scale, and make the outputs accessible to anyone with an internet connection. The name "Sibyl" is a nod to the oracles of ancient history - those who sought patterns in noise to glimpse likely futures. That's exactly what our models try to do: find the signal buried in decades of sports data.
Early 2024
Concept and Research
Initial research phase. Built the first version of our historical game database and began training baseline prediction models on NFL and NBA data.
Mid 2024
Model Validation
Expanded the dataset, implemented backtesting frameworks, and validated model performance against closing lines to measure CLV (Closing Line Value).
Late 2024
Beta Launch
Launched a closed beta with a small group of testers. Refined the app interface and added MLB and NHL coverage based on user feedback.
2025 - Present
Public Launch and Growth
Public launch with full feature set including Kelly Criterion stake sizing, CLV tracking, and real-time odds analysis across all four major sports.
The Technology
Our prediction engine is built on several interconnected components, each designed to extract maximum signal from available data. Here's what's running under the hood.
🧠
Machine Learning Models
Our core prediction models are trained on historical game data spanning 10+ years across all four major US sports leagues. We use ensemble methods combining gradient boosting and neural network approaches to generate win probability estimates and expected value calculations for each game.
📈
Real-Time Odds Analysis
We monitor odds across multiple major sportsbooks in real time. Our system flags discrepancies between our model's implied probability and market prices, identifying situations where the public or sharp action has created exploitable inefficiencies.
🎲
Kelly Criterion Stake Sizing
Every pick includes a recommended stake size based on the Kelly Criterion - a mathematically optimal bet-sizing formula that maximizes long-run bankroll growth while accounting for edge and uncertainty. We apply a fractional Kelly approach (typically 25-50%) to manage variance and protect against model error.
📍
Closing Line Value (CLV) Tracking
CLV is the gold standard metric for evaluating betting performance. We track CLV for every pick - comparing the odds at time of pick publication to the closing odds just before game time. Consistent positive CLV is the strongest indicator of genuine edge, and it's a metric we hold ourselves accountable to.
🔥
Feature Engineering Pipeline
Our models ingest hundreds of features per game: team and player statistics, rest days, home/away splits, weather data (for outdoor sports), injury reports, line movement patterns, public betting percentages, and more. The pipeline updates automatically as new data becomes available before tip-off or game time.
The Team
Sibyls Edge is built by a cross-disciplinary team with deep expertise in the fields that matter most for sports prediction.
Data Science Team
Our ML engineers and data scientists bring backgrounds in quantitative finance, applied statistics, and sports analytics. They design, train, and validate our prediction models.
Sports Analysts
Former sports journalists, professional bettors, and lifelong sports fans who understand the game beyond the numbers. They provide context and qualitative insight the models can't capture.
Engineering Team
Full-stack engineers responsible for the data pipelines, real-time odds ingestion infrastructure, web application, and the systems that keep everything running reliably.
Our Edge
The question everyone should ask of any prediction service is: why does your edge exist, and is it sustainable? Here's our honest answer.
1. We model what the market misprices
Opening lines are set by bookmakers to balance action, not necessarily to reflect true probability. Sharp money moves lines toward fair value, but there's often a window - particularly for less-followed games - where inefficiencies exist. Our models are calibrated to find these windows.
2. We don't bet against the closing line
Our approach is rooted in beating the closing line, not just recording wins. A team can win 55% of games and lose money long-term if they're taking bad prices. We focus on value - situations where our model's probability exceeds the implied odds by a meaningful margin.
3. Disciplined stake sizing prevents ruin
Even a genuinely profitable system can go bankrupt with poor bankroll management. Our Kelly-based sizing recommendations ensure that variance is manageable and that a losing streak doesn't wipe out accumulated gains.
4. We track everything honestly
Every pick is timestamped at publication. Every result is recorded. We don't cherry-pick records or omit losing picks from our performance summaries. Transparency is non-negotiable. Our CLV and win rate figures are calculated from the complete, unedited pick history.
OUR MISSION
"To give every sports bettor access to the same quality of analytical tools and information that institutional and professional bettors have always had."
We're building toward a future where sharp betting is accessible, transparent, and grounded in honest data - not hype, hot takes, or manufactured winning streaks. The numbers either hold up or they don't. We believe ours do.
Disclaimer: All performance statistics (62% win rate, +8.4% average edge) represent historical model backtesting results. Past performance does not guarantee future results. Sports betting involves risk. Never bet more than you can afford to lose.