The Mathematics of Juggling - WordPress.com · The Mathematics of Juggling Stephen Hardy March 25,...
Transcript of The Mathematics of Juggling - WordPress.com · The Mathematics of Juggling Stephen Hardy March 25,...
The Mathematics of Juggling
Stephen Hardy
March 25, 2016
Stephen Hardy: The Mathematics of Juggling 1
Introduction
Goals of the Talk
I Learn a little about juggling.
I Learn a little mathematics.
I Have some fun.
Stephen Hardy: The Mathematics of Juggling 2
Introduction
ReferenceMost of the material for this talk was gleaned from BurkardPolster’s great little book The Mathematics of Juggling:
Stephen Hardy: The Mathematics of Juggling 3
What is Juggling?
What is Juggling?
Juggling takes many forms...
Stephen Hardy: The Mathematics of Juggling 4
What is Juggling?
Examples
Jonathan Alexander http://adoreministries.com/, Kevin Rivoli http://www.syracuse.com/
Stephen Hardy: The Mathematics of Juggling 5
What is Juggling?
Diablo
Tony Frebourg - http://historicaljugglingprops.com/
Stephen Hardy: The Mathematics of Juggling 6
What is Juggling?
Flair Bartending
Ami Shroff - https://www.facebook.com/amibehramshroff
Stephen Hardy: The Mathematics of Juggling 7
What is Juggling?
Contact Juggling
Michael Moschen - http://www.michaelmoschen.com/
Stephen Hardy: The Mathematics of Juggling 8
What is Juggling?
Bounce Juggling
Dan Menendez - http://pianojuggler.com/
Stephen Hardy: The Mathematics of Juggling 9
What is Juggling?
Juggling is...
I Juggling is manipulating more objects than hands you areusing.
Bob Whitcomb, http://historicaljugglingprops.com/
I We are interested in toss juggling.
Stephen Hardy: The Mathematics of Juggling 10
What is Juggling?
Toss Juggling Rings
Anthony Gatto - http://www.anthonygatto.com/
Stephen Hardy: The Mathematics of Juggling 11
What is Juggling?
Toss Juggling Clubs
Jason Garfield - http://jasongarfield.com/
Stephen Hardy: The Mathematics of Juggling 12
What is Juggling?
Toss Juggling Knives
Edward Gosling - http://chivaree.co.uk/
Stephen Hardy: The Mathematics of Juggling 13
What is Juggling?
Toss Juggling (Passing)
Cirque Dreams Jungle Fantasy - https://en.wikipedia.org/wiki/Juggling_ring
Stephen Hardy: The Mathematics of Juggling 14
What is Juggling?
Toss JugglingWe will focus on a single person toss juggling with two hands.
Jason Garfield - http://jasongarfield.com/
Stephen Hardy: The Mathematics of Juggling 15
Simplifications
Distractions
We are only interested in the various patterns of throws andcatches.We will ignore:
I Showmanship (costumes, unicycles, balancing...)
I Props (kerchiefs, balls, rings, clubs, knives, bowling balls,torches, chainsaws...)
I Variations (inside, outside, piston two-in-one-hand...)
I Flourishes (pirouettes, throwing under the leg or behind theback...)
Stephen Hardy: The Mathematics of Juggling 16
Simplifications
Simplifying Assumptions
I The balls are juggled at a constant rate, throws alternatingbetween right and left hands.
I Patterns are periodic (otherwise it is not a pattern).
I At most one ball gets caught and thrown on every beat. If aball is caught, it is the one which is thrown on that beat. (Weignore the time between catch and throw).
These are simple, asynchronous juggling patterns.At the end of the talk we will mention some generalizations.
Stephen Hardy: The Mathematics of Juggling 17
Representing Juggling Patterns
Throws
I We use numbers to denote throws.
I A throw of height n lands n beats later.
I If n is odd it changes hands.
I If n is even it lands in the same hand it was thrown from.
I A juggling pattern is denoted by a string of numbers.
I The constant string n is the standard way to juggle n balls.
Stephen Hardy: The Mathematics of Juggling 18
Representing Juggling Patterns
Juggling PatternsThe constant string n is the standard way to juggle n balls.
I For odd n this is called a cascade.I For even n this is called a fountain – and you actually just
juggle n/2 balls in each hand independently.
I The circular pattern people think of when they think ofjuggling is called a shower. Since this pattern is asymmetricand requires throwing higher, is actually a bad way to startlearning.
Stephen Hardy: The Mathematics of Juggling 19
Representing Juggling Patterns
Throws (Continued)
The throws 0, 1 and 2 are a bit unusual.
I You juggle zero balls by doing nothing (sometimes performersuse this opportunity to clap or pirouette or catch a propthrown by someone else).
I You juggle one ball by fast horizontal throws back and forth.
I You juggle two balls by just holding them (or very smallvertical throws, but more often this is an opportunity forflourishes, or biting an apple).
Stephen Hardy: The Mathematics of Juggling 20
Representing Juggling Patterns
Throws (Continued)
I Humans can only throw so high accurately, even if ceilingsand wind aren’t an issue.
I When you start even 5 level throws are hard. I can manageabout a 7.
I The world records are around 13.
I Thus it makes sense we can cap our throws to a finite level.
Stephen Hardy: The Mathematics of Juggling 21
Examples
Two-Ball Patterns
I 31 – the 2-ball shower. A bad habit.
I 330 – training for 3 balls.
I 40 – two in one hand - training for 4 balls.
I 501 – the interesting and fairly challenging 2-ball pattern!
Stephen Hardy: The Mathematics of Juggling 22
Examples
Three-Ball Patterns
I 3 – lots of variations, inside/outside/tennis etc.
I 42 – two in one hand. A good opportunity for showmanship.
I 423 – a good first trick, time to bite the apple.
I 522 – how you juggle when you learn.
I 55500 – the flash - add a clap or pirouette to impress.
I 51 – the shower - be sure to practice both ways.
Stephen Hardy: The Mathematics of Juggling 23
Examples
More Three-Ball Patterns
I 504 – kind of weird.
I 441 – the asynchronous box.
I 4414413 – my favorite, fairly easy to do but impressivelooking.
I 50505 – the snake - good training for 5.
I 531 – very pretty when done properly.
Stephen Hardy: The Mathematics of Juggling 24
Examples
Four-Ball Patterns
I 4 – lots of variations: inside/outside/pistons, etc.
I 552 – kind of ugly, not great practice for 5.
I 55550 – better practice for 5.
I 53 – the four-ball half-shower.
I 71 – the four-ball shower.
I 5551 – fun.
I 534, 7531, 7131, 633, 741...
Stephen Hardy: The Mathematics of Juggling 25
Examples
Five-Ball Patterns
I 5
I 73 – the five-ball half-shower.
I 91 – the five-ball shower.
I 64 – three in one hand, two in the other.
I 645, 771, 726, 66661, 663, 744, 77731...
Stephen Hardy: The Mathematics of Juggling 26
Questions
Questions
I What strings of natural numbers are valid juggling patterns?
I How can we generate all valid juggling patterns?
I How many balls are needed to juggle a pattern?
Stephen Hardy: The Mathematics of Juggling 27
Questions
Observations
I The period of a pattern is the minimal length before itrepeats: 33333 = 3 has period 1, 441441 = 441 has period 3.
I Cyclic permutations give the same pattern up to chirality (51vs. 15 changes direction of the shower).
I Not all strings of natural numbers give valid juggling patterns.We can never have n(n − 1) for example:
Stephen Hardy: The Mathematics of Juggling 28
A Theorem
Notation
Let’s tackle the question: given a valid juggling pattern, how manyballs are needed to juggle it?
I We let t(i) be the height of of the throw on the i th beat.That ball lands on the i + t(i)th beat.
I We assume this is a periodic function with some period p:t(i + p) = t(i).
I We want to calculate balls(t), the number of balls needed tojuggle this pattern.
I We define σ(i) = i + t(i)− balls(t).
I This will be a permutation of Z exactly when t(i) is a validjuggling pattern.
I Conversely, given such a permutation we can recover t(i).
Stephen Hardy: The Mathematics of Juggling 29
A Theorem
An Example
Stephen Hardy: The Mathematics of Juggling 30
A Theorem
More Notation
I height(t) = maxi t(i) is the maximum throw height thatoccurs (again we assume this is finite).
I balls(t), the number of balls needed to juggle the pattern t,corresponds to the number of infinite orbits in our diagram(singleton orbits correspond to zero throws, and there are noother possibilities since balls cannot vanish or appear).
Stephen Hardy: The Mathematics of Juggling 31
A Theorem
Theorem (Average Theorem)
I The number of balls needed to juggle a valid pattern is theaverage height of the throws.
I If t is a valid juggling pattern with finite height, then
balls(t) = lim|I |→∞
∑i∈I t(i)
|I |
Where I = {a, a + 1, . . . , b} is an integer interval and|I | = b − a + 1 is the number of elements in that interval.
Stephen Hardy: The Mathematics of Juggling 32
A Theorem
The Proof
Let I be an integer interval with |I | > height(t). Then each balllands at least once in the interval I . Let O be an infinite orbit.
Stephen Hardy: The Mathematics of Juggling 33
A Theorem
Proof Continued
From the diagram, we can see that
|I | − height(t) ≤∑
i∈I∩Ot(i) ≤ |I |+ height(t)
Summing over all the orbits (singleton orbits do not contributebecause then t(i) = 0), we get
balls(t)[ |I | − height(t) ] ≤∑i∈I
t(i) ≤ balls(t)[ |I |+ height(t) ]
Dividing by |I | and taking the limit as |I | → ∞ we get the desired
result: balls(t) = lim|I |→∞
∑i∈I t(i)
|I |�
Stephen Hardy: The Mathematics of Juggling 34
A Theorem
Applications
I Corollary: If the average of a string of natural numbers is notan integer, then that string is not a valid juggling pattern.
I Example: 5342 is not a valid sequence because(5 + 3 + 4 + 2)/4 = 15/4 is not an integer.
I This averaging-to-an-integer condition is not sufficient: 543 isnot a valid juggling pattern, but (5 + 4 + 3)/3 = 4
I However, 534 is a valid juggling pattern.
I We do have a partial converse due to Hall:Theorem: If a string of natural numbers has an integeraverage, then some permutation of it is a valid jugglingpattern.
Stephen Hardy: The Mathematics of Juggling 35
A Theorem
Real Math
I Note that σ(i + p) = (i + p) + t(i + p)− balls(t) =i + t(i)− balls(t) + p = σ(i) + p
I
p∑i=1
(σ(i)− i
)=
p∑i=1
(t(i)− balls(t)
)=
p · balls(t)− p · balls(t) = 0
I Recall the affine Weyl group Ap−1 can be represented asbijective functions f : Z→ Z so that f (i + p) = f (i) + p, and∑p
i=1
(f (i)− i
)= 0.
I Thus the permutations coming from juggling patterns are asubgroup of the affine Weyl group Ap−1.
I In fact, Ehrenborg and Readdy used juggling to calculate thePoincare series of the affine Weyl groups!
Stephen Hardy: The Mathematics of Juggling 36
More Results
More Questions
I How can we generate all of the valid juggling patterns? Howmany are there?
I Site Swaps: We switch the landing sites of two differentthrows to get a new valid pattern with the same number ofballs.
Stephen Hardy: The Mathematics of Juggling 37
More Results
Site Swap Example
Stephen Hardy: The Mathematics of Juggling 38
More Results
More Algorithms
I If our pattern is not the constant pattern, then there must bea throw which is higher than the average and other throwwhich is lower than the average. By swapping the landingsites of a higher throw with a lower throw, we get a patternwhich is closer to a constant pattern.
I Since the throw heights and period are finite, in finitely manysteps we must be able to flatten our valid sequence to theconstant sequence with the same number of balls.
I The Flattening Algorithm: By applying site swaps andcycling permutations, we can take any valid juggling patternwith b balls to the constant sequence b.
Stephen Hardy: The Mathematics of Juggling 39
More Results
More Algorithms Continued
I Since site swaps and cyclic permutations are reversible, theseoperations actually generate all the valid juggling patternswith a given period and number of balls!
I Theorem: The constant pattern consisting of the throw brepeated p times can be transformed into any valid jugglingpattern of length p for b balls via site swaps and cyclicpermutations.
Stephen Hardy: The Mathematics of Juggling 40
More Results
ExamplesI The 3-ball period 2 patterns.
I The 3-ball period 3 patterns.
I Mathematics brought new patterns into juggler’s repertoires!
Stephen Hardy: The Mathematics of Juggling 41
More Results
More Results
I How many valid juggling patterns are there?
I Theorem: The number of all minimal b-ball juggling patterns(modulo cyclic permutations) of period p is
N(b, p) =1
p
∑d |p
µ(pd
) ((b + 1)d − bd
)Where µ is the Mobius function
µ(n) =
0, if n has repeated prime factors
1, if n = 1 or n has an even number of prime factors
−1, if n has an odd number of prime factors
Stephen Hardy: The Mathematics of Juggling 42
Final Thoughts
Generalizations
I Synchronous patterns – throw and catch with both handssimultaneously.
I Multiple hands – allow for passing between people.
I Multiplexing – throw and catch multiple balls in the samehand at the same time.
I Braids – attach each ball to a string then tie other endstogether. Then patterns correspond to braids!
Stephen Hardy: The Mathematics of Juggling 43
Final Thoughts
Final Thoughts
I Math is fun.
I Juggling is fun.
I Try playing withhttp://www.siteswap.net/JsJuggle.html
Thank You For Listening!
Stephen Hardy: The Mathematics of Juggling 44