BIG DATA

Examples:

• The million point of college tennis
• 0-4 makes up 70% of tennis

These assertions are not incorrect, but they say nothing of the individual nature of tennis, and result in shallow analysis.

Profiling players based on their 135 (serve), 246 (return) and 7+ (long rally) data makes more sense because it individualises the data.

BAD DATA

Example:

• Points won on first serve.

Points won on first serve is supposed to measure the effectiveness of a player’s first serve. But if a player’s first serve is in and the rally then goes 35 shots and the server wins the point, we could hardly call that a point won because the first serve was effective.

Equally, if the returner makes an error on shot 4 after a great return, the server wins the point but their first serve had nothing to do with the outcome.

The best way to measure the effectiveness of a player’s first serve is to measure how often they win the point on shot 1, 3 or 5, as that sequence of shots is a far more accurate measure of first serve effectiveness.

POINT ENDING DATA

We have always reported point ending data in tennis.

Example:

• Felix hit a forehand winner.

We know 3 things: the player, the type of shot and whether it was a winner or an error. The problem is we have no context.

Instead, using the 135 Framework, we want to know these 3 things about a rally before branching into other details:

1. Who was serving?
2. How long was the rally?
3. Who won the point?

So, Felix was the server. The rally was 3 shots. Felix hit a winner. 

Now we have the story of the point, or group of points, which is far easier to explain to junior players.