Shots made out of shots taken: 5/8 means 5 made from 8 tries. 2P%, 3P% and FT% are the share made.
Reb
Rebounds
Missed shots grabbed, at either end of the court.
Ast
Assists
Passes that lead straight to a basket.
Stl
Steals
Taking the ball off the other team.
Blk
Blocks
Opponents' shots knocked away.
To
Turnovers
Losing the ball without getting a shot off: a bad pass, a travel, an offensive foul and so on.
+/−
Plus-minus
The score difference while the player was on court: +6 means the team outscored its opponent by 6 points in those minutes.
PIR
Performance Index Rating
EuroLeague's official all-in-one score: the good things a player does minus the bad ones. EuroLeague Fantasy scores on it too.Points, rebounds, assists, steals, blocks and fouls drawn, minus missed shots, missed free throws, turnovers, shots that got blocked and fouls committed.
GP
Games played
Games the player has appeared in.
vs · @
In game logs, vs is a home game and @ an away game.
OT
Overtime
Five extra minutes, played when the score is level after 40.
Points per shot, counting two-pointers, three-pointers and free throws: the fairest single measure of how efficiently someone scores.Points ÷ (2 × (shots + 0.44 × free-throw attempts)). The 0.44 is there because free throws mostly come in pairs, so two of them use up about one chance to score.
eFG%
Effective field goal %
Shooting percentage with a made three counted as one and a half baskets, since it's worth 50% more.(Made shots + half the made threes) ÷ shots taken. Free throws aren't included; true shooting adds them.
FGA · 2PA · 3PA · FTA
Attempts
Shots taken: all field goals (twos and threes), two-pointers, three-pointers and free throws.
Qualified players
Shooting leaders need a minimum number of attempts, shown on each list. It grows as the season goes on, so a few hot nights can't top the table.
How strong a team is right now, as an Elo rating: 1500 is average, and home court is worth about 100 points.It's updated after every game since 2016-17, and beating a strong team or winning big moves it more. Every summer ratings move 30% back toward average; new clubs start just below it.
vs avg
The chance of beating an average EuroLeague team on a neutral court.
Last game · Season
How much the rating moved in the latest game, and since the season began.
Trend
The rating over the last ten games: green if it went up, red if it went down.
Form
Results of the last five games.
Top 6 · Play-in
Teams finishing 1st–6th go straight to the playoffs. 7th–10th play the play-in for the last two playoff places.
Season odds
Each team's chance of finishing 1st, making the top 6, the play-in, the playoffs and the Final Four, and of winning the title.The rest of the regular season, the play-in, the best-of-five playoffs and the Final Four are played out thousands of times with the current power ratings; the odds are how often each thing happened. Updated every night.
Projected
The average final win–loss record across those simulated seasons.
A possession is one team's turn with the ball (Poss counts them). Measuring per 100 possessions makes fast and slow teams, and long and short spells on court, comparable.Possessions are estimated from the play-by-play: shots, free throws, offensive rebounds and turnovers.
Off · Def · Net
Net rating
Points scored (Off) and allowed (Def) per 100 possessions. Net is the difference: +5 means outscoring opponents by 5 points every 100 possessions, and 0 is league average.
On/off swing
How much better or worse the team did per 100 possessions with the player on court than on the bench, in games the player played.It depends on teammates and opponents: starters often face the other team's best players, and a strong bench lifts the off-court numbers.
Adjusted impact
The points per 100 possessions a player adds compared with an average EuroLeague player, after accounting for every teammate and opponent on court with them. Split into offence and defence.It's fitted on every stint of the season (a spell with the same ten players on court), starting from the player's earlier seasons. The range (±) shows how sure the estimate is: the true value is 90% likely to be inside it.
Lineup
A group of players on court together. Best lineups ranks five-, three- and two-man groups by net rating.
Small sample
Not many minutes yet, so the number can still swing a lot. Read with care.
Our model's chance that each team wins. It combines power ratings, season efficiency (net rating) and home court, is updated every night and is locked at tip-off.
Expected margin
How many points the model expects the favourite to win by. Under half a point, the game is called a toss-up.
Ratings alone
What the power ratings by themselves would say, as a yardstick for the full model.
Backtest
How the model would have done in past seasons, each season predicted using only the seasons before it.
Track record
Real calls, saved before tip-off and graded after the game.
Calibration
Whether the percentages can be taken at face value: when the model says 70%, does the favourite win about 70% of the time?
Log loss
A score for how well-judged the chances were: confident right calls score well, confident misses cost a lot. Lower is better; always saying 50–50 scores 0.693.Chances are capped at 1% and 99% so one certain miss can't dominate.
Brier score
Another score for well-judged chances: the average squared gap between the chance given and what happened (100% or 0%). Lower is better; always saying 50–50 scores 0.250.
Jev AI
TypeSafe's decision model, as a second opinion. Before each game it gets the same kind of facts our model uses, but never our prediction, and gives the home team's chance of winning.
Polymarket
A prediction market where a share in the winning team pays $1, so a price of 77¢ means traders give that team about a 77% chance.Prices are saved every 15 minutes until tip-off. Games are graded at the last one, midway between the bid and the ask.
Bid · Ask · Spread
The bid is the best price a buyer offers right now, the ask the cheapest a seller accepts, and the spread is the gap. A spread over 10¢ means the market is too thin to trade.
Paper trading
Pretend trades on Polymarket's prices with no real money, to test whether the models could make a profit. Not betting advice.
Value rule
Buys the side the model rates higher than the market does: home if the model gives the home team a better chance than the price, away if worse.
Pick rule
Simply backs the team the model thinks will win.
Return
Profit or loss as a share of everything spent on trades, fees included.
Fees
Polymarket charges buyers 5% × price × (1 − price) a share: most on even games, least on heavy favourites.