Join — $29/month

The Accuracy Lab

Inspect what SLAITIQ published, when it was available and how much verified outcome evidence supports the record.

Slates with published records

9

Published projection snapshots

24

Player projection records across updates

7009

Verified results / evaluation players

30 / 1058

Projection accuracy by position

Evaluation uses the last eligible published projection before each player’s kickoff, one observation per player per slate, for players projected above zero. Both ingestion and publication must precede kickoff. A slate enters the accuracy table only when every player in that cohort has a verified result, including zero-point performances. Related slates can contain the same player; their observations are not independent games.

Complete cohorts only; lower absolute error is better
Model versionPositionPlayersMAEMedian absolute errorBiasMatched consensus MAE
SLAITIQ research v1K24.484.48-1.25Unavailable
SLAITIQ research v1TE61.981.18-1.23Unavailable
SLAITIQ research v1DST21.871.87-0.09Unavailable
SLAITIQ research v1WR103.482.03-2.33Unavailable
SLAITIQ research v1QB24.964.96-3.67Unavailable
SLAITIQ research v1RB83.121.59-0.91Unavailable

MAE is mean absolute projection error in fantasy points. Bias is prediction minus actual. Consensus comparisons require the same players and timestamps and a preserved baseline; a missing historical baseline is never reconstructed using later information. Legacy versions remain separate from current models.

Result verification

The first complete cohort is Pittsburgh at Cleveland on October 1, 2026. Base fantasy points were calculated from the official NFL final gamebook using DraftKings Showdown scoring, without the Captain multiplier. Zero and negative performances are retained. This is a small historical sample from an earlier model, not validation of the current model or a profitability claim.

Timestamped publication ledger

7009 pre-kickoff player projection records are preserved across these snapshots. Drafts and ownership-only copies are excluded. Repeated updates are audit history, not additional independent accuracy samples. Raw input files and provider estimates remain private.

Every preserved published projection snapshot, including earlier versions
SlateVersionPublished (Central)PlayersPre-kickoff records
2026 · Week 5 · TB @ DAL
2026-10-08
SLAITIQ NFL v1.15252
2026 · Week 5 · TB @ DAL
2026-10-08
SLAITIQ NFL v1.15252
2026 · Week 4 · ATL @ NO
2026-10-05
SLAITIQ NFL v1.15656
2026 · Week 4 · ATL @ NO
2026-10-05
SLAITIQ NFL v1.15656
2026 · Week 4 · ATL @ NO
2026-10-05
SLAITIQ NFL v1.15656
2026 · Week 4 · ATL @ NO
2026-10-05
SLAITIQ NFL v1.15656
2026 · Week 4 · ATL @ NO
2026-10-05
SLAITIQ anchor v15656
2026 · Week 4 · DET @ CAR
2026-10-04
SLAITIQ anchor v15050
2026 · Week 4 · KC @ LV
2026-10-04
SLAITIQ anchor v15858
2026 · Week 4 Afternoon Only
2026-10-04
SLAITIQ anchor v1214214
2026 · Week 4 Sunday Main
2026-10-04
SLAITIQ anchor v1629629
2026 · Week 4 Sunday Main
2026-10-04
SLAITIQ anchor v1629629
2026 · Week 4 Sunday Main
2026-10-04
SLAITIQ anchor v1629629
2026 · Week 4 Sunday Main
2026-10-04
SLAITIQ anchor v1629629
2026 · Week 4 · PIT @ CLE
2026-10-01
SLAITIQ research v15151
2026 · Week 4 · PIT @ CLE
2026-10-01
SLAITIQ anchor v15151
2026 · Week 4 · PIT @ CLE
2026-10-01
SLAITIQ anchor v15151
2026 · Week 4 · PIT @ CLE
2026-10-01
SLAITIQ anchor v15151
2026 · Week 3 Afternoon Only
2026-09-27
SLAITIQ median v1213213
2026 · Week 3 Sunday Main
2026-09-27
SLAITIQ median v1674674
2026 · Week 3 Sunday Main
2026-09-27
SLAITIQ median v1674674
2026 · Week 3 Sunday Main
2026-09-27
SLAITIQ median v1674674
2026 · Week 3 Sunday Main
2026-09-27
SLAITIQ median v1674674
2026 · Week 3 Sunday Main
2026-09-27
SLAITIQ median v1674674

What this record can establish

Publication timestamps establish what was available before play. Verified actuals establish prediction error. Neither a founder’s winnings nor an optimizer’s legality tests establish product profitability. Corrections to actual results retain their source and prior record.

Model methodology · Founder evidence · Open the live desk