Does Cincinnati Matter in Flushing?

After months of clay and grass tournaments, the best players on tour are finally competing on hard courts.  For many, Cincinnati is the extent of their North American hard court preparation leading up to the US Open.  No matter who wins this week, we’ll be tempted to anoint him the favorite in New York.  Should we?

Traditionally, Cincinnati features one of the strongest draws of the ATP season.  As the only tournament scheduled two weeks before the US Open, there are no alternatives for players preparing for the slam, and it still allows a week off.  This year’s draw, missing three top 10 players due to injury, is an aberration.

It’s no surprise, then, that the list of winners in Cincinnati is particularly impressive.  19 of the last 20 champions have career peak rankings of 1 or 2.  (The black sheep in the group is Thomas Enqvist, who “only” reached #4.)  Not only do the best in the world show up to play, they show up to win.

More than some warmups, Cincinnati seems to tell us who is in form.  Let’s see if tells us who is going to win the Open.

Since 1991, there have been four seasons when the same man lifted the trophy in Cincinnati and New York: Pat Rafter in 1998, Andy Roddick in 2003, and Roger Federer in 2005 and 2007.  Five more times, the Cincinnati winner reached the US Open final.  Not counting 1999, when Pete Sampras didn’t compete in Flushing, the Cincinnati champion has failed to reach the US Open round of 16 only twice in the last 21 years.

So, the Cincinnati winner has won the US Open about 20% of the time, and reached the final another 25%.  Sounds good, though not as good as we’d expect from the top seed.  On the other hand, Cincinnati winners aren’t always the top seed in New York, so we can’t expect them to perform at that level.

In fact, the Cincinnati winner has been the top seed in Flushing only six times.  On average, the Cinci champion has been seeded 4th in New York.  Compared to the performance we’d expect from a #4 seed, a 20% shot at winning the tournament, along with a nearly 1-in-2 chance of reaching the final, is extremely good.

Since 1991, #4 seeds at the US Open haven’t performed nearly so well during the final weekend as have Cincinnati champions.  Both groups have a roughly 6-in-10 chance of reaching the semis (#4 seeds: 57.1%, Cinci winners: 60%), but the #4 seeds have won only half of their semifinals, for a 28.5% chance of reaching the final, compared to the 45% of Cincinnati titlists.

The biggest difference is where it matters most: the final itself.  Cincinnati winners go on to win almost half of their US Open finals, winning 4 titles in 20 attempts, as we’ve seen.  But #4 seeds have won only 2 titles.  It’s not a huge sample, but if we expand our view to consider all four slams since 1991, the performance of #4 seeds stays about the same.

Much to my surprise, it seems that Cincinnati results do have something to say about the final rounds in Flushing.  This week’s winner isn’t exactly a lock to triumph in New York, but his performance in Ohio will tell us to expect that much more from him at the US Open.

3 thoughts on “Does Cincinnati Matter in Flushing?”

  1. An interesting article, insightful as always. Interesting also that 2/4 Cincinnati-US Open winners also won in Canada (Rafter, Roddick) and the only other Canada-Cincinnati winner in fairly recent times (Agassi) made the final in US Open.
    It was in reading an article about the recent Cincinnati final, that I came across the following statement:
    “Djokovic has always seemed particularly vulnerable towards the close of tight sets, and as far as I can tell he loses 7/5 more often than he should. I confess I have no sound statistical evidence to back this up, but as recently as the Olympics, he lost three such sets in his last two matches. Three years ago in the Cincinnati final he fell to Federer 6/1 7/5.”
    Full Article:
    I was wondering if you had any thoughts on this. Some players do seem to lose a surprising number of sets 7-5…or is this just an instance of selective memory.

    1. Carl Bialik did something on this a year ago:

      He uses it as something of a proxy for clutch, because 7-5 is the only way we can tell from the scoreline if a player lost a key service game (or won a key return game).

      Seems like a good thing to study further, especially with data on total points won, which would tell us how often a specific player should win or lose 7-5 sets. But not something I’ve yet done anything with.

Comments are closed.