Model comparisons
Build and compare models
Compare model forecasts with market prices and review graded games. Open the builder, Lab, your models or a remix.
Where we differ from the market
1 house plays · Week 6 · Sun, Oct 18 – Mon, Oct 19
Prices as of Oct 18, 4:57 PM ET
How to read the board
Model shows the published confidence. Gap shows the published model estimate minus the opening market estimate, in points. For moneylines, these are probability percentage points. A spread or total confidence is a score out of 100, not a win probability. A gap is a disagreement between estimates; it does not establish betting advantage. Best price is the best book quote at the play's line, or the price at pick time. Plays keep their order by model confidence, with ties by kickoff.
| # | Game | Play | Model | Gap | Best price |
|---|---|---|---|---|---|
| 1 | WAS @ SF | SF −7.5 | confidence 63/100 | +11.9 pp vs market | −103LowVig at −7.5 |
Swipe for model, gap and best price.
Sharp money, latest
2 games· as of Oct 18, 4:25 PM ET
| Game | Side | Tickets | Money | Open → now | Read |
|---|---|---|---|---|---|
| DAL @ GB | GB | 40% | 41.5% | -3 → -1.5 | No clear signal |
| WAS @ SF | SF | 79% | 84% | -5.5 → -7.5 | No clear signal |
Money share leads ticket share on the same side, checked against the opening and current line.
Where to go next
- Lab — ask the question first
Pick a catalogue question and read its sample, baseline and lift before you build anything. Answers are precomputed, so the number you see is the number that was graded.
- Bench — assemble the model
Snap bricks together into a model, see what each one contributes, and backtest it walk-forward against the same settlement ledger the house is graded on.
- My models — what you own
Every model you have built, with its record, why it passed or failed in words, and whether it is currently clearing breakeven.
- Remixes — start from someone else
Remix a published model, change one thing, and keep the lineage so the two records can be compared honestly.
A model's probability is a probability at a price, not a promise. Early samples are marked as early.