Pocket Signal combines multiple predictive models to turn NFL data into easy-to-read game forecasts and player outlooks.
Instead of asking one model to predict everything, Pocket Signal uses separate models for different parts of the game.
The result is a forecast that shows who is more likely to win, the expected scoring margin, and the expected total points.
Team performance, recent form, quarterback play, matchup statistics and other pregame information.
Separate models evaluate the winner, scoring margin and projected total.
The outputs are combined into one matchup forecast that users can explore.
Each model answers a different question about the game.
The winner model estimates each team's probability of winning the game.
The margin model estimates the expected point difference between the two teams.
The total model estimates the combined scoring output expected from both teams.
Pocket Signal uses the projected margin and projected total together to produce a coherent final-score projection.
This keeps the displayed score consistent with both the margin and total forecasts.
Signal strength represents confidence in the winner forecast — not a guarantee of the result.
The winner model sees a larger separation between the two teams.
Pocket Signal has a clear preference, but there is still meaningful uncertainty.
The matchup is closer and the winner forecast is more uncertain.
The winner and margin forecasts are produced by separate models, so close games can occasionally produce different answers.
Pocket Signal does not force the models to agree. The disagreement stays visible so users can see the additional uncertainty in the matchup.
Historically, the winner model has been slightly more accurate at selecting the outright winner, but the difference is small.
Market information provides another reference point for understanding a matchup.
Pocket Signal can compare its estimated win probability with the probability implied by the market.
The model's expected scoring margin can also be compared with the market's view of the matchup.
Explore weekly projections alongside recent performance, expected workload, opponent statistics and available sportsbook lines.
Quarterbacks have passing-yard, attempt and completion projections. Receivers and tight ends have receiving-yard, reception and target projections. Running backs have rushing-yard, receiving-yard and reception projections.
These are estimates, not exact outcomes. Consider other forecasts, injury updates and matchup information when forming your own outlook.
Blue chart bars show recorded game results; cyan shows the selected week's projection. Season and recent averages help you compare the forecast with the player's usual production.
Where available, the gold line marks the market benchmark. A player's past results against that line describe their history, not their probability of going over this week.
Compare what the opponent has allowed to the player's position across the season and its last three completed games. These figures combine production from all players at that position, not one individual.
The game environment adds projected team points, margin and total. These are supporting comparisons, not a breakdown of how the player model calculated its forecast.
Search by player or team, filter by position or game, and use the stars to build a watchlist in your browser. Sort by market gap to bring the largest projection-to-line differences to the top.
Availability and starting roles can change. Check current reports before relying on an outlook; a missing report does not confirm that a player is active.
The gap is Pocket Signal's projection minus the saved sportsbook line. Green means the projection is above the line; red means it is below.
Sorting by gap size uses the absolute difference, so a large negative gap can appear beside a large positive one. Across all positions, the main yardage stat is QB passing, RB rushing, or WR/TE receiving.
A larger gap is a larger disagreement, not a win probability or a guarantee. Players without a matching line appear last and show “No market line available yet.” Lines are snapshots and may have moved.
Explore Player Outlooks →The projection is 14.5 yards above the line.
The projection is 15.5 yards below the line.
The Track Record page shows results for the saved weekly Top 15 market-gap selections. If the projection was above the line, the player needs to finish above it; if it was below, the player needs to finish below it.
Hit rate excludes pending results, pushes and voids. Hypothetical units use the recorded odds and one unit risked per selection. Reconstructed archives are labeled separately from pregame snapshots.
View Player Market Gap Results →Pocket Signal preserves its original pregame forecasts before the result is known.
After the game, those forecasts are graded for winner accuracy, margin error and total error on the Track Record page.
View Track Record →Pocket Signal is designed to show what the models think, how confident they are, where they disagree, and how those forecasts perform over time.
Explore This Week's Games