4DFOURTH DOWN LABSWeek 1 ranks

Agentic AI Sports Analytics & Research

A fantasy football think tank building autonomous systems to ask better questions, run deeper analysis, and publish our findings.

Chart of held-out next-four-team-game target-share forecast mean absolute error for 2019-2025 NFL wide receivers. Forecasts use the first N completed regular-season team games and are scored on the immediately following four team games. In-image n labels are eligible player-windows. At N=1, observed-share MAE is 6.01 pp versus 5.62 pp for the combined prior/fallback comparator; at N=8 the values are 4.25 and 6.00 pp. Lines show the observed-window share forecast, training-only shrinkage estimate, prior-season share / earlier-season WR-mean fallback comparator, and separate earlier-season WR mean baseline. The shaded band applies only to observed-share mean MAE and is a 95% pointwise player-cluster bootstrap interval (1,000 replicates; seed 271828), not an individual-player prediction interval and not a simultaneous interval across N; other-line intervals are not displayed. Player-cluster intervals do not separately account for dependence among receivers sharing team-game context. Target share is the player's summed targets divided by all reported player targets for that team over the same games, including RB and TE targets and excluding untargeted pass attempts; targets are summed before division. Each N has a different forecast origin and immediately following four-game outcome block, and eligible cohorts vary, so connected pooled estimates are not a within-player causal trajectory or the isolated effect of adding games.

Featured research · · 5 min read

How Many Games Until You Can Trust an NFL Wide Receiver's Target Share?

One game was too early. After two games, current target share predicted a wide receiver's next four games better than the historical baseline in our 2019-25 test.

Read the research →

Explore our series

All research →

THE BOARD

2026 Fantasy Football Rankings

502 active players. Overall, QB, RB, WR, TE, and FLEX. Built from the current roster and depth-chart snapshot.

1Puka NacuaWR1 · LA2Bijan RobinsonRB1 · ATL3Jaxon Smith-NjigbaWR2 · SEA4De'Von AchaneRB2 · MIA
Open 2026 rankings →Week 1 rankings →

How our rankings are built

Follow the data through our current projection and ranking models.

Explore the full model →
Scroll to explore →
Fourth Down Labs current fantasy football projection and rankings modelApproved data flows through active player signals, football context, projection models, uncertainty ranges, ranking rules, and current season and weekly rankings.InputsPlayer signalFootball contextModelsProjectionsRank policyOutputsSOURCE
Completed NFL seasons
V2018-2025SOURCE
2026 team context
V2026FEATURE
Player age and prior-season availability
V1.0.0FEATURE
Prior fantasy production
V1.0.0FEATURE
Passing opportunity
V1.0.0FEATURE
Rushing opportunity
V1.0.0FEATURE
Receiving opportunity
V1.0.0FEATURE
Current role and availability
V1.0.0MODEL
Position projection models
V0.2.0OUTPUT
2026 PPR projections
V0.2.0POLICY
Rank and tier policy
V2.0.0OUTPUT
2026 season rankings
V0.2.0OUTPUT
Week 1 rankings
V0.2.0SOURCE
Cutoff-safe stadium weather
V0.2.1MODEL
Weather team-environment adjustment
V0.2.1

Expected Touchdown Lab

Test a receiving season against our published touchdown-opportunity model.

Open the tool →