The home team opens at −4.5 with decimal odds of 1.90. Later it is −6.5 at 1.90. A price-only chart would be flat, yet the selection has changed: the later line requires a larger winning margin. Basketball research needs the line and the price in the same record.
A useful observation contains the team, market period, handicap or total, price and timestamp. Add the overtime definition before joining the observation to a final score. Moneyline, spread and total markets describe different outcomes even when they share a fixture identifier.
Unchanged odds can conceal a two-point move
Consider an invented final score of 89–84 under the relevant market rules. The home team wins by five. It covers −4.5 by half a point but fails to cover −6.5 by one and a half. The two quotes both read 1.90, but they cannot be treated as observations of an unchanged selection.
Record “spread moved two points against the home selection; price unchanged” rather than “no movement”. If the archive contains alternative lines, a fixed-line comparison may be possible. If it does not, leave that comparison unavailable instead of constructing a price that was never quoted.
Choose the score that the market actually settles on
| Score basis | Score | Combined points | Against 170.5 |
|---|---|---|---|
| End of regulation | 82–82 | 164 | Under |
| Including overtime | 94–91 | 185 | Over |
bet365's US basketball rules include overtime in pre-game bets unless otherwise stated. Treat that as a documented example, not a rule for every bookmaker, jurisdiction or quarter market. The archive's settlement definition decides which score belongs in your calculation.
Keep pushes visible
At an integer total of 170, a combined score of exactly 170 is a push when the market returns the stake on equality. A research table should distinguish wins, losses, pushes and missing results. Counting a push as a loss changes the outcome label; removing it without explanation changes the denominator.
In an illustrative 100-record sample with 48 overs, 47 unders and five pushes, overs represent 48% of all records or 50.53% of the 95 decisive results. Neither percentage is a return calculation. The odds and treatment of every outcome would still be needed for that separate question.
Compare one competition and format first
A total-points band is easier to interpret within a consistent match format. Separate competitions with different scheduled lengths, and preserve whether each quote was taken before the game or during it. If you investigate a lineup announcement, retain its publication time rather than using a later explanation as if it were known at the opening.
The existing odds comparison workflow provides the basic matching steps. For basketball, make line and overtime coverage mandatory fields. That small addition prevents many apparent price patterns from being differences in market definition.
Sources and method
Original research examples prepared with AI assistance. Source rules, arithmetic and links were checked. Scores and prices are illustrative, not observed fixtures or performance claims. Bookmaker terms vary by jurisdiction, market and date.
- bet365 US: basketball settlement rules — accessed 23 October 2026