Every game sits in one of three states, and the pill beside the tip time says which. Confirmed means both clubs have announced their five. We check every minute from two hours before tip, so it lands within a minute of the announcement. Expected is the five that's been reported before the clubs say it. From Oct 30 to Nov 9 it matched the announced five on 149 of 172 club-games. Projected is our own call, for a game nobody has reported on yet.
A projected five starts from the club's last announced five. Anyone ruled out on the injury report, or no longer on the roster, is replaced by the player with the next-most expected minutes. We tested that against the five the club actually named on 156 club-games: it got 4.5 of the 5 starters right on average and all five on 91. Taking the five players with the most projected minutes got 3.9 right. A coach's habit beats a minutes model at this question.
Min is expected minutes. A named starter keeps his full number, because the five is the news that he's playing. Everyone else's is scaled by how many of the club's last 10 games he played, so a club's column adds up close to the 240 minutes a game has. Without that, the raw numbers ran to 290 a club, because every player carries his minutes if he plays. Scaling took the average miss from 7.0 minutes a player to 5.7. Pts, Reb, Ast and 3PM are our projections on the same basis.
The bench lists everyone we expect to play at least five minutes. At that cutoff the list runs about 11 players a club, close to the 11 who actually get in. Q, D and P beside a name are questionable, doubtful and probable on the NBA's official injury report. A late change is a club whose announced five differs from the one that was expected, with the books' current line for each player who moved. Books move those lines fast, so treat the number as where it stood when the page last refreshed.
Bucks
Knicks
