Plain-English definitions of the stats we cite on NBA game pages. If something on a bet page doesn't make sense, look it up here.
Each entry has the formula, the threshold we treat as "good," and one example so the abstract definition has something to grip.
🏀 The Four Factors — Dean Oliver's foundation
These four stats predict basketball wins better than anything else. Roughly ordered by importance: shooting > turnovers > rebounding > free throws.
Effective FG% eFG% target: 53%+
Field-goal percentage adjusted to credit 3-pointers properly.
eFG% = (FGM + 0.5 × 3PM) / FGA
Why it matters: regular FG% treats a 3 and a 2 identically. eFG% doesn't. It's the single best one-number shooting metric.
Example: 10/20 FG with 4/8 from 3 → eFG% = (10 + 0.5×4) / 20 = 60%, even though raw FG% is just 50%.
Turnover % TOV% target: under 13%
Turnovers per 100 possessions.
TOV% = TO / (FGA + 0.44 × FTA + TO)
Why it matters: a turnover is a possession with zero point value. Reducing TO% from 15% → 12% is worth roughly 3 points per game.
Offensive Rebound % OREB% target: 28%+
Percentage of available offensive rebounds the team grabs.
OREB% = OREB / (team OREB + opp DREB)
Why it matters: each offensive rebound is a "second chance" possession — worth ~1.1 points on average.
Free Throw Rate FTR target: 0.25+
Free-throw attempts divided by field-goal attempts.
FTR = FTA / FGA
Why it matters: getting to the line is high-EV. Teams that draw fouls also tend to stop the clock (good late) and put opponents in foul trouble.
Game-4 example: DET 12 FTA / 83 FGA = 0.14 — way under target. CLE 34 / 78 = 0.44 — elite. That's the game in one stat.
⚡ Pace + efficiency ratings
Pace PACE league avg ~100
Possessions per 48 minutes.
Why it matters: doesn't predict who wins, but predicts total scoring. Two efficient teams at high pace = total over. Two efficient teams at low pace = total under.
Offensive Rating ORtg elite: 115+
Points scored per 100 possessions.
Why it matters: separates "scoring a lot" from "scoring efficiently." A team scoring 110 at pace 90 is way better than one scoring 115 at pace 105.
Defensive Rating DRtg elite: under 110
Points allowed per 100 possessions.
Why it matters: defensive equivalent of ORtg. Combined with ORtg, gives Net Rating.
Net Rating NetRtg +5 = championship contender
ORtg minus DRtg. Point differential per 100 possessions.
Why it matters: best single number for "how good is this team." Strongly correlated with playoff success.
👤 Individual player stats
True Shooting % TS% elite: 60%+
Single-number scoring efficiency including 2s, 3s, and FTs.
TS% = PTS / (2 × (FGA + 0.44 × FTA))
Why it matters: a player at TS 55% on 20 shots is far more valuable than one at TS 45% on 25 shots. TS% beats raw FG%.
Usage Rate USG% primary option: 28%+
Percentage of team possessions a player uses (shot, turnover, or free throws).
Why it matters: tells you who's running the offense. High-usage low-efficiency is the bad combo. Player props anchor heavily on usage.
Plus/Minus ± noisy at small samples
Team's point differential while the player is on the floor.
Why it matters: useful in single games, less useful over a season (depends on lineups). Big +/- in a playoff game is a real signal.
Game-4 example: Mobley +30 (anchored both ends), Cunningham -23 (worst on team) — read those numbers and you know the game's shape.
Assist/Turnover Ratio AST/TO elite PG: 3.0+
Assists per turnover.
Why it matters: measures playmaking efficiency. A 2.0 is league average for guards. Under 1.5 is a problem.
🎯 Player Props — Reading the Different Game
Props are bets on individual player stats — points, rebounds, assists, threes, blocks, steals, combos (PRA). They're now the largest retail bet type by volume, and where books make their highest margins. The math is different from spreads/totals; the inputs are different too.
Common prop types
- Points — total points scored by the player.
- Rebounds — offensive + defensive rebounds combined.
- Assists — assists for the game.
- Threes made — 3-point field goals made.
- Pts+Reb+Ast (PRA) — combo prop, sum of all three.
- Steals + Blocks — often combined into one line.
- Double-double / Triple-double — yes/no markets.
Why props are different from game lines
- Higher house edge. Typical vig is ~4–5% per side (vs ~2–3% for spread/total). The book has more margin built in.
- Larger variance. One player's single-game outcome is noisier than a team's. A star can score 15 or 45 in the same matchup.
- Public-heavy markets. Most prop money is on the OVER of stars. Lines get shaded toward the under to compensate.
- More mispriced opportunities. Books can't price hundreds of props with the same care they put into the main game line.
- Sample size matters. 5–10 game samples are too small; lean on 20–30 game baselines + role context.
What stats inform a good prop bet
| Stat | Why it matters |
|---|
| Minutes projection | Foundation — no minutes, no production. Build this first. |
| Usage rate | % of team possessions the player uses. Sets ceiling. |
| Recent role (last 5–10 games) | Current trajectory, not season average. Books price this. |
| Position-specific opponent defense | Team DRtg hides matchup detail. Position split is what matters. |
| Pace | More possessions = more chances at every counting stat. |
| Home/road splits | Often huge differential, often mispriced. |
| Foul-trouble history | Bends minutes. Recent foul rate matters more than season. |
| Blowout risk | Garbage time crushes stars' production (subs in). |
| Series-specific adjustments | Defenses adapt by Game 4–5. Game-1 numbers stop applying. |
| Teammate injury status | Usage redistribution is the biggest prop-edge driver of the year. |
Pro tips for props
- Lean unders on stars — public-heavy overs shade lines low.
- Check alternate lines — sometimes a lower threshold has better EV than the standard one.
- Combo props (PRA) — books price correlations worse than single stats.
- First-half / first-quarter props — less efficient markets, less sharp attention.
- Defensive stat props — books price scoring better than blocks/steals/assists.
- Avoid Same-Game Parlays — exponentially negative EV by construction.
- Shop multiple books — prop lines vary more than game lines, sometimes by 10–20% in implied probability.
Full prop framework with the eight inputs we evaluate: prop analysis template →
Coming soon: MLB, NHL, NFL, Golf glossaries · Each game page links to its sport's glossary.