Latrinsorm
05-07-2014, 08:30 PM
The Heat crushing the Nets yesterday got me to thinking. Obviously a team that wins game 1 is in better position to win the series because it can go 3-3 and do so, and there's certainly a relationship between being a better team and winning. Over the past 11 postseasons, NBA teams have gone 121 of 165 (73.3%) after winning game 1. (The all-time mark of 77.3% is probably higher because of the shorter series that used to be played.) But does the strength of the win matter? Can we see any margin-based effects? As it turns out, yes!
http://img.photobucket.com/albums/v456/johnnyoldschool/NBAPlayoffFirstLoss_zps5ae4bce7.png
The x-axis should be read as "teams that won Game 1 by more than X points", so each additional point in the margin increases your chances of winning the series by 1% over the base 73%. I arbitrarily stopped at 12 because that happened to be the point in the sample beyond which less than 50 series qualified, I would assume that the curve levels off eventually.
Another thing I can do once the sample gets bigger is split out into home and away. Away teams definitely start with a much lower chance of winning the series, but I would guess that they get more % per point.
http://img.photobucket.com/albums/v456/johnnyoldschool/NBAPlayoffFirstLoss_zps5ae4bce7.png
The x-axis should be read as "teams that won Game 1 by more than X points", so each additional point in the margin increases your chances of winning the series by 1% over the base 73%. I arbitrarily stopped at 12 because that happened to be the point in the sample beyond which less than 50 series qualified, I would assume that the curve levels off eventually.
Another thing I can do once the sample gets bigger is split out into home and away. Away teams definitely start with a much lower chance of winning the series, but I would guess that they get more % per point.