AngelRace turns race data into explainable rankings, calibrated win probabilities and interactive Monte Carlo simulations — built for people who want to understand a race, not just be told what to think about it.
This gauge recomputes live from the AngelScore engine in your browser — the same weighted feature model described in the blueprint, running on synthetic demo data.
AngelScore is a transparent 0–100 competitive profile built from eleven weighted feature families — never a black box, never labeled as a guarantee.
AngelWin considers every runner together, so displayed win probabilities stay field-consistent — around 100% after scratches and updates.
AngelSim runs reproducible Monte Carlo simulations with stored seeds, so you can explore scenarios — scratches, pace shape, odds moves — before post time.
A 0–100 competitive profile from eleven feature families — horse performance, speed & pace, class, distance fit, surface, connections, form and post dynamics. Coverage falls, confidence falls with it.
A calibrated probability of winning within the current field, produced by comparing every runner's profile at once rather than scoring horses in isolation.
Running styles are classified E / E/P / P / S to project race shape, then AngelSim samples thousands of correlated outcomes to show a distribution, not a single guess.
Model probability versus market-implied probability, timestamped and refreshable — framed as Potential Value, never as a safe bet.
AngelRace ships as a virtual simulation and analytics product. AngelBet Lab never moves real money — bankrolls, stakes and payouts are entirely virtual, for understanding expected value and variance.
Scores, probabilities and simulated results are model output, not promises. Gamification rewards learning and calibration — never spend, streaks of losses chased, or riskier betting behavior.
Limited daily race cards, baseline AngelScore, full Academy access.
Full factor breakdown, AngelWin, AngelValue, saved picks, richer charts.
10K AngelSim runs, AI analyst Q&A, advanced scenarios, backtesting.