Poi as a Sport-Juggling Prop

Conventions, degrees of freedom, and the construction of comparable technical practice.

Boris Fontanov · 2 August 2026, Helsinki · continues On the Modality of Spinning

I keep returning to a small scene: three poi in the air, a rhythm everyone in the room would call a cascade, and two people arguing politely about whether they just watched the same trick. One saw backward no-beat. The other saw a pendulum that cleaned itself with a half-orbit. Both were right about the beat. Neither had named the kinetic task.

The previous essay argued that siteswap remains useful for poi as a rhythmic grid while leaving flexible-prop movement under-described. Siteswap records when an object must next be acted upon. How the object moves between interactions, how torque is maintained, which plane is occupied, and what the body does in the interval stay outside that grid. In informal practice that incompleteness is often harmless, because demonstration fills the gap; in graded or competitive systems the same gap becomes structural. A judge needs a finite statement: this performer completed this defined task under these conditions.

Poi resist that sentence. The prop may rotate while held, reverse during dwell, change plane without a release, arrive in several phases, or recover through tether deformation what would have failed with a rigid club. The problem is therefore how to conventionalise poi for sport without reducing them to clubs with soft shafts, still less to rings or balls.

Sport as controlled comparison

Sport juggling may be read as a restricted technical practice in which performances become comparable through shared tasks, declared conditions, repeatable counts, and observable criteria. The World Juggling Federation Skill Level Achievement Program is one working example of that structure: basic runs, named tricks, pirouettes, body throws, connections, and higher object counts, with pattern quality and body control treated as part of execution. Such scaffolding can be adapted to poi. Its difficulty assumptions for balls and clubs need a fresh reading against tethered mechanics.

With balls, rings, and clubs there is a broad distinction between the period in the hand and the flight of the object. Clubs require spin control, yet they remain approximately rigid bodies with a stable relationship between handle, centre of mass, and principal axis. Poi have no comparable fixed geometry. Head, tether, and handle may occupy different trajectories; the effective shape changes continuously. Comparison therefore needs agreement about which dimensions are fixed, which may vary, and which variations constitute separate skills.

The first restriction: juggling, not spinning

The initial convention should be categorical: sport-poi requirements must contain repeated throws and dwell, while continuous spinning alone stays outside that core. Weaves, flowers, stalls, contact sequences, wraps, and choreography without repeated throws and dwell remain available, but they do not belong to the sport; they may appear inside transitions, yet they are excluded from the core definition because they leave the flight-and-dwell structure that toss juggling shares.

Spin and weave already have a mathematical state treatment of their own. Iivari Varpanen’s toss-and-spin state graphs model continuous spinning structure; that work supports description, while the sport syllabus proposed here still has to choose a narrower throw-and-dwell surface for comparison.

Once repeated throws and dwell are required, the word cascade becomes partially recoverable: a three-poi pattern may share siteswap 3 with a three-ball cascade, but the word still names only the rhythm. Under that number lives a family of kinetic tasks — pendulum 3, backward no-beat 3, wallplane synchronous 3, backward 3 with an additional spin, forward 3, synchronous forward 3 — so the rhythmic identity is shared while the kinetic task is not.

Excess degrees of freedom

A rigid prop keeps a comparatively stable geometry. A poi has at least three moving elements (handle, tether, head), and the tether may be straight, curved, slack, winding, or briefly folded. Two catches that look similar at the handle may arrive with different head trajectories and different remaining angular momentum.

In ordinary siteswap reading, a 2 is often a hold: the object waits until its next scheduled action. For poi, dwell is rarely inert. During the interval before release the performer may maintain rotation, accelerate or decelerate the head, add a spin, reverse direction, redirect the plane, alter tether tension, or change the body’s orientation relative to the prop. The dwell phase belongs to the technical pattern.

Living practice already treats that fact as convention. In Bow’s public set titled for a hundred three-poi siteswaps, the accompanying catalogue frames every pattern in backward no-beat technique, treats the digit 2 as always spinning, and treats the digit 1 as passed with spin. The siteswap strings change from beginner cascade to high throws, while the modality and those digit semantics stay fixed across the set: a layered descriptor spoken aloud by performance.

Plane freedom compounds the problem. Practice usually names the two base orientations wheelplane and wallplane; relative to the body they read naturally as side and front or back planes. Movement may be organised in wheelplane, wallplane, horizontal, diagonal, overhead, or mixed planes — and those names are always read in the relation of body and prop, not of the room. A separate sport question arises when the body turns while the poi orbits stay nearly fixed in space: the body-relative name changes (side becomes front) even though the prop’s angular momentum was not redirected. Flexible props also permit corrections rigid props refuse: late catches absorbed through tether curvature, temporary plane breaks, additional orbits that restore timing. Those capabilities are valuable in practice, but in sport an undeclared recovery can look like a successful pattern: the judge sees the clean-up, not that the required control was already lost.

A layered descriptor, not a universal notation

A complete geometric notation for poi is imaginable and unsuitable as the ordinary language of a sport programme. Extensions such as Beever’s Generalised Siteswap, surveys of layered systems by Daniel Simu, and Rauli Katajavuori’s graphic notation already show that layered description is possible and that expressive power raises reading cost. Sport poi should use a limited descriptor: object count, siteswap or rhythmic pattern, movement modality, rotation and extra spin, phase or synchrony, plane orientation in space, body orientation, body-relative plane class, entry and exit, throw count.

objects: 3 poi
siteswap: 3
modality: backward no-beat
phase: asynchronous
in space: two parallel vertical planes
body-relative: wheelplane
entry: cold start
count: 30 throws
exit: clean collection

The descriptor records the parameters needed to distinguish graded tasks, leaving centimetres of motion to the reference video. Informal record culture already gestures in the same direction. On Juggling Edge poi leaderboards, trick names split cascade from cascade singles, halves, and flat fronts: siteswap alone was never enough for the people keeping throw counts.

A poi-specific difficulty order

Copying club or ring progression produces false assumptions. In ordinary curricula a one-up 180 may arrive before a one-up 360 because it asks for less body rotation. With poi, a 180 changes the performer’s orientation relative to the rotating prop. A backward-spinning poi thrown normally before a 180 becomes forward relative to the body after the turn. A 360 restores the original orientation and may preserve the original backward modality without further intervention. A natural one-up 180 is therefore often a backward-to-forward transition; a backward-to-backward 180 needs a flat throw, overspin, reversal, or post-catch correction. In poi, one-up 360 can be technically simpler than one-up 180.

The same asymmetry appears in the cascade family. A programme preference proposed here places pendulum 3 beside the foundation, then backward no-beat 3, wallplane synchronous 3, backward 3 with additional spin and synchronous relation, wallplane halfshower, forward 3, then synchronous forward 3. Pendulum-first remains provisional; the sharper living signal is that forward no-beat arrives after backward no-beat. The difference arises from the path of hands, elbows, and shoulders on the throw, from rotation and release phase, and from hand position at the catch, not from the spelling of the word forward.

Teaching culture already supports that sharper signal. Bow’s published No Beat Course opens on no-beat cascade before forward cascade, then windmill, basic siteswap, halfshower, and four poi. The hundred-siteswap video builds advanced rhythm under a single backward no-beat framing in its catalogue. Living curriculum is evidence that a sport programme can grow siteswap depth inside a fixed modality; it is still far from a federation table.

Planes in space, body names, and transition class

Names such as wheelplane and wallplane always belong to the relation of body and prop. For comparing transitions, sport needs one further record: whether the poi orbits stay nearly fixed in space while the body turns. The same physical orbit, renamed relative to the body, changes class when the performer turns; that is not a second owner of the prop, but a way to tell a body turn from a redirection of the planes. The distinction yields three transition mechanisms that can look identical at the destination and differ completely as skills.

In an orbit-preserving, body-mediated transition the poi planes stay approximately fixed in space while the body turns ninety degrees: the body-relative plane class changes — for example from wheelplane to wallplane. The orbits continue in space, but relative to the body the task is already different: a no-beat cascade in wheelplane and a no-beat cascade in wallplane are separate graded tasks, and a halfshower in either plane is a different pattern again. In a prop-mediated controlled plane break the body stays approximately fixed while the poi planes are redirected ninety degrees by specific hand movements: the same shift of body-relative plane class, reached by changing angular momentum. In a coupled transition both change together: a ninety-degree body-relative shift might be forty-five of body and forty-five of prop plane.

Control therefore means more than tidy planes. Advanced control includes the ability to retain a plane against disturbance and to break a plane while keeping authorship of the movement. A performer fails when the plane changes contrary to the declared task; a plane-preserving requirement also fails when the props are turned and body motion disguises the redirection.

Pirouette parity and numbers

Pirouettes should be divided into orientation-preserving and orientation-inverting classes. A turn through an even number of half-turns returns the performer to the original facing; an odd number inverts it. For normally rotating poi, even-half-turn pirouettes tend to preserve the body-relative direction class; odd-half-turn pirouettes tend to invert it. That default relation must be stated. Recognised odd-turn variants are separate graded skills: accept the direction inversion (backward → forward), keep the original direction with a flat throw, correct with overspin, or restore direction by a post-catch reversal. The phrase “one-up 180” by itself is too thin.

Numbers progression should keep three poi as the core vocabulary, four as intermediate density, and five as the highest principal number, while mixing fundamentals, siteswaps, pirouettes, body throws, and plane tasks across grades. Four begins after the three-poi foundation. Five distinguishes flash from stable control, acquisition from endurance, and modality change from mere survival. IJA Numbers Championships remain organised around balls, rings, clubs, and bounced balls; poi’s sport home, if it arrives, will be built rather than inherited. Poi juggling already has public recognition elsewhere: Chris Kelly’s IJA Tricks of the Month and David Cain’s history of poi in the juggling community record an “unexplored” juggling frontier and a late entry into juggling culture.

Equipment, freedom, and reference

Poi vary more than competition balls and often more than clubs or rings: tether length, diameter, and elasticity, head mass and diameter, handle mass and type, swivel presence, drag. Complete standardisation would ease comparison, but would sharply narrow the available choice of equipment and of the technical forms that go with it. A better first convention is declaration plus bounded classes: short, medium, and long tether; permitted head-mass range; limited elasticity; no automatic retractors; declared swivel use; equivalent props within one numbered pattern. Records should state the equipment, and variation should be visible rather than moralised.

Conventionalisation should stop before it becomes total bodily prescription. Exact stance, personal arm length, minor elbow position, aesthetic line, facial orientation, small tempo variation, and harmless height differences ordinarily remain free. Judging detects loss of the required technical structure and leaves anatomy and stylistic school outside the grade. Because poi movement is visibly continuous, judges are easily tempted to reward a preferred appearance. Judging criteria should separate pattern, modality, phase, plane, rotation, pirouette result, catch region, transition, correction policy, and authorship of the movement.

Practitioners see these flexible-prop failures clearly: an extra spin inserted to recover timing; a brief pendulum used to escape a no-beat pattern; an unintended body turn replacing a prop-plane change; a stall resetting phase; a wrap absorbing a late catch; a short weave or contact carry standing in for a required reception. The faults still need shared names, or judging collapses into familiarity and private intuition.

Written definitions alone are insufficient. Every graded requirement needs a clear reference execution, a slow technical view, a normal-speed qualification view, a diagram of body-relative names and plane orientation in space, common failure modes, and at least one workable alternative bodily realisation where appropriate. For plane transitions, two camera angles matter, because a frontal view can hide depth drift; what is needed is the visible boundary between acceptance and rejection, not the perfect demonstration alone.

Attestation, limits, selective limitation

A technical grade should attest an exact body of evidence under a stated programme version. The certificate records completed tasks, leaving the wider value of the practitioner outside its scope. Poi carry large technical branches outside toss juggling. Contact, tether topology, spinning composition, and group movement may coexist with a modest sport grade; high numbers may coexist with thin compositional or pedagogical skill. The certificate carries only a limited set of facts: programme version, grade, completed requirements, equipment, evidence, reviewers, date, and whether the claim is a historical achievement or a current capability. The title means little without the video: it merely points to the evidence by which the claim can be checked.

No convention will describe all available poi juggling. Unusual tether lengths, mixed-prop ensembles, contact-to-toss hybrids, deliberate slack, compositional difficulty, and continuous transformations without stable named states will remain underrepresented. Because poi permit many paths between two states, an exact standard may begin to reward compliance with the chosen route rather than general control. That danger shrinks when the syllabus keeps fixed requirements beside equivalent-transition comparisons, plane-preserving and plane-breaking variants, constrained improvisation, and optional experimental distinctions.

The excess flexibility of poi is the source of the discipline’s technical identity. Rules that erase active dwell, rotation, plane transformation, body-relative reclassification, and tether-mediated correction leave poi looking like soft clubs. Rules that leave those quantities undefined leave performances hard to compare. What is needed is not to forbid that flexibility wholesale, but to limit it selectively: for each task it should be clear in advance which freedoms are fixed, which remain open, and which transformations are being tested. The sport object of poi is always narrower than the physical prop and the whole available practice around it.

Working assumptions

A first stable sport-poi programme can begin from a short list of working assumptions — held open to revision and discussed in public when they change.

Only repeated throw-and-dwell patterns belong to the core discipline.
Three poi define the foundation; four intermediate density; five the highest principal number.
Backward no-beat 3 is the foundational qualifying cascade; pendulum 3 and wallplane synchronous 3 are proposed basic companions.
Backward 3 with an additional spin remains rhythmically 3 and is graded separately.
Forward 3 arrives after backward no-beat; synchronous forward arrives later.
One-up 360 is introduced before one-up 180; for odd-half-turn pirouettes it is stated in advance whether direction inversion is accepted or must be corrected.
Plane orientation in space and body-relative class (wheelplane, wallplane) are recorded separately; orbit preservation through a body turn and prop-mediated plane breaks are different skills.
Siteswap is the rhythmic layer, not the complete trick definition.
Every requirement has a reference execution and examples of typical faults; the equipment in use is stated explicitly.
A certificate confirms completion of specific tasks, not the value of the practitioner as a whole.

These conventions are intentionally incomplete: their purpose is to make the first shared comparisons possible.

Closing

Bringing poi into sport juggling asks for more than a new prop column beside balls, rings, and clubs. Poi transform the relation between event and movement: the throw no longer contains the whole technical act, and the period before release may hold sustained rotation, redirection, acceleration, phase correction, and changes in the relation between body and plane, so the same siteswap may name several tasks with different difficulty.

For that reason poi need a layered sport language in which the rhythmic pattern identifies when interaction occurs; modality, how the prop rotates; phase, the relation among hands and poi; plane, where movement occurs; body orientation determines how that plane is perceived; and transition class specifies whether the body, the prop planes, or both are transformed.

Object-first research in contemporary circus juggling has long insisted that a trick is movement and the physical object together: shape, material, and the history of why a prop is the size it is. Sport poi inherits that insistence as equipment declaration and digit semantics. Methodologies that break patterns into components and recombine them (reduct-construct in Erik Åberg’s terms) already describe how practitioners multiply vocabulary under a shared rhythm. The sport task is to choose which of those recombinations become comparable grades.

The greater flexibility of poi is a reason to construct the discipline carefully from the beginning. A working convention should stay narrow enough to support comparison and modest enough to admit that it governs only one region of practice. Technical convention is a surface for comparison, not an ontology of poi; poi can enter sport juggling when their freedoms are declared, and the central task is to determine which of their many possible movements are being compared, and to make that comparison honest.

This essay continues On the Modality of Spinning.