Track record

Full-field results include wrecks. Race scores and older lineup grades remain in the detailed report.

Scoring

DraftKings

Sample

67 races

Projected vs actual correlation

0.29

Baseline: Salary and odds comparisons are not published in this summary.

Before-lock record: These tests do not verify forecasts saved before lock.

NASCAR Xfinity results, kept separate from Cup.

Scoring

DraftKings

Sample

56 races

Projected vs actual correlation

0.38

Baseline: Salary and odds comparisons are not published in this summary.

Before-lock record: These tests do not verify forecasts saved before lock.

Weekly player projections are separate from game picks and draft rankings. Recent-season replays do not prove pre-lock capture.

Scoring

DraftKings Classic

Sample

18 holdout weeks

Projected vs actual correlation

0.68

Baseline: Top-player capture: model 65.0%; recent form 62.0%. This compares player selection, not salary-constrained lineup returns.

Before-lock record: These tests do not verify forecasts saved before lock.

These tests do not grade the ensemble currently served by the DraftKings optimizer.

Scoring

DraftKings / FanDuel

Sample

166 DK tournaments; 166 FD tournaments

DK projected vs actual correlation

0.32

FD projected vs actual correlation

0.32

Baseline: A salary or odds benchmark is not available in this report.

Before-lock record: These tests do not verify forecasts saved before lock.

The archive does not certify the exact model and odds served before each card. It does not establish an edge over moneyline rankings.

Scoring

DraftKings

Sample

143 cards

Projected vs actual correlation

0.28

Baseline: Top-player capture: model 67.0%; recent form 64.0%. This compares player selection, not salary-constrained lineup returns.

Before-lock record: These tests do not verify forecasts saved before lock.

The replay covers the earlier rate model and players who played. It does not validate current rotation and injury adjustments.

Scoring

DraftKings

Sample

541 slates

Projected vs actual correlation

0.75

Baseline: Top-player capture: model 86.0%; recent form 85.0%. This compares player selection, not salary-constrained lineup returns.

Before-lock record: These tests do not verify forecasts saved before lock.

Earlier figures used incomplete scoring. They stay withheld until revalidation is complete.

Scoring

DraftKings

Sample

Figures withheld

Baseline: Current-scoring validation is required before publishing comparisons.

Before-lock record: These tests do not verify forecasts saved before lock.

The optimizer and its detailed evaluation remain in preview.

Scoring

Not publicly validated

Sample

No public DFS record

Baseline: Not published.

Before-lock record: No public forward DFS record is claimed.