A few weeks back I wrote a post on whether momentum exists in college football, which ultimately concluded that what we mistake as "momentum" is more likely a naturally occurring mathematical phenomenon.
I aimed to test this idea on college basketball as well by looking at overtime games. I'm gauging the existence of "momentum" if the team that tied the game to send it to overtime then goes on to win that game. This isn't as easy an argument to make as scoring streaks in college football, since there are a whole host of reasons why either the leading or trailing team would win the game in overtime (team that ties wins: momentum, better team prevailing over a longer 45 minutes, etc; team that gives up lead wins: better team had the lead originally, trailing team ran out of gas, etc).
There are notable examples on both sides. During this ongoing season, Nevada came back from 14 down to New Mexico in 64 seconds to force and then win the game in OT. On the flip side, one of the more famous moments of the Duke-Carolina rivalry was Jeff Capel's buzzer beating half-court shot to force double-OT in 1995. Of course, the reason I include this is because it didn't matter: Carolina won the game anyway, 102-100.
I looked at every overtime game for both season-to-date in 2016-17 and the full season of 2015-16 (data courtesy of Basketball-Reference), resulting in 578 games. I then took the play-by-play and/or game recaps of these games and checked which team tied the game to force OT against which team ultimately won the game. After all that, the team that tied the game has a record of 287-291, or 49.65%.
With a record this close to .500, it's not even worth performing a test against the null hypothesis that "the team to tie the game is more likely to go on to win in OT", but I'll do it anyway. It returns a z-score of -0.17, with an associated p-value of 0.434, which is nowhere near any threshold of significance. As before with college football, it seems much more likely that we're fitting a narrative of "momentum" to something that instead occurs naturally by chance.
I aimed to test this idea on college basketball as well by looking at overtime games. I'm gauging the existence of "momentum" if the team that tied the game to send it to overtime then goes on to win that game. This isn't as easy an argument to make as scoring streaks in college football, since there are a whole host of reasons why either the leading or trailing team would win the game in overtime (team that ties wins: momentum, better team prevailing over a longer 45 minutes, etc; team that gives up lead wins: better team had the lead originally, trailing team ran out of gas, etc).
There are notable examples on both sides. During this ongoing season, Nevada came back from 14 down to New Mexico in 64 seconds to force and then win the game in OT. On the flip side, one of the more famous moments of the Duke-Carolina rivalry was Jeff Capel's buzzer beating half-court shot to force double-OT in 1995. Of course, the reason I include this is because it didn't matter: Carolina won the game anyway, 102-100.
I looked at every overtime game for both season-to-date in 2016-17 and the full season of 2015-16 (data courtesy of Basketball-Reference), resulting in 578 games. I then took the play-by-play and/or game recaps of these games and checked which team tied the game to force OT against which team ultimately won the game. After all that, the team that tied the game has a record of 287-291, or 49.65%.
With a record this close to .500, it's not even worth performing a test against the null hypothesis that "the team to tie the game is more likely to go on to win in OT", but I'll do it anyway. It returns a z-score of -0.17, with an associated p-value of 0.434, which is nowhere near any threshold of significance. As before with college football, it seems much more likely that we're fitting a narrative of "momentum" to something that instead occurs naturally by chance.
No comments:
Post a Comment