Pro Statistics and
League Statistics

Chris Foy

Craig O'Shannessey has revolutionized the way knowledgeable coaches think about rally length in pro tennis with statistics showing that 1 and 2 hit points are the norm and that long rallies of 10 plus balls are actually rare. He's already written extensively about it on Tennisplayer. (Click Here.)

But do those numbers apply to the rest of us? To find out I did a statistical study of women's league play in Atlanta, Georgia. The answer was fundamentally yes, but with a critical difference when it came to serve. To start, let me explain how I first discovered Craig's data analysis and how that inspired my own study.

Disbelief

At a coaches' symposium, I found myself sitting in Craig presentation in disbelief. He asked the conference room full of coaches a straightforward question. "What was the most common rally length at the 2015 Australian Open?"

Hands began to shoot up all around the room as if a third grade teacher had just asked a basic math question. "4" was shouted from a few rows behind me. "6" was blurted from the front of the room. I heard another coach yell out, "3."

Then Craig gave the real answer, "1." The room got silent fast. More than 200 tennis coaches in the audience did not believe what Craig was saying.

"Far and away, 1 is the most common rally length," Craig repeated to a very quiet group.

Then he explained is methodology. A double fault counted as a zero shot rally, an ace or any unreturned serve count as a 1 shot rally. A returned serve in which the server missed the second ball is a 2 shot rally, and so on.

Next he flashed the raw data up on the screen. 30% of all points at the Australian Open were unreturned serves. The other numbers were equally hard to believe. 15% of all points were three shot rallies. Less than 1% of points were more than 12 shots--or six shots per player. In total 70% of all points for were shots or less.

Someone shouted, "Wait, what about women's tennis?" A coach in the third row asked, "Yeah, but this is on hard courts; what about clay?"

Craig then proceeded to pull over 100,000 data points from all four majors, from both the men's and the women's games and the story was identical.

The objections kept coming. "Look Craig, this cannot be true with the girls 12s," insisted one coach. "What about women recreational players?" asked another.

Then Craig made a bold claim, "Ladies and Gentlemen, we all play the same game. This data will be equal across all surfaces and levels of the game."

I was in shock. Level of play, gender, or court surface had no statistically significant effect on the way a tennis point played out?

I left dumbfounded and intrigued. But I noticed that all of Craig's data dealt with performance players either pro, college or competitive juniors.

As a pro in Atlanta, I coach a lot of women's Atlanta Lawn Tennis Association league teams. I wondered if we might get a different statistical picture if we charted ALTA matches for rally length.

So that spring I got a clipboard, a piece of paper and a pencil. I labeled my paper into columns; 0 through 12+. I sat down on the balcony that overlooks our clay courts at Druid Hills Golf Club and started counting shots.

I charted two statistics. First rally length. Second, service holds and service breaks.

I was already interested in the question of holding serve based on conversations with my fellow staff pros. We had come to the conclusion that in recreational women's tennis it wasn't about holding serve, but rather holding return.

My Data

The conclusion from my data showed the rally length the numbers matched Craig's fairly closely. 70% of all points at the Australian Open were 0-4 shots. In the B- group in our league, with the largest number of data points I charted, the number was 67% of all points in that same 0-4 shot range.

In Australia 20% of all points were 5-8 shots. In Atlanta the number was 26%. In Australia 10% of all points were 9 shots or more. In Atlanta, the number was actually lower, with only 7% of all points going 9 shots or more.

But before we conclude that all tennis is exactly the same across all levels, let's take a closer look. Remember, at the pro level, 30% of all points are unreturned serves. It wasn't surprising to find at the women's league level this number was 13%, less than half.

This data also showed other differences. The 0-4 shot category does in fact account for about two-thirds of the game at both the pro and recreational levels.

But the further breakdown of the league points is a different story. The serve is the most important shot in the pro game with somewhere in the neighborhood 70% of all points going to the server.

However, my data shows an entirely different story in Atlanta. When looking at more than 2,500 points across four levels that I charted, I calculated the serving team had a statistical advantage of exactly 0%. That's right. Zero.

The data showed that it makes no difference at all whether you serve or return. You have about an equal chance of winning any given game.

The Same Game?

So my conclusion is that Craig is not wrong to say, "We all play the same game." Across all levels of the game the points are short. But the serve makes the anatomy of a pro point very different from a recreational one. For league play it just as important to "hold return" as to "hold serve."

So as a coach how does this change the way we work with players? My conclusion is that improving your level has less to do with groundstrokes or even volleys and more to do with improving your serve and return. One more game holding serve or one more game holding return can be the difference in a set or a match.

Most league practices focus on long groundstroke exchanges, but as we saw long rallies are only around 10% of actual league competitive points.

My data shows that the real skills to work on are the first four to five shots because that's about 80% of the game. I now emphasize the serve, the return and the first volley.

The goal should be trying to produce a miss from your opponent on the first or second ball you get a racket on because that may be the only opportunity you have before the point is over.

I don't mean that you should try to blast a winner every time you touch a ball, but instead try to keep your opponents from getting easy balls like a waist high forehand.

To work on this I created a drill game called "Finish in 4". Players play out a point while keeping track of the number of shots. If the rally goes to the fifth ball, the point stops and must be replayed.

The first week I started using this drill, some people thought this would result in games that never ended. But we found that only a fraction of points were replayed.

The women's teams started to love the sense of urgency it created to take control of a point early. One men's team I charted had two matches going and in 15 minutes, just 7 points went to the fifth shot.

It all suggests - like Craig O'Shannessy says - that the game of tennis is not only about forehands and backhands, it's more about serving and returning, and working on them is what practice should look like.