|
Before beating Marist in overtime on Sunday, the Temple Owls were having a rough weekend in Puerto Rico. On Thursday, they were on the cusp of securing a quality win against Providence in the first round of the Puerto Rico Tip-Off Classic. The Owls enjoyed a comfortable lead for most of the game and led by 10 with six minutes to go before the Friars closed the game on a 14-2 run to win 66-64. The following night, Temple appeared to recover from that disappointing finish with a monster first half. After Chris Clark's half-court heave at the halftime buzzer, the Owls owned a 25-point lead against the College of Charleston.
It wasn't enough. The Cougars went on a 45-18 run after the break, taking a lead of their own before the final media timeout, and ultimately winning 79-76 on their own long-range buzzer beater. The game will probably hold up as the biggest halftime comeback of the season. It was the largest halftime comeback in the history of the C of C basketball program, and the largest one at the D-I level of college hoops in at least nine seasons.
Just how improbable was the reversal of fortunes in this game? As I alluded to over the summer on my blog, I've been collecting information to start answering questions like this. The project ultimately will be able to provide us with the probability of a team winning the game given the time left, the pace of the game and the game's location. It's a project that will likely take months, even a year or two, to complete. Using the halftime data I have for every game since the 2000 season, we can get an idea of just how unusual Charleston's win was.
The data collected for over 40,000 games from the last eight seasons is enlightening and illustrates just how rare this comeback was. The largest halftime deficit overcome during that time was 21 points. This was accomplished on seven occasions, most recently on January 13, 2007 when Mercer overcame a 56-35 deficit against Campbell to eventually win 94-93.
The Mercer/Campbell game provides the disclaimer for just how simplistic the analysis is at this point. The halftime score doesn't do the Bears' comeback justice--they were actually down by 23 after Campbell's Russ Gibson hit a jumper at the 15:59 mark of the second half. Perhaps a better example of how halftime data doesn't always tell the best story is Kentucky's legendary 1994 comeback at LSU. The Wildcats were down by 31 with 15:34 left in the second half, after being down by just 16 at intermission. In the context of the halftime score, Kentucky's comeback was not that amazing--second-half comebacks of 16 points or more have occurred an average of 11 times a season since the 2000 campaign. Obviously, knowing how the second half unfolded, it was an amazing comeback.
So with an understanding that this isn't a study of the most incredible turnarounds ever, only the most impressive relative to the halftime deficit, let's put all the data on the table. The following list includes games from the 1999-2000 season through Saturday.
Halftime
Lead Record Pct
26+ 1097-0 1.000
25 214-1 .995
24 271-0 1.000
23 299-0 1.000
22 403-0 1.000
21 431-7 .984
20 531-6 .989
19 639-6 .991
18 701-15 .979
17 866-19 .979
16 885-35 .962
15 1088-41 .964
14 1169-66 .947
13 1322-116 .919
12 1438-137 .913
11 1571-165 .905
10 1649-244 .871
9 1702-330 .838
8 1845-432 .810
7 1874-527 .781
6 1934-651 .748
5 2014-775 .722
4 1946-1018 .657
3 1800-1122 .616
2 1732-1240 .583
1 1618-1444 .528
This table at least gets us in the ballpark of establishing the likelihood of various halftime comebacks. Charleston's performance has happened only once out of 215 times a team has faced a 25-point halftime deficit over the past eight years. This comeback has happened 0.47% of the time a team faces such a deficit. another way to look at it is that a team facing a deficit of 22-25 points has only overcome that once in 1188 attempts, or 0.09% of the time.
However, that's not how the data should be interpreted. The majority of those large halftime leads occurred when one team was obviously better than its opponent. It's pretty rare for teams fairly close in strength, like Temple and Charleston, to produce such a lopsided first half. It also doesn't distinguish the site of the game. Teams with a lead at home are more likely to win than being in a similar situation on the road or on a neutral court. So when the College of Charleston sat in the locker room of the Coliseo de Puerto Rico on Friday, their chances of defeating Temple were probably closer to 1 in 100 than 1 in 1000. Nonetheless, it was a rare feat, one we may not see duplicated for years.
Those who follow the college game closely know that this is not the first time Fran Dunphy has been on the wrong side of a historic comeback. On February 9, 1999, his Penn team took a 33-9 advantage to the half against Princeton. The Tigers came back to win 50-49. It's pure speculation right now, but given the pace of that game and the fact that it was a home game for Penn, I wouldn't be surprised if that contest has a case for being one of the unlikeliest halftime comebacks ever.
Unfortunately, I haven't been able to find halftime data for the '99 season. Otherwise I could say with certainty that Dunphy has been on the wrong end of the two biggest second-half comebacks over the past ten seasons. Given that there wasn't a comeback of greater than 21 points since the 2000 season, it seems improbable that there was one of 24 points or more over the last month-and-a-half of the '99 season.
Think about it: There have been about 42,000 games involving a D-I team since that unusual night at the Palestra. Dunphy has coached in just 245 of them, and it's one of his teams that is the first to exceed what occurred on that night. If there's any consolation for Dunphy, it's that his team is nowhere close to being a part of the biggest halftime comeback ever. According to the NCAA Record Book, the honor belongs to Duke, which overcame a 29-point deficit against Tulane to win 74-72 in a 1950 "Dixie Classic" game.
Ken Pomeroy is an author of Basketball Prospectus.
You can contact Ken by clicking here or click here to see Ken's other articles.
|