Qualifying and quantifying 'clutch'
Meet Benjamin Tom.
He's a complicated fellow, who played on Cal-Berkeley's national championship club tennis team. He also graduated with a degree in Business Administration -- and in Molecular & Cellular Biology. He was wafting along at the Georgetown University School of Medicine when he suddenly wondered if he really wanted to be a doctor. He asked the school's administration for a leave of absence, which was granted. Now, he's pursuing his dream -- sports analytics.
His parents, naturally, are very proud.
Tom happened to be coaching ESPN analyst Pam Shriver's son, George, on a Southern California under-10 team, when the two got to talking.
"Pam and I were brainstorming," Tom explained, "about how to look at the numbers differently in tennis."
He was aware of how some analytics people quantified NBA clutch play, measuring a player's contributions in the last two minutes of a game and when the score differential was five or less. Could the two come up with a similar model for tennis?
Here is what Tom and Shriver settled on for defining clutch moments in tennis:
• All break points
• All set points
• All tiebreaker points
• All deuce or advantage points when a total of eight or more games have been played, including 30-all and 40-30.
• All points played in the deciding set, when the game differential is no greater than three -- the third set for women and the deciding set for men, either three or five.
So while you've been enjoying Serena's march through the Grand Slams, Tom has been busy breaking down that trek -- in excruciating detail.
And here's what he's discovered:
In winning all seven matches at Wimbledon, Serena outperformed the field in 9 of 10 statistically comparable categories, many by a substantial margin. While the other 127 players average 67 percent of first-serve points won, Serena won 80.3 percent. The field averages winners 15.3 percent of the time, while Serena was at a robust 24.4. Her edge in ace percentage was 18.3 to 5.
But that is reasonably predictable stuff.
Tom also calculated how Serena did measured against herself, comparing her total points to "clutch" points. Of the 962 points she played at the All England Club, 695 were defined as non-clutch points, with 267 deemed to be "clutch."
Serena's already extraordinary play elevated to a higher level in the crucible of clutch. She was better in 8 of 10 categories, including dramatic improvements in winners versus unforced errors, first-serve percentage and aces.
Looking at Serena's 26 matches Grand Slam matches, Tom isolated her third-set statistics versus the overall stats for all those matches. As you might expect, Serena plays markedly better in the third set.
Serena surpasses herself in 7 of 8 categories and is even (60 percent) in first-serve percentage. Of particular note, the ratio of winners to unforced errors (which goes from 1.4 to 1.67) and percentage of total points won (56 to 60) improve appreciably.
"It would be the equivalent of having an NBA or major league player, who already leads the league in nearly every statistic, get even better if we only measure their stats in overtime or extra innings," Tom said.
And then there is the ultimate comparison -- the third-set numbers versus the overall numbers of those 11 three-set matches.y.
Across the board, the margins increase and, in some cases, double or more.
The critical category of break-point conversions, for instance, improves from a 1 percent increase to 6 percent. Likewise, second-serve winning percentage, receiving points won and winner/unforced error ratios all soared when it mattered most.
"We have always felt that Serena has been the greatest of clutch players," Shriver concluded. "With Ben Tom's analytics, we can put more value and context to her amazing levels under at the biggest stages, especially in final sets."