North Carolina beat Clemson 59 consecutive times in Chapel Hill, the longest such streak in NCAAB, dating back to 1926. Admittedly, the annual component became less impressive once Miami and Va Tech joined the conference in 2004 and we stopped playing Clemson at home every year. But beating ANYONE 59 times in a row is impressive, regardless of the gulf between two programs. Even throughout fluctuations in UNC's strength (we beat them by 18 at home in the worst year (2001-02) in the program's history), the streak persisted until this year.
So just how unlikely was it? Short of calculating every team's RPI (a notoriously unreliable metric) back to 1926, I tried to approximate it through a few different methodologies.
If the talent disparity each year was equal to 2005's title team (21 points better than Clemson), each win would have a 97% chance: 0.97 ^ 59 = 16.6%, or about 1 in 6. Not that crazy.
Taking the other extreme, if every year was equal to the 2002 team, where Clemson actually was slightly BETTER, factoring in home court would yield a 63.94% win probability: 0.6394 ^ 59 = 0.0000000003%, or 1 in 288,014,740,777. That's 288 billion. This is pretty close to each team being evenly matched and adding ~5 points for home court (65% win probability in each game). Which isn't a particularly satisfying answer either: this isn't representative of reality.
Over the full head-to-head records (home/away/neutral), Carolina led 128-20 (86.5%) going in to this season's game. Using this win probability gives: 0.865 ^ 59 = 0.019%, or 1 in 5,249.
Even so, this streak had a 0.9% chance of ending (1 in 111) late on January 11 - until that unlikely event happened too.
No comments:
Post a Comment