Our Prediction Models
Data-driven models that identify mispriced contracts across prediction markets. Each model is backtested and continuously refined.
Active Models
Daily High Temperature
LIVEPredicts daily high temperatures for major US cities, identifying mispriced contracts on Kalshi temperature markets.
CITY COVERAGE
MODEL FEATURES
- Real-time NWS data integration
- Historical pattern analysis
- Multi-bracket probability modeling
- Automated signal generation
HOW IT WORKS
Our proprietary temperature forecasting model combines multiple data sources including NWS forecasts, historical patterns, and real-time atmospheric data to generate probability distributions for daily high temperatures. We compare these probabilities against Kalshi market prices to identify trading opportunities.
Daily Low Temperature
LIVEPredicts daily low temperatures for major US cities, targeting overnight and early morning temperature brackets on Kalshi.
CITY COVERAGE
MODEL FEATURES
- Overnight cooling pattern analysis
- Cloud cover & radiative cooling modeling
- Wind and humidity factor integration
- Early morning signal generation
HOW IT WORKS
Our low temperature model specializes in predicting overnight minimums, which are influenced by different atmospheric factors than daily highs. By analyzing cloud cover, wind patterns, humidity, and radiative cooling conditions, we identify mispriced contracts in the daily low temperature markets.
Pro Basketball Player Props
LIVEPredicts player prop outcomes (points, rebounds, assists, etc.) for pro basketball games, identifying mispriced contracts on Kalshi player prop markets.
PROP TYPES COVERED
MODEL FEATURES
- Recent form & season average blending
- Opponent defensive rating adjustments
- Rest day & back-to-back penalties
- Pace & blowout risk factors
- Real-time market price integration
HOW IT WORKS
Our basketball player props model analyzes player performance data including recent form (last 5 and 15 games), season averages, matchup-specific defensive ratings, rest days, home/away splits, pace factors, and blowout risk to generate probability distributions for various statistical outcomes. We compare these against Kalshi market prices to identify high-edge trading opportunities.
Pro Hockey Game Totals
LIVEPredicts game totals (over/under) for pro hockey games, with heavy emphasis on goalie analysis and rest factors.
GOALIE TIER SYSTEM
MODEL FEATURES
- Goalie tier classification system
- Back-to-back rest penalties (+12% GA)
- Team goals for/against rates (GF60, GA60)
- Pace and schedule adjustments
- Goalie confirmation tracking
HOW IT WORKS
Our hockey game totals model is fundamentally goalie-centric. We analyze goalie tiers (elite to emergency), save percentages, back-to-back situations, and rest factors alongside team offensive and defensive metrics. Goalie rest—especially B2B second-night starts—is the biggest edge source, with goalies allowing ~12% more goals on no rest.
Golf Tournament Picks
LIVEPredicts tournament outcomes (winner, top 5/10/20, make cut) for PGA Tour events using Strokes Gained analysis and course fit modeling.
MARKET TYPES
MODEL FEATURES
- Strokes Gained skill decomposition (OTT, APP, ARG, PUTT)
- Course-specific skill weight profiles
- Player-course fit scoring (0-100%)
- Multiple prop types (Winner through Make Cut)
- Configurable SG vs course fit blend
HOW IT WORKS
Our golf model blends two key factors: overall player skill (Strokes Gained Total) and course-specific fit. We decompose player skills into four components—driving, approach, short game, and putting—then weight them based on each course's demands. A bomber may dominate TPC Scottsdale but struggle at Augusta's precision layout. The 70/30 SG-to-course-fit blend captures both raw talent and situational edge.
Build Your Own Custom Model
Fine-Tune Model Parameters to Match Your Strategy
Our models come with sensible defaults, but every trader has different risk tolerances and market views. With Model Customization, you can adjust over 40 parameters to create a personalized prediction model that aligns with your trading style.
Adjust Confidence Thresholds
Set your own edge requirements for low, medium, and high confidence signals.
Tune Forecast Weights
Control how much weight is given to different weather models (HRRR, GFS, ECMWF).
Backtest Your Configuration
Run historical backtests to see how your custom parameters would have performed.
CUSTOMIZABLE PARAMETERS
Save Multiple Configurations
Create and save multiple custom configurations to test different strategies side by side.
Coming Soon
Baseball Player Props
Player prop predictions for MLB including strikeouts, hits, and home runs
Crypto Price Targets
Daily and weekly price threshold predictions for BTC and ETH
Election Forecasts
State and national election probability modeling
Want to See Our Models in Action?
Sign up for free to access live signals from our temperature prediction model and get notified when new models launch.
Go to Dashboard