The Mathematics of Juggling and Site-Swap Notation

This article explores how mathematical pioneers Claude Shannon and Ronald Graham transformed the art of juggling into a quantifiable science. By establishing foundational equations and expanding the discrete mathematics of the site-swap notation system, these mathematicians helped create a universal language for jugglers. This system not only classified every existing juggling trick using simple sequences of numbers, but also used combinatorics to predict and invent entirely new patterns that had never been physically attempted.

Claude Shannon’s Foundational Juggling Theorem

Before juggling patterns could be written as simple numbers, the physical constraints of juggling had to be defined mathematically. In the 1970s, Claude Shannon—best known as the father of information theory—developed the first formal mathematical theorem of juggling.

Shannon observed that juggling is strictly bound by time and space. He formulated an equation that relates the number of objects being juggled to the time those objects spend in the air and in the juggler's hands:

(F + D) / (V + D) = B / H

  • F (Flight time): The time a ball spends in the air.
  • D (Dwell time): The time a ball spends in a hand.
  • V (Vacant time): The time a hand spends empty.
  • B (Balls): The total number of props.
  • H (Hands): The number of hands used.

Shannon’s theorem proved that to juggle more objects, a juggler must either throw the objects higher (increasing flight time) or move their hands faster (decreasing dwell and vacant time). This time-based relationship laid the groundwork for the development of site-swap notation.

The Birth and Expansion of Site-Swap

Building on Shannon's concept of tracking throws over discrete beats of time, the site-swap notation system was independently developed in 1985 by Paul Klimek, Bruce Tiemann, and Mike Day. Shortly after its inception, mathematician and avid juggler Ronald Graham—who served as president of the International Jugglers' Association—recognized the profound mathematical depth of the system.

Graham, alongside mathematicians like Joe Buhler, David Eisenbud, and Colin Wright, formalized the combinatorial mathematics behind site-swap. While the original inventors created the notation, Graham and his colleagues used graph theory, abstract algebra, and combinatorics to prove how and why the system worked, publishing seminal papers that brought site-swap to the global scientific and juggling communities.

How the Site-Swap System Works

Site-swap strips away the complex variables of hand movements and body positions, focusing entirely on the rhythm of throws and the order in which props are caught.

The system assumes that throws happen on a steady beat, alternating between the left and right hands. Each throw is assigned a single number, which dictates how many beats into the future that specific prop will be thrown again.

  • 0: An empty hand on that beat (no throw).
  • 1: A quick, flat pass directly from one hand to the other.
  • 2: A hold, where the ball stays in the same hand for a beat.
  • 3: A standard throw that crosses to the other hand, at the height of a normal three-ball cascade.
  • 4: A higher throw that lands back in the same hand that threw it.
  • 5: A high throw that crosses to the opposite hand, used in a five-ball cascade.

Odd numbers always cross to the opposite hand, while even numbers always return to the same hand. Because juggling patterns repeat, a trick is written as a repeating sequence of numbers. A standard three-ball cascade is simply written as 333 (or shortened to 3), because every throw is identical.

The Mathematics of Inventing New Patterns

Graham and his fellow researchers utilized two primary mathematical rules within site-swap to test if a random sequence of numbers could actually be juggled without two balls colliding in the same hand at the same time.

1. The Averaging Rule

To find out how many balls are required to juggle a specific site-swap pattern, you simply calculate the average of the numbers in the sequence.

  • Take the pattern 531.
  • Add the numbers together: 5 + 3 + 1 = 9.
  • Divide by the length of the sequence (3 beats): 9 / 3 = 3.
  • Therefore, 531 is a three-ball pattern. If the average is not a whole number, the pattern is mathematically impossible to juggle.

2. The Collision Test

Even if the average is a whole number, two balls might still land in the same hand on the same beat (e.g., the pattern 432 averages to 3, but is impossible). To test for collisions mathematically, you add the value of each throw to the beat number on which it was thrown, and then apply modular arithmetic (dividing by the length of the sequence and looking at the remainder). If all remainders are unique, the pattern is valid and juggleable.

By feeding these two rules into computers, mathematicians generated thousands of valid numerical sequences. Jugglers then looked at these computer-generated numbers and worked backward to figure out how to physically throw them. Classic modern tricks like 441 and 531 were never conceived naturally by performers; they were discovered on paper using site-swap math first, fundamentally changing the evolution of physical juggling.