Glossary: what every column means

Every number on PropSims comes from 10,000 simulations of each game, compared with posted lines where they exist. These are the terms behind the column headers. Tap or hover the small next to any header on the site to see the same definition.

Props · Ranges · Fantasy · DFS · Basics · Track record

Props

Line: The number the sportsbook or pick'em app sets for a stat. You bet on whether the player finishes over or under it.
Prop: The stat the bet is about, such as receiving yards, receptions or an anytime touchdown.
Price (odds): American odds for each side. -110 means you bet $110 to win $100; +120 means a $100 bet wins $120.
Book: Where the line comes from (a sportsbook such as FanDuel, or a pick'em app such as Underdog or PrizePicks), with its price when it has one.
Median / Our projection: Our middle outcome: in half of our 10,000 simulated games the player finishes above this number, and in half below it.
Mean / Average: The average across all our simulated games. It is usually a bit higher than the median because a few huge games pull it up.
Gap vs line (SDs): How far our projection sits from the line, measured in "typical swings" of our simulations (standard deviations). Near 0 means the line is close to our number; +1 means a full typical swing above it.
Lean / Sim lean: Which side of the line our simulated median falls on. It describes our simulations; it is not a recommendation to bet.
Side: Over or under. On the board it is the side with the higher estimated EV at the posted price.
Sim probability (our chance): The share of our 10,000 simulated games in which that side wins, straight from the model with no market input. Raw model numbers are often too confident.
Fair probability (fair chance): Our best estimate of the real chance the bet wins: the sportsbook's no-vig chance, nudged by our simulations by an amount fitted on past seasons. It stays close to the market because our model alone has not proven it beats the market.
Market / No-vig probability: The sportsbook's implied chance with its built-in cut (the "vig") removed, so over and under add up to 100%. For pick'em apps it is the win rate you need to break even.
EV (estimated expected value): Roughly how much you would win or lose per $1 bet over the long run if our fair chance is right. +5% means about 5 cents per dollar; small positive EVs are estimates, not proven edges.
Edge: Our fair chance minus the market's no-vig chance. Positive means we think that side wins a little more often than its price implies.
Line time: When we captured this line. Lines move, so an older line may no longer be available at that number or price.
Stale line: The line was captured well before this update and may have moved or been taken down since.
Kickoff: Scheduled game start, in Eastern Time. Picks are only logged before kickoff.
# (row): Position in the current sort order (estimated EV by default). It is a list position, not a rating.

Ranges

Floor (p10): A bad-day outcome: only 1 of every 10 simulated games is worse than this.
p15: A below-average day: about 1 simulated game in 7 is worse than this. Paired with p85, the middle 70% of outcomes sit between them.
Ceiling (p85): A good-day outcome: only about 1 simulated game in 7 is better than this.
Ceiling (p90): A big-day outcome: only 1 simulated game in 10 is better than this.
p95: A blow-up game: only 1 simulated game in 20 is better than this.
Range: A bar showing the spread from floor (p10) to ceiling (p85), with the median marked. Wider means less predictable.

Fantasy

Scoring (PPR / Half-PPR / Standard): The fantasy format: PPR gives 1 point per catch, half-PPR 0.5 and standard 0. Everything else (yards, touchdowns) scores the same.
Rank: Where the player ranks at his position this week in that scoring format, by our projection (QB1 is the top quarterback).
Call (Start / Borderline / Sit): Our start/sit verdict for a 12-team league (1 QB, 2 RB, 2 WR, 1 TE, 1 flex), based on his rank at the position. Backups and limited-role players are always Sit.
Half-PPR / PPR / Std median: The player's median fantasy points in that scoring format: half of our simulated games are above it, half below.
Key stats (median): His most important stats for the week at the median of our simulations, such as passing yards for a QB or receiving yards for a WR.
Head-to-head win %: How often this player outscores the other one across the same simulated games, so shared-game effects are included. Ties count as half.
Backup / Limited role: His projection is near zero because of his depth-chart role; it only matters if the starter ahead of him is out.
Injury tag (Q / D / O): Official injury status at our last update: Q = questionable, D = doubtful, O = out. Out players are removed from the simulations.

DFS

Proj: The number the optimizer builds lineups with for this player: our mean, median or 85th percentile depending on your setting, or your own number if you typed one.
My proj: Type your own projection to replace ours for this player; the optimizer then uses yours.
Salary: The player's DFS price on FanDuel or DraftKings for this slate. Your lineup must fit under the salary cap.
Value (points per $1K): Projected fantasy points for every $1,000 of salary. Higher means more points for the money.
Own%: The estimated share of DFS lineups that will include this player. Some figures are our estimates rather than published numbers.
Exp (exposure): The share of your generated lineups that include this player.
Min % / Max %: Limits you set on how often a player may appear across all the lineups you generate.
Lock / Exclude: Lock forces a player into every lineup; Exclude keeps him out of all of them.
Slates: How many DFS slates list a salary for this player (FanDuel and DraftKings, classic and Showdown).

Basics

Player: The player, with his position and team. Click a name for his own page where one exists.
Pos: Position: QB quarterback, RB running back, WR wide receiver, TE tight end, D/DST team defense.
Team / Opp: The player's team and this week's opponent.
Week: NFL regular-season week.

Track record

Track record: Our public log of every positive-EV pick the board showed, written before kickoff and graded after the game, so anyone can check our results.
Logged: When the pick was written to our log (always before kickoff), or how many picks were logged.
Lines (source): Which group of lines the picks were graded against, for example sportsbooks or pick'em apps.
Over / Under split: How many of the logged picks were overs and how many were unders.
Avg. est. EV: The average estimated EV of those picks when they were logged. It is what we expected, not what happened.
Graded / n: How many picks (or log records) are in the group. Small samples swing a lot, so treat early results with caution.
Hit rate / Win rate: The share of graded picks that won, with pushes left out.
W-L-P: Wins, losses and pushes. A push is when the result lands exactly on the line and the stake is refunded.
Units: Profit or loss if you bet 1 unit (the same stake) on every pick at the logged price.
ROI: Return on investment: total profit divided by the total amount staked. +5% means 5 cents profit per dollar bet.
Actual: What the player actually recorded in the game.
Result: Whether the pick won, lost or pushed once the game was graded.
Close / CLV: The line just before kickoff. If the line moved toward our side after we logged the pick (closing line value), that is a common sign the pick was good.
Model P(over) / Predicted / Actual over rate: A calibration check: the chance we gave overs in each band (Predicted) next to how often overs actually hit. Close numbers mean our probabilities are honest.
Log file / Chain / SHA-256: A digital fingerprint of each log file. Every entry is chained to the one before, so editing a past pick would break the chain and change the fingerprint.

Model opinion, not advice. For entertainment and informational purposes only. 21+. How the model works · Track record