Reading source coverage
Passes thrown to him, caught or not.
The passes nflverse’s play-by-play names him as the intended receiver ontargetsThe share of his team’s targets that went to him.
Targets ÷ team targetstarget_shareHow far downfield the passes thrown to him travelled, from the line of scrimmage to where the ball reached him, added up over every target, caught or not. A throw behind the line counts as negative.
The sum of the air yards on his targetsair_yardsThe share of his team’s air yards that were thrown to him: how much of the downfield passing game ran through him.
Air yards ÷ team air yardsair_yards_shareHow far downfield his targets were thrown, on average.
Air yards ÷ targetsadotTarget share and air yards share in one number, with targets counting for more. Josh Hermsmeyer’s formula.
1.5 × target share + 0.7 × air yards sharewoprReceiving yards for each air yard thrown his way. A short-area receiver who gains yards after the catch scores high; a deep threat who catches fewer of his targets scores low. It swings a lot from week to week.
Receiving yards ÷ air yardsracrHis rushing attempts.
The runs nflverse’s play-by-play credits to himcarriesThe share of his team’s carries that he took.
Carries ÷ team carriescarry_shareHis receiving yards measured against how often his team threw. It gives credit to a receiver who produces in an offense that doesn’t pass much.
Receiving yards ÷ team pass attemptsyards_per_team_pass_attemptFantasy points with a point for every catch.
Standard points + 1 per catchpprFantasy points with half a point for every catch.
Standard points + 0.5 per catchhalf_pprFantasy points with nothing for catches, as nflverse scores them.
Passing yards ÷ 25 + 4 per passing TD − 2 per interception + rushing and receiving yards ÷ 10 + 6 per rushing, receiving or return TD + 2 per two-point conversion − 2 per fumble loststandardPass attempts, completions, passing yards and touchdowns, interceptions, rushing yards and touchdowns, catches, receiving yards and touchdowns, and fumbles lost are nflverse's own counts for the game, added up over a season. Pass attempts don't include sacks.
The share of his team’s targets and carries inside the opponent’s 20 that went to him.
(targets + carries inside the 20) ÷ team targets and carries inside the 20red_zone_sharePasses thrown to him on snaps inside the opponent’s 10.
Targets with the line of scrimmage at the 10 or closertargets_inside_10The share of his team’s targets inside the opponent’s 10 that went to him: who gets the ball when a touchdown is close.
Targets inside the 10 ÷ team targets inside the 10target_share_inside_10His runs on snaps inside the opponent’s 10.
Carries with the line of scrimmage at the 10 or closercarries_inside_10The share of his team’s carries inside the opponent’s 10 that he took.
Carries inside the 10 ÷ team carries inside the 10carry_share_inside_10Passes thrown to him on snaps inside the opponent’s 5.
Targets with the line of scrimmage at the 5 or closertargets_inside_5The share of his team’s targets inside the opponent’s 5 that went to him.
Targets inside the 5 ÷ team targets inside the 5target_share_inside_5His runs on snaps inside the opponent’s 5.
Carries with the line of scrimmage at the 5 or closercarries_inside_5The share of his team’s carries inside the opponent’s 5 that he took: the goal-line back.
Carries inside the 5 ÷ team carries inside the 5carry_share_inside_5Passes thrown to him into the end zone.
Targets whose air yards reach the goal line (air yards ≥ yards to go to the end zone)end_zone_targetsTargets plus carries inside the 10: the two kinds of work worth the most fantasy points.
Targets + carries inside the 10high_value_touchesWhat a league-average player would have scored, full PPR, with exactly his targets and carries: each one valued by where on the field it came and how far the ball was thrown.
The sum over his targets and carries of the model’s expected catches, yards and touchdowns, scored 1 per catch, 0.1 per yard and 6 per touchdownxfp_pprPoints he scored on his targets and carries, minus what they were expected to be worth: who is beating his role, and who is leaving points behind.
Actual points on his targets and carries − xFPfpoe_pprHis share of his team’s expected points: how much of the offense’s scoring chances ran through him.
xFP ÷ team xFPxfp_shareHow much each pass thrown his way raised his team’s expected points, on average. Above 0 helps the offense.
The sum of EPA on his targets ÷ his targetsepa_per_targetHow much each of his runs raised his team’s expected points, on average. Most backs are below 0: running usually loses value against passing.
The sum of EPA on his carries ÷ his carriesepa_per_carryThe share of his targets and carries that raised his team’s expected points.
Targets and carries with EPA > 0 ÷ targets and carriessuccess_rateThe share of his targets and carries that went for a big gain: a run of 10+ yards or a catch of 20+.
(runs of 10+ yards + catches of 20+ yards) ÷ (targets + carries)explosive_rateYards he gained after the catch beyond what nflfastR’s model expected from where and how he caught it, per catch.
(yards after catch − nflfastR’s expected YAC) ÷ catchesyac_over_expectedHow much each of his called passes raised his team’s expected points: throws, sacks and scrambles.
The sum of nflfastR’s QB EPA on his dropbacks ÷ his dropbacksepa_per_dropbackHow much more often his passes were completed than nflfastR’s model expected from each throw’s depth, spot and situation, in percentage points.
The average of nflfastR’s CPOE over his pass attemptscpoeThe share of his dropbacks that raised his team’s expected points.
Dropbacks with QB EPA > 0 ÷ dropbacksdropback_success_ratexFP asks what a league-average player would have scored with the same work. Every target and carry falls into one cell, and the cell's value is the average result of every play like it: catches, yards and touchdowns, so one model scores PPR, half-PPR and standard.
Model ledger-xfp-v1 (2006-2022, k=30). It is ours, fitted on nflverse play-by-play by a script anyone can rerun, so its numbers won't match another site's xFP. It uses nothing from Fantasy Points or any paid model.
On the team pages, from the same play-by-play. Neutral situations: win probability between 20% and 80%, the first three quarters, and outside the first half's last two minutes.
Offensive snaps that stood, a game: passes, sacks, scrambles and runs.
Plays ÷ gamesplays_per_gameHow often the offense called a pass, in every situation.
Dropbacks ÷ playspass_rateHow often the offense called a pass when the score didn’t force its hand.
Dropbacks ÷ plays, neutral situations onlyneutral_pass_rateHow often an average offense would have passed in the same neutral situations, by nflfastR’s model of down, distance, field position, time and score.
The average of nflfastR’s xpass over the same neutral playsexpected_pass_rateHow much more often the offense passed than an average offense would have in the same neutral situations. Above 0 leans pass.
Neutral pass rate − expected pass rate, on the same playsproeSeconds of game clock from one snap to the next, in neutral situations. Lower is faster.
Game-clock seconds between consecutive snaps of the same drive ÷ those gaps, starting from a neutral snapneutral_pace_seconds| Numbers | Seasons | Source | License |
|---|---|---|---|
| Passing | 1999 on | nflverse weekly player stats | CC BY 4.0 |
| Rushing | 1999 on | nflverse weekly player stats | CC BY 4.0 |
| Catches and receiving yards | 1999 on | nflverse weekly player stats | CC BY 4.0 |
| Targets | 1999–2002 and 2009 on | nflverse weekly player stats | CC BY 4.0 |
| Air yards | 2009 on | nflverse weekly player stats | CC BY 4.0 |
| Fantasy points | 1999 on | nflverse weekly player stats | CC BY 4.0 |
| Red zone, inside the 10 and inside the 5 | 1999 on | nflverse play-by-play | CC BY 4.0 |
| End-zone targets | 2006 on | nflverse play-by-play | CC BY 4.0 |
| Expected points (xFP) | 2006 on | nflverse play-by-play | CC BY 4.0 |
| EPA, success and explosive plays | 1999 on | nflverse play-by-play | CC BY 4.0 |
| YAC over expected and CPOE | 2006 on | nflverse play-by-play | CC BY 4.0 |
| Pass rate, neutral pass rate and pace | 1999 on | nflverse play-by-play | CC BY 4.0 |
| Pass rate over expected (PROE) | 2006 on | nflverse play-by-play | CC BY 4.0 |
We checked the gaps against nflverse's raw weekly files: 2003–2008 have under 70 targets a season against about 16,000 in the seasons either side, and 1999–2002 carry air yards on only 400 to 500 player-games a season (about 8,000 air yards, against 140,000 a season from 2009). A share built on that would look like a real number and mean nothing, so we leave it out.
nflverse publishes its own target share, air yards share, WOPR and RACR in the same file. Theirs divide by every target on the team; ours divide by the targets to its quarterbacks, running backs, receivers and tight ends, which leaves out the odd throw to a lineman or a defender (under 100 a season). In 2012, our target share was within half a percentage point of nflverse's in 99% of player-games, and our air yards share in 93%. We compute our own so a season's share has a defined denominator (his team's totals in the games he played) and so the CSV can show it.
#: what the table is, the credit and license, and a link here. Spreadsheets show them as rows. In pandas, read it with comment="#"; in R, comment.char = "#".team_targets, team_air_yards, team_carries, team_pass_attempts), so any number can be checked.red_zone_share on, the columns come from play-by-play, with their own team totals (team_targets_inside_10, team_xfp_ppr and so on). The team pages have their own CSV, with the play counts behind each rate.First-read share, designed targets, catchable throws, drops, play action, RPO and blitz rates need FTN's charting (2022 on, CC BY-SA 4.0), which comes next in its own labelled section. Snap share and per-snap rates need nflverse's participation files (2016–2025, published after each season). Routes run (and anything per route) aren't in any free data: free participation data only records the route of the player who was thrown to, so anything we publish there will be labelled as a stand-in, per pass snap rather than per route.