An opening total is the bookmaker’s first estimate of the combined points two teams will score in a game. Also called the opening over/under, it sets a single threshold: over bettors expect more combined points, while under bettors expect fewer. The winner and margin do not matter for this wager. section 17.0 section 17.3
The number is designed to attract balanced money on both sides. If wagering is uneven, the line can move before betting closes: upward to encourage under bets or downward to encourage over bets. The resulting number is the closing total. section 17.0
Grizzlies at Pacers
On October 17, 2018, the opening total was 209. A bettor choosing the over expected at least 210 combined points; a bettor choosing the under expected 208 or fewer. The line later closed at 204.5, while the teams combined for 194 points. section 17.0 section 17.4.2
To evaluate an opening total after a game, add road and home points, then calculate its absolute error: the distance between combined points and the opening total. Calculate the same distance for the closing total. The smaller distance identifies which estimate was nearer to the actual result; ties are recorded separately. section 17.4.2 section 17.5.1
oddsdf3 %>%
mutate(versusPTS =
case_when(diff_ptsT_opentotal >
diff_ptsT_closetotal ~ "closetotal",
diff_ptsT_closetotal >
diff_ptsT_opentotal ~ "opentotal",
diff_ptsT_opentotal ==
diff_ptsT_closetotal ~ "same")) -> oddsdf3
oddsdf3$versusPTS <- factor(oddsdf3$versusPTS)
This transformation labels each game according to whether the opening total, closing total, or neither was closer to the combined points scored. section 17.4.2
Across the 2018–19 NBA regular season data, closing totals were nearer to combined points more often than opening totals, by roughly 10% in count comparisons and about 2% in average variance. That pattern held whether the total moved up or down, supporting the idea that wagering adds information to the initial estimate. section 17.5 section 17.5.1 section 17.6
