One Shot, Twenty-One Balls: A Computer-Assisted Proof of a Single-Stroke Total Clearance in an Idealized Model of Snooker
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Computer Science. The day in computer science: software, systems and theory on arXiv.
Tom: We're looking at the paper: "One Shot, Twenty-One Balls".
Jane: Here's the gist of the paper. One stroke can pocket all twenty-one object balls in an idealized model of snooker, and this successful set has a positive Lebesgue measure in the natural shot space pg1.
Tom: This finding is significant because it shows that while a total clearance is not a fluke of measure zero, it occupies genuine volume in the shot space pg1.
Jane: The Mathematical Model The authors construct a fully explicit model called Model M to analyze billiard dynamics pg3.
Tom: This model specifies the table dimensions, ball radii, and the dynamics between events such as deceleration and collision laws pg3.
Paper summary: Tom: The paper "One Shot, Twenty-One Balls: A Computer-Assisted Proof of a Single-Stroke Total Clearance in an Idealized Model of Snooker" shows that total clearance is possible in this idealized model and forms a set with positive Lebesgue measure >
Jane: They prove it using a computer construction to establish the existence of these margins and show that total clearances are not just random events with zero measure >
Lu: The implication for theoretical work is pointing that tradition at a bar bet, showing how the mathematical framework can guarantee things that are otherwise just folklore or intuition >
Conclusion: Tom: So we're wrapping up this look at the paper "One Shot, Twenty-One Balls: A Computer-Assisted Proof of a Single-Stroke Total Clearance in an Idealized Model of Snooker." Basically, they’ve used a computer to prove that in this idealized snooker model, you can pocket all twenty-one object balls with just one stroke.
Meng: And from my side, I see the issue immediately. The paper mentions that while the set of shots leading to twenty-one balls has positive measure, the actual neighborhood around that successful shot is "extraordinarily thin." It's like trying to hit a needle in a haystack when the needle exists somewhere in that hay >
Lalam: That’s exactly what they’re pointing out. The mathematical work guarantees those margins exist at all, but for twenty-one balls, those margins are so fine they can't be seen or played by a human cue ball with any realistic precision >
Jane: So the authors are essentially saying that trick-shot artists operate in this weird middle ground between something that is mathematically real and something that is physically unobservable.
Tom: Exactly. The paper doesn't claim it’s a "holy grail" of snooker; they’re just showing that the mathematical structure allows for it. It turns the conversation around from "Can this be done?" to "How small does it have to be to be real?"
Lu: And they set up some serious follow-up work on that. They suggest looking at subset simulations, where you check if you can get one or two balls, then three, and so on. If they can find a way to prove the decay law holds better for smaller numbers, that would give us more hope about the scaling >
Meng: From an engineering view, I'm interested in how much computational power it takes to actually verify those witness shots they built. It seems like the construction itself is very sensitive, so verifying it at this level requires a lot of computation >
Jane: It’s interesting because it shows that sometimes the most important part of a math paper isn't just the big result, but how they frame the limits and what comes next for real-world application. The question becomes whether we can bridge that gap between theory and physics.
Israel Central Bureau of Statistics
cs.CG, math.DS, nlin.CD
Submitted: 2026-07-14
Updated: 2026-11-15
Comments: 20 pages, 1 figure. v2: the witness of v1 was found not to survive exact arithmetic and is replaced by a certified one; adds the width and the lower edge of the ribbon, a second configuration with a wide window, and the semialgebraic structure. Code, data and certificates: https://github.com/avnerkantor/one-shot-twenty-one-balls (doi:10.5281/zenodo.23816267)
License: http://creativecommons.org/licenses/by/4.0/
Importance score: 85/100
The gist: One stroke can pocket all twenty-one object balls in an idealized model of snooker, and this successful set has a positive Lebesgue measure in the natural shot space pg1.
Key concepts
- Positive Lebesgue Measure
- This mathematical concept indicates that the set of successful shots leading to twenty-one pockets occupies a genuine volume in the shot space. It means there are infinitely many such shots clustered together, not just isolated points, suggesting the clearance is mathematically possible within the model.
- Witness Configuration
- A witness is a specific, verified configuration and shot that achieves the desired outcome—pocketing all twenty-one balls in one stroke. The paper constructs this using a single chain of collisions where every ball ends up in its target pocket, proving the clearance is physically achievable.
- Measure Zero vs. Unobservably Small
- This distinction highlights the core finding: while the set of successful shots has positive measure (mathematically possible), it is 'unobservably small' because sensitivity compounds multiplicatively. This means that although a clearance exists mathematically, its practical existence for a human player is negligible.
- Model M
- This is a fully explicit mathematical model used to simulate the physics of the billiard table. It precisely defines table dimensions, ball sizes, and the rules governing collisions like deceleration and friction. This rigorous model allows the authors to prove existence and analyze shot dynamics.
Terminology
Summary
One stroke can pocket all twenty-one object balls in an idealized model of snooker, and this successful set has a positive Lebesgue measure in the natural shot space pg1. p. 0 We
examine the claim in an idealized but fully specified model of billiard dynamics. Within the
model we exhibit an admissible configuration of the twenty-two balls and a stroke of the
cue ball that pockets all twenty-one object balls, and we show that the set of such strokes
has positive Lebesgue measure in the natural shot space: total clearances are not flukes
of measure zero but open events. Claim (iii) is true beyond reasonable
doubt, and we will quantify just how true. Claim (i) is false in any reasonable idealization
of the game, and the point of this article is that it fails in a robust way in principle: the
successful strokes form a set of positive measure, so a total clearance is not a lone knife-edge
coincidence but occupies genuine volume in the shot space.Where the paper says it
Where the paper says it
The Mathematical Model
The authors construct a fully explicit model called Model M to analyze billiard dynamics pg3. This model specifies the table dimensions, ball radii, and the dynamics between events such as deceleration and collision laws pg3. p. 1 Including spin would enlarge the shot space by two or more dimensions and
complicate the collision law, but it would not change the logic of anything below: the existence
theorem needs only that the final state depends continuously on the shot near one well-chosen
shot, and that argument is indifferent to the dimension of the shot space or the detailed smooth
collision law. A witness in the spinless model, struck with zero spin, remains a witness in any
enlarged model whose dynamics agrees with Model M on the zero-spin slice. In other words, for the literal opening rack, N(c0, x)
is not defined by the model until an arbitrary tie-breaking convention is chosen. We therefore
work throughout with the ε-separated opening configuration c
ε
0
, the regulation positions with
the reds spread by ε = 0.Where the paper says it
Where the paper says it
Existence of a Witness
The paper establishes the existence of a witness configuration and shot that achieves a total clearance, which is robustly called regular pg3. Lemma 2 provides a construction for this witness using a single-carrier chain where the cue ball strikes twenty-one object balls in succession pg4. p. 2 The witness is a single-carrier chain. The cue ball is the only carrier: it strikes
twenty-one object balls in succession, and at each collision the struck ball departs along the
line of centers aimed at a pocket center while the cue is deflected onward to the next ball. The cue ball is the only carrier: it strikes
twenty-one object balls in succession, and at each collision the struck ball departs along the
line of centers aimed at a pocket center while the cue is deflected onward to the next ball. No
object ball ever strikes another; every pocketing is a struck ball sent directly to a pocket by
the cue, and the cue itself is not pocketed.Where the paper says it
Where the paper says it
The Measure vs. p. 0 We
examine the claim in an idealized but fully specified model of billiard dynamics. Within the
model we exhibit an admissible configuration of the twenty-two balls and a stroke of the
cue ball that pockets all twenty-one object balls, and we show that the set of such strokes
has positive Lebesgue measure in the natural shot space: total clearances are not flukes
of measure zero but open events.Where the paper says it
Theorem 1 proves that for an admissible configuration c∗, the set of shots leading to N=21 contains a nonempty open set, implying positive Lebesgue measure pg4. However, Remark 4 notes that this guaranteed neighborhood is extraordinarily thin
and unobservably small because sensitivity compounds multiplicatively across twenty-one successive cuts pg8. This illustrates the gap between positive measure
and unobservably small
pg1.
Computational Decidability
The paper separates two computational points regarding the conjecture pg4. First, the conjecture is decidable in principle by simulation, as a single certified trajectory can prove it via Lemma 1 pg4. p. 4 Two computational points deserve separation. First, the conjecture is decidable in principle
by simulation: a single stroke x
∗
found by any search whatsoever, then certified by interval
arithmetic to be regular with N = 21, proves it, by Lemma 1. Second, the conjecture is not decidable by
direct Monte Carlo: no feasible number of random strokes distinguishes an event of probability
10−40 from an impossible one. Sampling can only ever bound P(21) from above; a single
certified trajectory settles existence.Where the paper says it
Where the paper says it
Rarity Estimation
The authors estimate the probability P(k) that a uniformly random stroke pockets exactly k object balls using direct simulation pg5. The observed decay of P(k) suggests an extrapolated value for P(21) around 10-12, conditional on Conjecture 1 pg6. p. 4 First, the log-linear decay visible over the populated range (a straight line of slope b ≈ −0. 57 per ball on the log10 scale, the largest observed
value being k = 7) invites extrapolation to k = 21, and we report the extrapolated value,
∼ 10−12, but its logical status must be kept straight: extrapolation cannot distinguish a tiny
positive P(21) from an exact zero, so the number is meaningful only conditional on Conjecture 1, and even then it assumes the decay law persists through a regime the data never probes. From the opening break it is, we believe, also possible, and settling
that belief is a well-posed finite computation that current search methods cannot yet perform
(Section 4). And it will never be observed, because its probability, if positive, is protected by
an exponential decay that our experiments measure in its shallow end (Section 5).Where the paper says it
Where the paper says it
Discussion and Conclusion
The folklore is summarized as three true sentences regarding the total clearance pg7. The paper concludes that a total clearance is possible in the model, forms a set of positive measure there, but will never be observed due to exponential decay pg7. p. 6 6 Discussion
The folklore is best restated as three true sentences. A total clearance in one stroke is possible
in the model, and forms a set of positive measure there—robust in principle, though for the
twenty-one-ball witness that robustness is of unobservably small width (Theorem 1 and the
remark following it). 6 Discussion
The folklore is best restated as three true sentences. A total clearance in one stroke is possible
in the model, and forms a set of positive measure there—robust in principle, though for the
twenty-one-ball witness that robustness is of unobservably small width (Theorem 1 and the
remark following it). And it will never be observed, because its probability, if positive, is protected by
an exponential decay that our experiments measure in its shallow end (Section 5). The bar bet,
in other words, confuses measure zero with unobservably small.Where the paper says it
Where the paper says it
Where the paper says it
The Witness
Figure 2 provides a verified witness showing the cue ball striking all twenty-one object balls in one stroke pg8. The verification protocol requires that exactly the planned collision pairs occur and that all twenty-one object balls are pocketed pg8. p. 7 All 21 object balls pocketed (verified end to end); minimum pairwise ball separation 154 mm; pocket-entry speeds from 0. The verification protocol: the layout is replayed in the event-driven integrator; we require
that exactly the planned collision pairs occur, that all twenty-one object balls are pocketed,
and we report the minimum gap between consecutive events, the minimum pocket-entry speed,
and the minimum clearance attained by any pair of balls not designed to touch. All 21 object balls pocketed (verified end to end); minimum pairwise ball separation 154 mm; pocket-entry speeds from 0. The verification protocol: the layout is replayed in the event-driven integrator; we require
that exactly the planned collision pairs occur, that all twenty-one object balls are pocketed,
and we report the minimum gap between consecutive events, the minimum pocket-entry speed,
and the minimum clearance attained by any pair of balls not designed to touch. All 21 object balls pocketed (verified end to end); minimum pairwise ball separation 154 mm; pocket-entry speeds from 0. The verification protocol: the layout is replayed in the event-driven integrator; we require
that exactly the planned collision pairs occur, that all twenty-one object balls are pocketed,
and we report the minimum gap between consecutive events, the minimum pocket-entry speed,
and the minimum clearance attained by any pair of balls not designed to touch. All 21 object balls pocketed (verified end to end); minimum pairwise ball separation 154 mm; pocket-entry speeds from 0. The verification protocol: the layout is replayed in the event-driven integrator; we require
that exactly the planned collision pairs occur, that all twenty-one object balls are pocketed,
and we report the minimum gap between consecutive events, the minimum pocket-entry speed,
and the minimum clearance attained by any pair of balls not designed to touch. The verification protocol: the layout is replayed in the event-driven integrator; we require
that exactly the planned collision pairs occur, that all twenty-one object balls are pocketed,
and we report the minimum gap between consecutive events, the minimum pocket-entry speed,
and the minimum clearance attained by any pair of balls not designed to touch. All three
margins are far above double-precision rounding scales, which is the sense in which Theorem 1
is computer-assisted; an interval-arithmetic certificate would be a finite, routine upgrade.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Reproduction
All results are produced by short, dependency-free Python scripts around a single eventdriven implementation of Definition 2 pg8. This includes the engine, the build-and-verify constructor, and the Monte Carlo analysis pg8. p. 7 py (the build-and-verify constructor,
which grows the chain one collision at a time inside the engine and emits the witness of Appendix A), mc break. py (the Monte Carlo, the sensitivity grid, and the table above), and
make figures. Remark 3 (Status of the verification). The witness is verified numerically, in double precision: the shot x
∗
is replayed in an event-driven integrator of Definition 2 and produces exactly
twenty-one pocketings, with the individual margins that enter regularity (pocket-entry speeds,
inter-ball clearances, event-time gaps) all strictly positive and many orders of magnitude above
the rounding scale.Where the paper says it
Where the paper says it
References
The paper cites foundational work on dynamical systems and geometry in billiards pg9. Specifically, it references Sinai's work on dynamical systems with elastic reflections pg9. It also cites Tabachnikov's book on Geometry and Billiards pg9, and Ballard's work on discrete mechanical systems with perfect unilateral constraints pg9. Additionally, it references Kiefl's Python package for realistic billiards simulation pg8. The paper also cites a reference to pooltool [4] for cross-checking the witness in a full-physics engine pg8. p. 7 Seeds are fixed in the scripts. Readers who want a full-physics cross-check
including spin and sliding friction can replay the witness in an engine such as pooltool [4]; by
Remark 1 the zero-spin witness is the right object to test.Where the paper says it
Model Instance
The specific parameters used for the model are listed in the Model instance section pg8. These parameters include L = 3. p. 3 Lemma 1 guarantees that the success set around a
regular witness is open, hence of positive Lebesgue measure; it says nothing about how large
that measure is. For our twenty-one-collision witness the guaranteed neighborhood is in fact
extraordinarily thin: scanning the shot parameters (V0, φ) outward from x
∗
, the interval on
which all twenty-one balls still fall shrinks below the resolution of a double-precision grid (a
change of order 10−6 m/s in V0 already spoils the clearance). The table is the closed rectangle T = [0, L]×[0, W] with
L = 3. 569 and W = 1. The table is the closed rectangle T = [0, L]×[0, W] with
L = 3. 569 and W = 1. 569 and W = 1. 778 (meters; the playing area of a full-size table). 569 and W = 1. 778 (meters; the playing area of a full-size table). 778, R = 0. 02625, a = 0. 778, R = 0. 02625, a = 0. A ball is pocketed (and removed from the table) at
the first instant its center enters the open disk of radius ρ = 0. 055 about a pocket point. A ball is pocketed (and removed from the table) at
the first instant its center enters the open disk of radius ρ = 0. 055 about a pocket point. A shot is a pair x = (V0, φ) ∈ Ω = (0, Vmax] × [0, 2π)
with Vmax = 12 m/s: the cue ball is set in motion with speed V0 in direction φ. Given an
admissible c and a shot x, the dynamics of Definition 2 runs until all balls are at rest or
pocketed; N(c, x) ∈ {0, 1, . What mathematics adds is the guarantee that the margins exist at all, and a clear
account of why the guarantee for the break position, claim (ii), remains open: it sits exactly
where verification is easy and search is hopeless, a combination that will be familiar to anyone
who has thought about needles in high-dimensional haystacks. A The witness
cue
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15 16
17
18
19
20
21
Verified 21-ball single-stroke clearance (cue path in red)
Figure 2: The verified witness: the cue ball (open) strikes all twenty-one object balls (dark) in
one stroke, its path shown in red; each struck ball’s route to its target pocket is dotted. All 21 object balls pocketed (verified end to end); minimum pairwise ball separation 154 mm; pocket-entry speeds from 0. The verification protocol: the layout is replayed in the event-driven integrator; we require
that exactly the planned collision pairs occur, that all twenty-one object balls are pocketed,
and we report the minimum gap between consecutive events, the minimum pocket-entry speed,
and the minimum clearance attained by any pair of balls not designed to touch. Remark 3 (Status of the verification). The witness is verified numerically, in double precision: the shot x
∗
is replayed in an event-driven integrator of Definition 2 and produces exactly
twenty-one pocketings, with the individual margins that enter regularity (pocket-entry speeds,
inter-ball clearances, event-time gaps) all strictly positive and many orders of magnitude above
the rounding scale. Remark 3 (Status of the verification). The witness is verified numerically, in double precision: the shot x
∗
is replayed in an event-driven integrator of Definition 2 and produces exactly
twenty-one pocketings, with the individual margins that enter regularity (pocket-entry speeds,
inter-ball clearances, event-time gaps) all strictly positive and many orders of magnitude above
the rounding scale. The verification protocol: the layout is replayed in the event-driven integrator; we require
that exactly the planned collision pairs occur, that all twenty-one object balls are pocketed,
and we report the minimum gap between consecutive events, the minimum pocket-entry speed,
and the minimum clearance attained by any pair of balls not designed to touch. All three
margins are far above double-precision rounding scales, which is the sense in which Theorem 1
is computer-assisted; an interval-arithmetic certificate would be a finite, routine upgrade.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
The Shot Parameters
The specific shot parameters used for the witness are detailed pg8. The cue is set at (0. p. 0 We
examine the claim in an idealized but fully specified model of billiard dynamics. Within the
model we exhibit an admissible configuration of the twenty-two balls and a stroke of the
cue ball that pockets all twenty-one object balls, and we show that the set of such strokes
has positive Lebesgue measure in the natural shot space: total clearances are not flukes
of measure zero but open events. We
examine the claim in an idealized but fully specified model of billiard dynamics. Within the
model we exhibit an admissible configuration of the twenty-two balls and a stroke of the
cue ball that pockets all twenty-one object balls, and we show that the set of such strokes
has positive Lebesgue measure in the natural shot space: total clearances are not flukes
of measure zero but open events. Remark 3 (Status of the verification). The witness is verified numerically, in double precision: the shot x
∗
is replayed in an event-driven integrator of Definition 2 and produces exactly
twenty-one pocketings, with the individual margins that enter regularity (pocket-entry speeds,
inter-ball clearances, event-time gaps) all strictly positive and many orders of magnitude above
the rounding scale. Remark 3 (Status of the verification). The witness is verified numerically, in double precision: the shot x
∗
is replayed in an event-driven integrator of Definition 2 and produces exactly
twenty-one pocketings, with the individual margins that enter regularity (pocket-entry speeds,
inter-ball clearances, event-time gaps) all strictly positive and many orders of magnitude above
the rounding scale. Remark 3 (Status of the verification). The witness is verified numerically, in double precision: the shot x
∗
is replayed in an event-driven integrator of Definition 2 and produces exactly
twenty-one pocketings, with the individual margins that enter regularity (pocket-entry speeds,
inter-ball clearances, event-time gaps) all strictly positive and many orders of magnitude above
the rounding scale. What mathematics adds is the guarantee that the margins exist at all, and a clear
account of why the guarantee for the break position, claim (ii), remains open: it sits exactly
where verification is easy and search is hopeless, a combination that will be familiar to anyone
who has thought about needles in high-dimensional haystacks. A The witness
cue
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15 16
17
18
19
20
21
Verified 21-ball single-stroke clearance (cue path in red)
Figure 2: The verified witness: the cue ball (open) strikes all twenty-one object balls (dark) in
one stroke, its path shown in red; each struck ball’s route to its target pocket is dotted. What mathematics adds is the guarantee that the margins exist at all, and a clear
account of why the guarantee for the break position, claim (ii), remains open: it sits exactly
where verification is easy and search is hopeless, a combination that will be familiar to anyone
who has thought about needles in high-dimensional haystacks. A The witness
cue
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15 16
17
18
19
20
21
Verified 21-ball single-stroke clearance (cue path in red)
Figure 2: The verified witness: the cue ball (open) strikes all twenty-one object balls (dark) in
one stroke, its path shown in red; each struck ball’s route to its target pocket is dotted.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Margins
The margins for the verified witness are explicitly detailed pg8. All twenty-one object balls pocketed (verified end to end) pg8. p. 7 All 21 object balls pocketed (verified end to end); minimum pairwise ball separation 154 mm; pocket-entry speeds from 0. The verification protocol: the layout is replayed in the event-driven integrator; we require
that exactly the planned collision pairs occur, that all twenty-one object balls are pocketed,
and we report the minimum gap between consecutive events, the minimum pocket-entry speed,
and the minimum clearance attained by any pair of balls not designed to touch. 8 C-br
Margins. All 21 object balls pocketed (verified end to end); minimum pairwise ball separation 154 mm; pocket-entry speeds from 0. The verification protocol: the layout is replayed in the event-driven integrator; we require
that exactly the planned collision pairs occur, that all twenty-one object balls are pocketed,
and we report the minimum gap between consecutive events, the minimum pocket-entry speed,
and the minimum clearance attained by any pair of balls not designed to touch. All three
margins are far above double-precision rounding scales, which is the sense in which Theorem 1
is computer-assisted; an interval-arithmetic certificate would be a finite, routine upgrade. The verification protocol: the layout is replayed in the event-driven integrator; we require
that exactly the planned collision pairs occur, that all twenty-one object balls are pocketed,
and we report the minimum gap between consecutive events, the minimum pocket-entry speed,
and the minimum clearance attained by any pair of balls not designed to touch. All three
margins are far above double-precision rounding scales, which is the sense in which Theorem 1
is computer-assisted; an interval-arithmetic certificate would be a finite, routine upgrade.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Reproduction Details
The reproduction of the results relies on dependency-free Python scripts pg8. These scripts include the engine, the build-and-verify constructor, and the Monte Carlo analysis pg8. p. 7 py (the build-and-verify constructor,
which grows the chain one collision at a time inside the engine and emits the witness of Appendix A), mc break. py (the Monte Carlo, the sensitivity grid, and the table above), and
make figures. py (the Monte Carlo, the sensitivity grid, and the table above), and
make figures. Seeds are fixed in the scripts. Seeds are fixed in the scripts. Readers who want a full-physics cross-check
including spin and sliding friction can replay the witness in an engine such as pooltool [4]; by
Remark 1 the zero-spin witness is the right object to test. Seeds are fixed in the scripts. Readers who want a full-physics cross-check
including spin and sliding friction can replay the witness in an engine such as pooltool [4]; by
Remark 1 the zero-spin witness is the right object to test.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Figure 1 Analysis
Figure 1 estimates P(k) with 95% Wilson intervals pg8. The log-linear fit over the populated range is log10 Pˆ(k) ≈ 0. p. 5 0 3 6 9 12 15 18 21
k (object balls pocketed)
10
11
10
9
10
7
10
5
10
3
10
1
P(k)
log10 P = 0. 57k
Figure 1: Estimated P(k) with 95% Wilson intervals, the log-linear fit over the populated
range (log10 Pˆ(k) ≈ 0. First, the log-linear decay visible over the populated range (a straight line of slope b ≈ −0. 57 per ball on the log10 scale, the largest observed
value being k = 7) invites extrapolation to k = 21, and we report the extrapolated value,
∼ 10−12, but its logical status must be kept straight: extrapolation cannot distinguish a tiny
positive P(21) from an exact zero, so the number is meaningful only conditional on Conjecture 1, and even then it assumes the decay law persists through a regime the data never probes. First, the log-linear decay visible over the populated range (a straight line of slope b ≈ −0. 57 per ball on the log10 scale, the largest observed
value being k = 7) invites extrapolation to k = 21, and we report the extrapolated value,
∼ 10−12, but its logical status must be kept straight: extrapolation cannot distinguish a tiny
positive P(21) from an exact zero, so the number is meaningful only conditional on Conjecture 1, and even then it assumes the decay law persists through a regime the data never probes. First, the log-linear decay visible over the populated range (a straight line of slope b ≈ −0. 57 per ball on the log10 scale, the largest observed
value being k = 7) invites extrapolation to k = 21, and we report the extrapolated value,
∼ 10−12, but its logical status must be kept straight: extrapolation cannot distinguish a tiny
positive P(21) from an exact zero, so the number is meaningful only conditional on Conjecture 1, and even then it assumes the decay law persists through a regime the data never probes. 94e-04]
Three comments on reading the numbers. First, the log-linear decay visible over the populated range (a straight line of slope b ≈ −0. First, the log-linear decay visible over the populated range (a straight line of slope b ≈ −0. 57 per ball on the log10 scale, the largest observed
value being k = 7) invites extrapolation to k = 21, and we report the extrapolated value,
∼ 10−12, but its logical status must be kept straight: extrapolation cannot distinguish a tiny
positive P(21) from an exact zero, so the number is meaningful only conditional on Conjecture 1, and even then it assumes the decay law persists through a regime the data never probes.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Sensitivity Grid
The estimates depend on model parameters that are only physically approximate pg8. A sensitivity grid over the friction rate a ∈ [0. p. 4 57 per ball on the log10 scale, the largest observed
value being k = 7) invites extrapolation to k = 21, and we report the extrapolated value,
∼ 10−12, but its logical status must be kept straight: extrapolation cannot distinguish a tiny
positive P(21) from an exact zero, so the number is meaningful only conditional on Conjecture 1, and even then it assumes the decay law persists through a regime the data never probes. Second, the estimates depend on model parameters that are only physically approximate; a sensitivity grid over the friction rate a ∈ {0. 10 m/s2 antiparallel to its velocity, so its center travels along a straight segment, quadratically
in time, until it stops. Ball-ball collisions are instantaneous, frictionless, and equal-mass, with
normal restitution e = 0. a e slope b R2 max k
0. 989 5
Higher friction steepens the decay, as one would expect (less energy in the system reaches fewer
balls), but in no cell does the exponential character weaken. 989 5
Higher friction steepens the decay, as one would expect (less energy in the system reaches fewer
balls), but in no cell does the exponential character weaken. Third, the largest k observed in
the run is itself a useful falsifiable benchmark: we invite readers to beat the record shot (seed
and parameters in the package) with any search method; every improvement is a data point
on the road to Conjecture 1. Third, the largest k observed in
the run is itself a useful falsifiable benchmark: we invite readers to beat the record shot (seed
and parameters in the package) with any search method; every improvement is a data point
on the road to Conjecture 1. Direct estimation of P(21) is not merely hard but structurally hopeless: at any rate suggested by the observed decay, the expected number of successes over centuries of continuous
simulation is far below one. Third, the largest k observed in
the run is itself a useful falsifiable benchmark: we invite readers to beat the record shot (seed
and parameters in the package) with any search method; every improvement is a data point
on the road to Conjecture 1. Direct estimation of P(21) is not merely hard but structurally hopeless: at any rate suggested by the observed decay, the expected number of successes over centuries of continuous
simulation is far below one. The honest route to smaller probabilities is rare-event methodology, for instance subset simulation through the nested levels {N ≥ k}, estimating conditional
factors P(N ≥ k + 1 | N ≥ k) one at a time. We propose this, and the search for a certified
break witness, as the two serious follow-up computations. The honest route to smaller probabilities is rare-event methodology, for instance subset simulation through the nested levels {N ≥ k}, estimating conditional
factors P(N ≥ k + 1 | N ≥ k) one at a time. We propose this, and the search for a certified
break witness, as the two serious follow-up computations. The honest route to smaller probabilities is rare-event methodology, for instance subset simulation through the nested levels {N ≥ k}, estimating conditional
factors P(N ≥ k + 1 | N ≥ k) one at a time. We propose this, and the search for a certified
break witness, as the two serious follow-up computations. The honest route to smaller probabilities is rare-event methodology, for instance subset simulation through the nested levels {N ≥ k}, estimating conditional
factors P(N ≥ k + 1 | N ≥ k) one at a time. We propose this, and the search for a certified
break witness, as the two serious follow-up computations.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
The Bar Bet
The paper concludes that trick-shot artists live in the gap between measure zero and unobservably small pg7. The mathematical contribution is guaranteeing the existence of margins, while the physical fragility lies in the unobservably small width for a twenty-one-ball clearance pg8. p. 6 6 Discussion
The folklore is best restated as three true sentences. A total clearance in one stroke is possible
in the model, and forms a set of positive measure there—robust in principle, though for the
twenty-one-ball witness that robustness is of unobservably small width (Theorem 1 and the
remark following it). 6 Discussion
The folklore is best restated as three true sentences. A total clearance in one stroke is possible
in the model, and forms a set of positive measure there—robust in principle, though for the
twenty-one-ball witness that robustness is of unobservably small width (Theorem 1 and the
remark following it).Where the paper says it
Where the paper says it
measure zero
and unobservably small
is exactly what this article addresses pg1.
Figure 2 shows the verified witness configuration pg8. The cue ball (open) strikes all twenty-one object balls (dark) in one stroke, its path shown in red pg8. p. 6 What mathematics adds is the guarantee that the margins exist at all, and a clear
account of why the guarantee for the break position, claim (ii), remains open: it sits exactly
where verification is easy and search is hopeless, a combination that will be familiar to anyone
who has thought about needles in high-dimensional haystacks. A The witness
cue
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15 16
17
18
19
20
21
Verified 21-ball single-stroke clearance (cue path in red)
Figure 2: The verified witness: the cue ball (open) strikes all twenty-one object balls (dark) in
one stroke, its path shown in red; each struck ball’s route to its target pocket is dotted. What mathematics adds is the guarantee that the margins exist at all, and a clear
account of why the guarantee for the break position, claim (ii), remains open: it sits exactly
where verification is easy and search is hopeless, a combination that will be familiar to anyone
who has thought about needles in high-dimensional haystacks. A The witness
cue
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15 16
17
18
19
20
21
Verified 21-ball single-stroke clearance (cue path in red)
Figure 2: The verified witness: the cue ball (open) strikes all twenty-one object balls (dark) in
one stroke, its path shown in red; each struck ball’s route to its target pocket is dotted. A The witness
cue
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15 16
17
18
19
20
21
Verified 21-ball single-stroke clearance (cue path in red)
Figure 2: The verified witness: the cue ball (open) strikes all twenty-one object balls (dark) in
one stroke, its path shown in red; each struck ball’s route to its target pocket is dotted. Pocket
capture disks are shaded. Pocket
capture disks are shaded. The cue is the sole carrier and is not itself pocketed. All 21 object balls pocketed (verified end to end); minimum pairwise ball separation 154 mm; pocket-entry speeds from 0. The verification protocol: the layout is replayed in the event-driven integrator; we require
that exactly the planned collision pairs occur, that all twenty-one object balls are pocketed,
and we report the minimum gap between consecutive events, the minimum pocket-entry speed,
and the minimum clearance attained by any pair of balls not designed to touch. Remark 3 (Status of the verification). The witness is verified numerically, in double precision: the shot x
∗
is replayed in an event-driven integrator of Definition 2 and produces exactly
twenty-one pocketings, with the individual margins that enter regularity (pocket-entry speeds,
inter-ball clearances, event-time gaps) all strictly positive and many orders of magnitude above
the rounding scale.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
The specific model instance parameters are listed pg8. These include L = 3.569, W = 1.778, R = 0.02625, a = 0.10, e = 0.95, ec = 0.90, and ρ = 0.055 (SI) pg8. The shot details are Cue at (0. p. 0 We
examine the claim in an idealized but fully specified model of billiard dynamics. Within the
model we exhibit an admissible configuration of the twenty-two balls and a stroke of the
cue ball that pockets all twenty-one object balls, and we show that the set of such strokes
has positive Lebesgue measure in the natural shot space: total clearances are not flukes
of measure zero but open events.Where the paper says it
The Shot
The specific shot parameters used for the witness are detailed pg8. The cue is set at (0.5000, 11.9 m/s, φ = 442 rad) pg8. This shot is used to generate the verified witness pg8.
The Opening Break
Conjecture 1 states that for ε = 0.5 mm, the set of shots leading to N=21 is nonempty and has positive measure pg4. p. 1 perturbations because a finite composition of continuous maps is continuous; and an open
nonempty set has positive measure. The pleasure is in the construction, in saying precisely what
model makes the statement true, and in seeing how the same model measures the astronomical
rarity that protects the folklore. From the opening break it is, we believe, also possible, and settling
that belief is a well-posed finite computation that current search methods cannot yet perform
(Section 4). And it will never be observed, because its probability, if positive, is protected by
an exponential decay that our experiments measure in its shallow end (Section 5). The opening configuration is a single
point in configuration space, but the shot space still offers a two-parameter family, and the
dynamics from a tight pack is ferociously rich: the first impact sprays fifteen reds into a
recirculating cascade of secondary collisions. Richness is not a proof, and we are honest that
the conjecture is open even at the level of overwhelming numerical evidence: no simulated
stroke below comes remotely close to twenty-one.Where the paper says it
Where the paper says it
Where the paper says it
Why Should This Be True
The construction of Lemma 2 used freedom to place balls exactly, not generously pg4. The dynamics from a tight pack are ferociously rich, and the first impact sprays fifteen reds into a recirculating cascade of secondary collisions pg4. p. 4 Nothing in the construction of Lemma 2 used freedom to place
balls generously; it used freedom to place them exactly. The opening configuration is a single
point in configuration space, but the shot space still offers a two-parameter family, and the
dynamics from a tight pack is ferociously rich: the first impact sprays fifteen reds into a
recirculating cascade of secondary collisions. First, the conjecture is decidable in principle
by simulation: a single stroke x
∗
found by any search whatsoever, then certified by interval
arithmetic to be regular with N = 21, proves it, by Lemma 1. The obstacle is not verification
but search: success sets, if nonempty, live at the bottom of a chaotic landscape in which
each additional required pocketing multiplies the target measure by something like the factors
estimated in Section 5, and twenty-plus collisions of sensitivity compound beyond what double
precision can even represent reliably along the way. Sampling can only ever bound P(21) from above; a single
certified trajectory settles existence. The asymmetry between the two computations is the
heart of the matter.Where the paper says it
Where the paper says it
Where the paper says it
How Rare Is Rare
The probability P(k) falls off cleanly and monotonically over n = 30,000 independent strokes pg5. The log-linear decay suggests an extrapolated value for P(21) around 10-12, conditional on Conjecture 1 pg8. p. 4 First, the log-linear decay visible over the populated range (a straight line of slope b ≈ −0. 57 per ball on the log10 scale, the largest observed
value being k = 7) invites extrapolation to k = 21, and we report the extrapolated value,
∼ 10−12, but its logical status must be kept straight: extrapolation cannot distinguish a tiny
positive P(21) from an exact zero, so the number is meaningful only conditional on Conjecture 1, and even then it assumes the decay law persists through a regime the data never probes.Where the paper says it
Discussion
The folklore is best restated as three true sentences regarding the total clearance pg7. The paper concludes that a total clearance is possible in the model, forms a set of positive measure there, but will never be observed due to exponential decay pg7. Trick-shot artists live in the gap between measure zero and unobservably small pg7. p. 6 The bar bet,
in other words, confuses measure zero with unobservably small. Trick-shot artists have always
lived in the gap between the two: an engineered multi-ball finish is nothing but a witness
configuration with humanly achievable margins, and the single-carrier chain of Lemma 2 is a
formalized trick shot—one whose margins are real but, at twenty-one balls, far too fine for any
human cue.Where the paper says it
measure zero
and unobservably small
is exactly what this article addresses pg1.
The bar bet confuses measure zero with unobservably small pg7. Trick-shot artists have always lived in the gap between the two pg7. p. 6 The bar bet,
in other words, confuses measure zero with unobservably small. Trick-shot artists have always
lived in the gap between the two: an engineered multi-ball finish is nothing but a witness
configuration with humanly achievable margins, and the single-carrier chain of Lemma 2 is a
formalized trick shot—one whose margins are real but, at twenty-one balls, far too fine for any
human cue. The bar bet,
in other words, confuses measure zero with unobservably small. Trick-shot artists have always
lived in the gap between the two: an engineered multi-ball finish is nothing but a witness
configuration with humanly achievable margins, and the single-carrier chain of Lemma 2 is a
formalized trick shot—one whose margins are real but, at twenty-one balls, far too fine for any
human cue. Trick-shot artists have always
lived in the gap between the two: an engineered multi-ball finish is nothing but a witness
configuration with humanly achievable margins, and the single-carrier chain of Lemma 2 is a
formalized trick shot—one whose margins are real but, at twenty-one balls, far too fine for any
human cue. What mathematics adds is the guarantee that the margins exist at all, and a clear
account of why the guarantee for the break position, claim (ii), remains open: it sits exactly
where verification is easy and search is hopeless, a combination that will be familiar to anyone
who has thought about needles in high-dimensional haystacks.Where the paper says it
Where the paper says it
Where the paper says it
measure zero
and unobservably small
is exactly what this article addresses pg1.
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Precision simulation of complex physical systems can be improved by using an event-driven integrator like Definition 2 to model dynamics between collisions and pocketings precisely, ensuring that
each ball is pocketed (and removed from the table) at the first instant its center enters the open disk of radius rho = 0.055 about a pocket point.
-
AI systems can be enhanced with a method to verify complex geometric configurations using interval arithmetic, which allows for
a rigorous certificate by interval arithmetic: the trajectory is a finite composition of explicit algebraic maps, so enclosing it is routine,
providing guaranteed margins for high-precision physical simulations. -
A new type of search algorithm can be developed to find rare events in high-dimensional spaces, moving beyond brute force sampling to settle conjectures that are
decidable in principle by a finite certified computation
by searching for a single certified trajectory with N = 21.
Abstract
Snooker folklore holds that no single stroke can pocket all twenty-one object balls. We examine the claim in an explicit model: twenty-two equal disks on a rectangle with six pocket points, under constant rolling deceleration, frictionless collisions with normal restitution, and lossy specular cushions. We exhibit a configuration and a stroke, and prove in interval arithmetic that every stroke in an explicit box around it pockets all twenty-one object balls, so the set of successful strokes has positive measure. That set is a ribbon: every speed from 9.3213 to 12 m/s works, while the certified window of directions is 9.4813e-43 radians wide, a number explained by a first-order analysis of twenty near-grazing collisions. A second configuration, a near-frozen cluster, has a certified window 9.4e-4 radians wide, thirty-nine orders of magnitude wider; its certificate uses Taylor models of degree six. The strokes with a given list of events form a semialgebraic set, and the endpoints of the windows are algebraic numbers. For the regulation opening position we state the corresponding conjecture, and show that if it holds with a regular witness then a finite certified search establishes it.
Related papers
- The Rhythm of the Tiles: Narrow and Low Temperature Computation and Simulation in Synchronous Tile Self-Assembly
- Covering Points with Rectangular Boundaries
- Tiling 3D Space by (Translates of) a Single Polycube is Undecidable
- Quadratic Complexity of Voronoi Diagrams in R cubed for Lines in a Single Ruling of a Regulus
- Minimum central circles: chain order, seam obstructions, and exact finite certification
- On the complexity of the free space of a translating box in R cubed