Round Tipping
Leaderboard →
You're building the User Model
Your predicted scoreline is treated as a simulation — the same way the BS Machine runs thousands of simulations to generate its probability distributions. Across all users, your collective picks form a crowd model that we pit against the BS Machine. The more accurate your scoreline prediction, the more your simulation contributes to beating the machine.
How to use the sliders
Use the Margin slider to pick the winner and by how much — slide left for a home win, right for an away win. Use the Total Points slider to predict the combined score — it defaults to this season's average. The scores update automatically.
User vs User scoring
Each game is worth up to 10 points — margin and total are scored independently, each out of 5. Scoring follows a bell curve with a standard deviation of 6 points — one converted try. A perfect margin prediction scores 5, a perfect total scores 5. Being one converted try off still scores well. Two tries off and you're near zero.
Perfect → 5 pts
±6 pts error → ~3 pts
±12 pts error → ~0.7 pts
User Model vs Machine Model
After each round, the crowd's picks are compared against the BS Machine's predictions using probability scoring. For each game, if the BS Machine assigned a 36% chance to the actual result, it scores 0.36. If the crowd model implied a 45% chance, it scores 0.45. Scores accumulate across all games — whoever is closer to 1.0 per game wins the round.
Picks lock at kickoff. You can update your prediction for any game right up until the ball is kicked.