Learn
Understand every number
Every stat, label and method the site uses, grouped by topic.
97 terms
Basketball basics
Core counting stats and scoreboard ideas.
- PTSPointsPoints scored. PPG is points per game.
- REBReboundsMissed shots recovered. RPG is rebounds per game.
- ASTAssistsPasses that directly lead to a made basket. APG is assists per game.
- STLStealsDefensive takeaways credited as steals.
- BLKBlocksShots rejected by a defender.
- TOVTurnoversPossessions lost without a shot attempt. Raw turnovers per game measure volume. They rise with minutes, touches, and usage, and are not a skill grade by themselves.Learn
- MINMinutesPlaying time. MPG measures availability and role size, not shooting or defensive skill.
- DIFFPoint differentialAverage scoring margin (points scored minus points allowed). Positive means outscoring opponents.Learn
- +/−Plus/minusTeam scoring margin while the player is on the floor. Strongly affected by teammates and opponents.Learn
- PaceHow many possessions a team typically plays per game. Faster pace raises counting stats.Learn
- GPGames playedGames played this season.
Shooting
How shots become points, and how efficiently.
- FG%Field goal percentageMade field goals ÷ field goal attempts. Treats twos and threes the same.Learn
- 3P%Three-point percentageMade threes ÷ three-point attempts.Learn
- FT%Free throw percentageMade free throws divided by free throw attempts.
- TS%True shooting percentageScoring efficiency that accounts for 2-point shots, 3-point shots, and free throws.Learn
- eFG%Effective field goal percentageShooting efficiency that gives extra credit for made threes (without free throws).Learn
- 3PArThree-point attempt rateShare of field goal attempts that are threes.Learn
Usage & role
How often a player handles the offense.
- USG%Usage rateShare of team possessions a player uses (shots, free throws, turnovers) while on the floor. Higher usage means more on-ball responsibility, not automatically better play.Learn
- AST/TOAssist-to-turnover ratioAssists divided by turnovers. Higher usually means cleaner creation when the sample is adequate.Learn
- ORB%Offensive rebound percentageEstimated share of available offensive rebounds a player or team grabs.Learn
Team efficiency
Team scoring, defense, and margin.
- On/offTeam net rating with a player on the floor minus with him off. It follows the whole lineup, so teammates and opponents move it.Learn
- ORtgOffensive ratingPoints scored per 100 possessions. On player pages, season boards may supply only an approximate individual estimate. Missing ratings stay unavailable, never fabricated.Learn
- DRtgDefensive ratingPoints allowed per 100 possessions. Lower is better.Learn
- NETNet ratingORtg minus DRtg, scoring margin per 100 possessions.Learn
- Opp PPGOpponent points per gamePoints allowed per game. Lower is better on defense.Learn
Impact models
Third-party plus-minus style estimates (DARKO, RAPTOR).
- DARKOEstimated points per 100 possessions above or below an average player (predictive impact).Learn
- RAPTORRAPTOR (pts/100). Open historical data, blank for seasons after 2021-22 when it stopped being published.Learn
- O-RAPTOROffensive RAPTOR (points per 100).Learn
- D-RAPTORDefensive RAPTOR (points per 100).Learn
- WARRAPTOR wins above replacement for the season (minutes-scaled).Learn
- PERPlayer efficiency ratingAll-in-one box score rate. League average is about 15. Higher means more production per minute from counting stats.
- BPMBox plus/minusEstimated point impact per 100 possessions from box score stats. Zero is average.
- OBPMOffensive box plus/minusOffensive half of box plus/minus.
- DBPMDefensive box plus/minusDefensive half of box plus/minus.
- VORPValue over replacement playerWins added above a cheap replacement player, scaled to minutes played.
- WSWin sharesCredit for team wins from box score production. Roughly one win share equals about one win contributed.
- WS/48Win shares per 48Win shares rate per 48 minutes so different playing times compare more fairly.
- PIEPlayer impact estimateNBA share of positive game events (scoring, boards, assists, steals, blocks).
- O-DPMDARKO offenseOffensive half of DARKO impact.Learn
- D-DPMDARKO defenseDefensive half of DARKO impact.Learn
Proprietary stats
Original DRBL numbers built from play-by-play. They show who is playing at a high level and who delivered the most season value.
- DRBL-LDRBL leverageLeverage-weighted impact for clutch moments. A DRBL diagnostic for high-stakes possessions.Learn
- DRBL ΔDRBL component disagreementShows when DRBL diagnostics disagree with each other. Useful for spotting mixed signals.Learn
- DRBL ±DRBL uncertaintyLegacy uncertainty diagnostic. Validated DRBL/100 uses a published shrinkage step instead.Learn
- DRBL/100How good was the player's impact rate? DRBL's main ranking number for quality per 100 possessions.Learn
- R1 PtsR1 PointsThe point-credit ledger behind WAR1. Same player order as WAR1, different units. Usually hidden on main boards.Learn
- WAR1How much season value did they pile up? DRBL's wins-style total above the role-matched R1 baseline.Learn
- DRBL-OOffensive half of DRBL's play-by-play impact split. Strong context for scoring and creation.Learn
- DRBL-DDefensive half of DRBL's play-by-play impact split. Strong context for suppressing scoring.Learn
- DRBL-PPossession-level diagnostic (parent of DRBL-O / DRBL-D). Do not add with LN and B to “rebuild” DRBL/100.Learn
- DRBL-LNLineup-context diagnostic: who you played with. Not proven off-ball value, and not part of a P+LN+B sum.Learn
- DRBL-BBox-score / behavior diagnostic (usage, creation, shot mix). Not camera tracking; not part of a P+LN+B sum.Learn
- R1The baseline DRBL compares players to: a role-aware expectation, not a classic “replacement player.”Learn
DRBL systems
How DRBL ranks seasons, resumes, and games.
- How DRBL worksPossession → expected scoring → player credit → rate + season total.Learn
- ValidationDRBL validationHeld-out and out-of-time tests for DRBL, not a claim that it beats other public metrics.Learn
- Historical dataDRBL historical dataWhy older seasons can have box scores without a published DRBL number.Learn
- LimitationsDRBL limitationsWhat DRBL is not: causal WAR, full off-ball measurement, or a proven best-in-public model.Learn
- CPICareer Production IndexDRBL’s transparent counting composite for Career Resume peak/prime/longevity, not DARKO or RAPTOR.Learn
- Career ResumeDRBL summary of a career’s Peak, Prime, Longevity, and Trajectory from qualifying seasons.Learn
- PeakCareer PeakThe single qualifying season with the highest Career Production Index (CPI) for this player.Learn
- PrimeCareer PrimeLongest contiguous run of qualifying seasons at or above 90% of the player’s own peak CPI.Learn
- LongevityCareer LongevityQualifying seasons at or above 70% of the player’s own Career Resume peak CPI, not merely years in the league.Learn
- Longevity-onlyLongevity-only seasonA qualifying season at 70–89% of peak CPI: longevity-level production outside the prime band.Learn
- Contiguous primePrime contiguityCareer Resume prime is the longest unbroken run of ≥90% seasons. A gap below 90% splits prime windows.Learn
- DevelopmentDevelopment / EmergenceDescriptive rise toward peak on the career arc, not a formal Career Resume v1 scoring band.Learn
- Career ArcHow Peak, Prime, Longevity, and trajectory phases fit together as overlapping performance bands, not exclusive buckets.Learn
- Career-selfCareer-self comparisonPeak/Prime/Longevity use % of this player’s own peak CPI, not a league or board percentile.Learn
- Rank My SeasonsRanks a player’s seasons using only pairwise comparisons and Copeland points.Learn
- Rank Team SeasonsRanks a franchise’s seasons with pairwise Team Season Compare and Copeland points.Learn
- Team CompareTeam Season CompareCompares team seasons head-to-head on board metrics, tolerances, and category plurality, without a composite score.Learn
- Season ComparePlayer Season CompareHead-to-head player seasons by category edges, the building block for Rank My Seasons.Learn
- CopelandCopeland rankingRanking method: win a pairwise matchup = 1 point, essentially even = 0.5, loss/unavailable = 0.Learn
- Game LabBox-score story of a game: what stood out, what changed, and team context, without PBP claims.Learn
- Season evidenceRepresentative games (largest win, highest scoring, etc.) that illustrate a season. Descriptive only, not “most important.”Learn
- Scoreboard onlyScoreboard-onlyDRBL knows the game from schedule/score data, but detailed box-score rows are not available yet.Learn
- Season baselineThe team's average performance across qualifying games from the same season, used to describe how unusual a single game was.Learn
- ASK DRBLNatural-language questions routed to existing DRBL analyzers, not free-form AI.Learn
- PercentilePercentilesShare of a defined population below this value. Always check which population DRBL used.Learn
- Historical impactDRBL only treats season-true impact archives as historical. Live overlays are not backfilled history.Learn
Labels & status
What DRBL means by even, unavailable, contested, and more.
- EvenEssentially evenThe two values are within DRBL’s comparison tolerance, so neither side gets a meaningful edge.Learn
- UnavailableRequired data is missing or not reliable enough for this comparison.Learn
- InsufficientInsufficient evidenceSome data exists, but the sample does not meet DRBL’s qualification threshold.Learn
- Not eligibleThis season does not meet the minimum participation rules for the analysis.Learn
- IncompleteIncomplete seasonThe season is still in progress or lacks enough games to treat as a completed season.Learn
- ContestedContested rankingPairwise wins form a cycle, so the ranking has no clean transitive ordering.Learn
- CloseClose topThe top-ranked options are separated by only a small Copeland gap under the current methodology.Learn
Transactions
Source events vs structured trades, and why genealogy waits.
- Source eventA transaction note kept word for word. Reporting context, not a structured asset ledger.Learn
- Related clusterRelated event clusterMultiple source events safely linked as related reporting. Still not a verified trade package.Learn
- Structured txStructured transactionVerified asset-level transaction. Currently 0 in production, so trade genealogy stays blocked.Learn
- TPETrade ExceptionA CBA mechanism that may let a team absorb qualifying salary without matching outgoing salary, only when a structured TPE ledger exists.Learn
- Salary fitWhether a player's salary amount fits a cap mechanism (e.g. a TPE). Fit is not eligibility and not legality.Learn
- LegalityTrade legalityWhether a full proposed transaction satisfies applicable CBA and roster rules. Requires a deterministic validator, never inferred from fit lists.Learn
- Draft capitalOwned draft picks and related rights from a structured pick ledger, not inferred from transaction notes.Learn
Sentiment
How the Sentiment tab measures fan and media tone, and where each number comes from.
- FanFan laneTone of fan discussion about a player or team, shown as 0–100% where 50% is neutral. It is measured from fan blog headlines, social posts and video comments when a subject has enough of them, and uses curated values otherwise.
- MediaMedia laneTone of news coverage about a player or team, 0–100% with 50% neutral. It is measured from publisher headlines when a subject has enough of them. It is never merged with the fan lane.
- Headline toneA language model's rating of one headline as positive, neutral or negative toward each player it names. Players named only in the summary get a word list score tuned for basketball ("waived" reads negative, "extension" positive). Against hand labels the model matched about 7 times in 10 and the word list about half, so single ratings can be wrong.
- CuratedCurated sentiment valueA hand-written prototype number that shows how the board works. It is not a measurement, and it is dated so you can see how old it is.
- Coverage floorThe minimum amount of source material before a lane gets a score: 3 headlines or 5 posts in the last 7 days. Below it the lane is blank, not neutral.