Sports analytics has a jargon problem: half the words get used loosely and the other half get used to sound smart. Here's what we mean, exactly, when we use each of these — so nothing on a MatchWiz board should ever read like code. Link to any term directly with its anchor.
The model's language
Matchup score
Our 0–100 rating of a player's spot on a given day — how a hitter profiles against tonight's starter and bullpen (or a pitcher against tonight's lineup), with park, weather, playing time, and recent form folded in. Higher means a better spot today; it's a rating of the matchup, not of the player's talent. See: How MatchWiz works.
Projection
The model's expected number for a specific stat in a specific game — 5.8 strikeouts, 4.3 team runs. Every projection is posted before the game and graded against the real box score after it, at the line the market offered.
Edge / gap (Δ)
The distance between our projection and the market's posted line. A starter projected for 6.9 strikeouts against a 5.5 line has a 1.4-K gap. The gap — not the raw projection — is the signal: it measures how much the model and the market disagree. See: Plays of the Day.
Lean
The side of a line our number lands on — over or under, NRFI or YRFI — before any judgment about whether the gap is big enough to matter. Every board shows its leans; only backtest-qualified gaps graduate into plays.
Play of the Day
The single spot on a board where our projection diverges furthest from the book's line — far enough that the same disagreement was profitable in the season backtest. One per market, posted only after odds are in and the lineup is final, then graded at the posted line. See: Plays of the Day board.
Graded / the ledger
What happens to every number we publish: after the game, the projection is scored against the real result at the price that was actually posted, and the outcome is logged win, loss, or push. The season records on every board come from that ledger — misses included.
Locked at first pitch
Once a game starts, its projections, lines, and leans can't change on our boards. Nothing gets quietly rewritten to look smarter after the fact — what you see on an archived date is what we said before the game.
Backtest
Replaying a strategy against past games to see if it would have made money. We only surface plays from patterns that graded profitable over a full-season backtest — and we validate on held-out data the model didn't tune on, so it can't just memorize the past.
Handedness splits
A player's numbers separated by the throwing hand he's facing — lefty bats vs righties, and so on. The single biggest matchup input in baseball: plenty of hitters are stars against one hand and liabilities against the other.
Park factor
How much a ballpark inflates or suppresses scoring and homers relative to a neutral field. Coors Field turns fly balls into souvenirs; Oracle Park eats them. Projections shift accordingly.
Units
A stake-sized measuring stick — one unit is whatever a standard bet means to you. Records in units (+8.2u) are comparable across bettors and bankrolls in a way dollar figures aren't. We grade every play in units at the posted price. See: Responsible gaming.
Flat 1u vs ★ recommended units
Two ways we grade the same plays. Flat = one unit on everything, the honest baseline. Recommended = each play sized 1–3 units by the size of its gap, because bigger disagreements have historically outperformed. Both records are posted so you can compare.
ROI
Return on investment: units won divided by units staked, at the posted prices. A +6% ROI means every 100 units risked returned 106. Hit rate tells you how often; ROI tells you whether it actually paid after the vig.
Hit rate
The percentage of graded calls that landed on the right side. Useful, but incomplete on its own — a 55% hit rate can lose money at bad prices, and a 45% one can profit at long odds. Always read it next to ROI.
Market terms
Moneyline
A bet on which team wins, priced in American odds: −150 means risk 150 to win 100 (the favorite), +130 means risk 100 to win 130 (the underdog). The sharpest, most efficiently priced market in baseball. See: Moneyline & totals board.
Implied probability
The win chance a price is charging you for. Negative line: odds ÷ (odds + 100), so −150 implies 60%. Positive line: 100 ÷ (odds + 100), so +130 implies about 43.5%. Both sides of a game sum past 100% — the excess is the book's margin.
Fair odds
The price a probability deserves with no vig in it. If a first inning is truly 58% to stay scoreless, fair NRFI odds are about −138. We post fair odds beside the market's price so the gap between them is visible at a glance. See: NRFI board.
Vig / juice
The book's built-in margin — the reason both sides of a coin flip are −110 instead of +100. It's the house edge you pay on every bet, and the hurdle any model has to clear before a right call becomes a profitable one.
Line
The posted number you're betting against — a strikeout total of 5.5, a team total of 4.5, a moneyline price. Lines move as money and news arrive, which is why every play we grade records the line that was actually up when it surfaced.
Closing line & CLV
The final line before first pitch — the market's most informed number. Closing line value (CLV) means consistently beating that final number, which sharp bettors treat as the best early proof of a real edge. We grade at the posted line instead, because that's the number you could actually have taken.
Prop (proposition bet)
A bet on a player stat instead of the game result — hits, home runs, strikeouts, total bases. Props are priced with less money and attention than game lines, which is why model-vs-market gaps show up there more often. See: Batter props guide · Pitcher props guide.
Team total
The over/under on one club's runs, separate from the full game total. A corner market that gets set off the game total and less attention after that — which is exactly where our per-club run projections have found their edge. See: Team totals board.
Game total (over/under)
The over/under on both teams' combined runs. Efficient and sharply managed — we publish our projected total beside it as transparency rather than pretending at an edge the grades don't support.
NRFI / YRFI
No Runs First Inning / Yes Runs First Inning — a bet on whether the 1st inning stays scoreless. Only the top of each order bats, so it's really a leadoff-hitters-versus-starter matchup that settles in about twenty minutes. See: NRFI board & guide.
Push
A tie with the line — the pitcher lands exactly on 6 strikeouts against a 6.0 line. Stakes are returned, and pushes sit in their own column in our records rather than counting as wins or losses.
Parlay
One bet that needs several legs to all win. The payout multiplies, but so does the book's margin — you pay vig on every leg at once, which is why parlays are the most profitable product books offer. Them, not you.
Chalk
The heavy favorite, or in DFS the highest-owned obvious play. Betting chalk isn't wrong — paying the wrong price for it is. In tournaments, chalk that fails is what separates fields.
“Lock”
A bet sold as unlosable. No such thing exists — baseball's best teams lose sixty times a year, and anyone selling certainty is selling entertainment. It's the one term on this page we hope you only ever see in quotes. See: Why we won't sell you a lock.
K-gap
Our shorthand for the strikeout edge: the distance between our lineup-adjusted strikeout projection for a starter and the book's posted K line. The largest backtest-qualified K-gaps become strikeout Plays of the Day. See: Strikeout board.
DFS terms
DFS (daily fantasy sports)
Contests where you draft a fresh salary-capped lineup every slate and score points from real box-score stats. No season-long commitment — every day is a new draft, which is why daily projections matter so much. See: MLB DFS guide.
Salary / value (points per $1K)
Every DFS player costs salary from a fixed cap, so raw projected points aren't enough — what matters is points per thousand dollars of salary. A 9-point hitter at $3,800 beats a 10-point hitter at $5,600. Value is the only stat that builds full lineups. See: Lineup optimizer.
Cash game
DFS formats where roughly half the field profits — 50/50s, double-ups, head-to-heads. You're trying to beat the median score, so cash lineups favor floor: steady projections, confirmed lineup spots, no heroics.
GPP (tournament)
Guaranteed prize pool tournaments — huge fields, top-heavy payouts. Beating the median isn't enough; you need a near-perfect lineup, so GPPs reward ceiling and being different from the crowd, not safety.
Stack
Rostering several hitters from the same team's lineup. Baseball scoring is bursty — a big inning pays the whole group at once, since runs and RBIs chain through consecutive hitters. Stacking is the single biggest structural lever in MLB DFS. See: Stacking, explained.
Ownership / leverage
The share of the field rostering a player. In cash games ownership barely matters; in GPPs it's everything — a great game from a 40%-owned star moves you nowhere, while the same game from a 5%-owned bat passes a third of the field.
Lineup lock
The moment contests close and rosters freeze — typically the slate's first pitch. The most common avoidable DFS loss is rostering a player who isn't in the day's actual lineup, which is why our boards flag confirmed lineups all morning. See: DFS projections board.
Missing a term?
If something on a board still reads like code, that's a bug in our writing, not your reading. Tell us and we'll add it here.
Definitions describe how MatchWiz uses each term. Research signals, not betting advice — 21+, play responsibly (1-800-GAMBLER).