First Round · 2007-08
7 games · Boston Celtics advanced
Every game
Game 1Apr 20
ATL 81@
BOS 10421·19·15·26 / 29·20·24·3191%Game 2Apr 23ATL 77@BOS 9620·22·16·19 / 24·28·24·2092%Game 3Apr 26BOS 93@ATL 10226·30·18·19 / 32·24·28·1828%Game 4Apr 28BOS 92@ATL 9724·24·27·17 / 29·22·14·3232%Game 5Apr 30ATL 85@BOS 11019·24·21·21 / 27·31·23·2987%Game 6May 2BOS 100@ATL 10332·18·32·18 / 20·29·30·2432%Game 7May 4ATL 65@BOS 9916·10·17·22 / 27·17·35·2087%
ATL 81@BOS 10421·19·15·26 / 29·20·24·3191%Game 2Apr 23ATL 77@BOS 9620·22·16·19 / 24·28·24·2092%Game 3Apr 26BOS 93@ATL 10226·30·18·19 / 32·24·28·1828%Game 4Apr 28BOS 92@ATL 9724·24·27·17 / 29·22·14·3232%Game 5Apr 30ATL 85@BOS 11019·24·21·21 / 27·31·23·2987%Game 6May 2BOS 100@ATL 10332·18·32·18 / 20·29·30·2432%Game 7May 4ATL 65@BOS 9916·10·17·22 / 27·17·35·2087%Each game opens its own page with the period breakdown, the team totals and the full player box score.
What the model made of it
Games it favoured correctly
4 of 7
Mean probability on the winner
64.1%
Brier over the series
0.2128
A backtest, and a sample of 7. The model was refit on games strictly earlier than each of these, so it never saw the result — but a record over a handful of games is not evidence in either direction. The one that means something is on the accuracy page, over 25,749.