Model summary unavailable
No backtest rows are loaded yet.
Sport
Stay in the same Linelabs flow and switch the live sport context.
Results & Transparency
How to interpret the current signal quality
This page is meant to make the latest loaded evaluation easier to trust. It summarizes what the current model runs support, where the edge still looks modest, and how the product should be framed today. The goal is not to show off every metric. It is to make the current read easier to interpret before you act.
Signal Performance
Data freshness unavailable
Latest loaded evaluation
Walk-Forward Accuracy
-
vs baseline -
Walk-Forward AUC
-
vs baseline -
Weeks / Games
-
Walk-forward only
Most Recent Model
-
Loading...
Current Proof State
How the latest scorecard supports current claims
These proof states are based on the current loaded board and latest available evaluation rows. When the data is thinner or less current, the claims step down automatically.
Current board coverage
Board-tier proof is loaded when current tier data is available.
Execution context
Execution-context proof is shown when the live board supports it.
Latest evaluation
Scorecard proof is only claimed when recent evaluation rows are available.
Surfaced Pick History
What recently settled surfaced picks show
This is separate from model evaluation. It summarizes recently settled picks that were actually surfaced on the board, with sample size and timing shown directly.
Posted-pick history is only shown when settled lifecycle data is available. It is not a tout board.
No settled surfaced-pick history loaded
Current Interpretation
What the latest evaluation supports
- In the latest loaded runs, totals currently read as the stronger model-backed surface.
- In the same runs, spreads still read better as market-led, price-aware leans.
- Freshness, line movement, and sportsbook context still matter as much as the raw signal.
Claim Safety
What we will not overstate
- We do not describe current signals as guaranteed winners.
- We do not treat spread output as strong standalone model alpha.
- We do not hide stale data behind generic confidence language.
Use This With The Board
Check freshness first
If the latest evaluation is stale, treat confidence as provisional.
Use spreads carefully
Spread picks work best as market-informed context, not as aggressive edge claims.
Lean more on totals
Totals currently carry more of the model-backed differentiation in the stack.
What To Do Next
This page is not a tout-style win-rate board. Use it to read the latest evaluation honestly, then go back to the live board to decide whether the current number is still worth betting.
Read lane quality first
Use the scorecard to understand which lane currently looks stronger before you weigh any individual pick.
Keep the claim level modest
A small edge over baseline can still matter, but it should not be framed like dominant model alpha.
Pair proof with price context
Results tells you how to read the lane. The live board still decides whether the current number is worth playing.