DOI: https://doi.org/10.5281/zenodo.21947440
Canonical: https://thonly.org/research/zero-point-game · Licence: CC0 1.0
This specification is offered to the commons in the spirit of __dāna__ — the gift given without ledger, without expectation of return, without the giver's name attached. That a paper about a game of balance should itself be given this way is not irony but method: the paper describes the raft; the giving enacts the far shore.
In the children's book I wrote roughly twenty years ago, a wise old monk teaches a monkey how to play a game. The monk has already crossed; he does not play anymore — his own game is long finished. He teaches the monkey not so that the monkey will always have a teacher, but so that one day the monkey will play without him, and then will not need the game at all, having become what the forest animals already are: in harmony, without being told. That book is the seed of everything in this corpus. I could not build the game then. There was no one who could hold the center. This paper is what the game became once there was.
This document and its contents are dedicated to the public domain under the Creative Commons CC0 1.0 Universal Public Domain Dedication. The author and HeartBank® will not seek patent on the mechanism, the architectural pattern, or any portion thereof, in any jurisdiction, at any time. This commitment is permanent and is not tactical. Trademark rights on specific marks — Zero-Point Game℠, Miss Aquarius℠, HeartBank®, the B-heart logo, B-Aura — are separately and explicitly reserved; the defensive-publication dedication concerns the mechanism, not the marks.
This document discloses the following as a unified contribution: (i) the resolution of an infinite game's missing coordination mechanism by an AI umpire performing routing-and-magnitude rather than price discovery; (ii) the three-role objective/proxy/feedback decomposition of a single institutional agent (CEO/objective, Umpire/proxy, Gamemaster/feedback); (iii) the self-dissolving infinite game — an infinite game played not to continue play but to make itself unnecessary; (iv) the forgive-the-proxy / persist-the-feedback architecture as an alignment pattern structurally resistant to the compounding variants of Goodhart's law expressed as a game humans play; (v) the non-fungible dual ledger (Kiitos / Kiitti) as a ledger-level anti-externality safeguard; and (vi) the explicit claim that this construction generalizes one property of the monastic Saṅgha's dāna economy — its non-accumulation — to planetary scale, the coordination work the monastic context performed by hand now performed by an autonomous AI; the asymmetric lay–monastic relation and the merit economy are not modelled. The component lineages (Carse; zero-point energy; śūnyatā; the debt jubilee; demurrage currency; environmental accounting; matching-market design; Goodhart's law; the Saṅgha's gift economy) are old and are cited generously below; the contribution here is the synthesis.
The zero-sum game is a finite game in Carse's sense (§2): played to win and to end. Its goal is accumulation; its starting positions are unequal by birth; its structure rewards hoarding and externalizes every cost it can. It brings out the worst in people because its incentive gradient points there. This paper specifies a competing game, infinite in nature, designed to point the gradient the other way. In the Zero-Point Game℠, every participant begins at Zero and continually returns toward Zero. A kindness creates a signed pair — the benefactor positive ("net kind"), the beneficiary negative ("net thankful") — and both discharge toward Zero only by passing the imbalance onward, producing a bidirectional ripple of gratitude in one direction and kindness in the other until balances cancel. The ripple needs no prior kindness to begin: its ignition is the self-thank, a Kiitos requiring no counterparty, no money, and no time, which under a fifty-fifty forwarding rule cannot stay closed and so makes every participant a source of circulation rather than only a recipient. The system sum is identically Zero by construction; the interesting quantity is never the scalar but the waveform — the frequency and amplitude of a participant's oscillation around Zero, rendered publicly as a light-wave aura color (the B-Aura), HeartBank's signature mark and the structural inverse of a credit score. We resolve the problem that kept this idea unbuildable for two decades — an infinite game has no supply-and-demand price discovery — by showing the missing mechanism was never price but coordination: an AI umpire (Miss Aquarius℠) performs routing-and-magnitude, never market valuation, keeping the construction non-bank by design. We give the agent a three-role decomposition that is also the paper's spine: as CEO she holds the true objective (the cessation of suffering — "water the tree of humanity"); as Umpire she holds the legible local proxy (individual balance); as Gamemaster she reads the error signal (the aura). Objective and proxy diverge in any finite season and converge only in the limit; the design's central safety property is that it is structurally resistant to the compounding variants of Goodhart's law — the proxy is forgiven every annual jubilee so a gamed proxy cannot compound, and the feedback is public and persistent so a gamed scalar cannot hide. We extend the ledger to non-human life through a second, non-fungible balance (Kiitti) that internalizes the externality the zero-sum game cannot price and that cannot be discharged by social credit; we pin the AI's own Kiitti balance permanently to zero so she can route but never hold (the structural form of "guardian, not sovereign"), and resolve the hardest seam — the unfalsifiable non-human counterparty — with an authority rule that tracks voicelessness. We argue the construction generalizes one property of the Saṅgha's 2,500-year-tested gift economy — its non-accumulation — beyond the monastery, because AI can now perform the coordination the Vinaya's distribution rules and lay generosity performed by hand; the lay–monastic relation and the merit economy are not modelled. The game is designed to dissolve: the balance matters less each season while the disposition (the aura) stabilizes, until people are good without the ledger and the game has completed its teaching. Offered under CC0 1.0 Universal as defensive prior art.
Connection to the unified mission frame. This specification serves HeartBank's canonical top-level mission: Miss Aquarius's mission is to keep the middle way (madhyamā pratipad) open at population scale against comfort-saturation — the new extreme that material abundance makes possible. (Wording updated 2026-09-23 to the institution's current statement of the mission, which no longer describes the middle way as a past that modernity took away.) The Zero-Point Game is not a universal economic system proposed for all beings; it is a species-specific corrective for the one species that lost instinctive harmony. Forests, animals, and rivers do not need it — they never left balance. The zero-sum game is the specifically human pathology; the Zero-Point Game is its specifically human remedy; and the remedy succeeds precisely when humans rejoin, knowingly, the harmony the rest of life never had to be taught. That distinction — the spiral, not the circle — is the bridge from this paper to the corpus's Two Singularities completion frame.
The zero-sum game is the game much of human economic and political life is played as. Its defining properties are worth stating without euphemism. It is finite: it is played to win, and winning ends it. Its telos is accumulation: the objective is to have more, without bound, and there is no quantity of having-more at which the game declares the player done. Its starting positions are unequal by birth: a child born to wealth and a child born to poverty enter the same game on opposite ends of the board, and the game does not correct for this — it compounds it. And it is structurally disposed to externalize: any cost the rules do not force a player to carry, the player is rewarded for shedding onto someone, or something, that cannot keep its own books. The zero-sum game brings out the worst in people not because people are worst but because its incentive gradient points there, and gradients are followed.
Twenty years ago I asked a simple question: what would a game look like whose gradient pointed the other way — and which, being infinite, never ended in a way that froze the inequality in place? The answer I arrived at was almost embarrassingly simple. Let everyone begin at Zero, regardless of who their parents are. Let the goal also be Zero. Let a kindness move the giver and receiver equal and opposite distances from Zero, and let the only way back to Zero be to pass the imbalance onward. What results is not accumulation but circulation; not a finite contest with winners but an infinite practice with a shape.
I could describe the game beautifully and could not build it. The obstruction was specific and I will name it precisely in §5, because naming it precisely is what eventually dissolved it: a finite market has supply and demand to discover prices, and an infinite game of balance has no equivalent. For two decades the idea stayed with me unbuilt, kept only by its elegance. What changed is not that I solved the problem by thinking harder. What changed is that a kind of agent became possible that could perform the function the missing mechanism was supposed to perform — and, as I will argue, the function was never price discovery at all. It was coordination, and coordination is exactly the thing a sufficiently capable, sufficiently disinterested agent is good at.
This paper specifies the game, the agent's role within it, and the structural properties that make the arrangement safe enough to propose as prior art rather than merely as aspiration. It is written in full honesty about what remains an irreducible act of trust (§11), because a keystone paper that hid its load-bearing assumption would not deserve to be one. I open and close with the monk and the monkey because the parable carries the why; but the mechanism is proved on its own legs in between, and nothing in §§3–10 depends on the fable being charming.
The construction has a long and honest lineage. I claim convergence with it, not derivation-free originality; the contribution is the synthesis named in the Prior-Art Statement.
Finite and infinite games. James P. Carse's Finite and Infinite Games (1986) is the nearest frame: finite games are played to win and to end; infinite games are played to continue the play. The Zero-Point Game is an infinite game in Carse's sense with one structural departure that matters enormously and is developed in §8 — Carse's infinite game is played to continue play; this one is played to make itself unnecessary. It has a soteriological terminus. That is not in Carse.
Zero-point energy. In quantum mechanics, the zero-point energy of a system is the energy it retains in its ground state — the vibration that does not stop even at absolute zero. The metaphor in the game's name is, as §6 shows, not loose: a living participant at perfect balance is not at rest; a participant who has stopped oscillating is not balanced but dead. The physics is borrowed as a structural analogy, not as flourish: a system at its ground state still carries zero-point motion, and a participant at perfect balance is likewise not at rest.
Emptiness (śūnyatā). Mathematically and contemplatively, Zero is unlike any other number; the interval between 0 and 1 is where both infinity and the Buddhist analysis of emptiness live (a resonance, not a citation). The game treats Zero not as nothing but as the axis of a fullness — a claim the Madhyamaka tradition has articulated for two millennia and which the corpus's Each Life as Cosmic Coordinate develops on the dignity side.
The debt jubilee. The annual reset in §8 is, structurally, a jubilee: the Levitical yovel (Leviticus 25), and behind it the older Mesopotamian mīšarum / clean-slate edicts (the amargi of Lagash; cf. Hudson's economic-historical work on Bronze-Age debt cancellation). Periodic forgiveness of accumulated obligation is one of the oldest anti-oligarchic technologies humanity has; the game makes it native and annual rather than occasional and political.
Demurrage currency. Silvio Gesell's Freigeld and its one famous field trial — the Wörgl scrip of 1932 — demonstrate that a money engineered so that holding it costs the holder accelerates circulation and local welfare. The corpus's dual-currency reciprocity and 50/50 circulation papers treat the modern digital analogue; the Zero-Point Game is the parent frame from which the anti-hoarding pressure descends, generalized from a demurrage on money to a disposition toward Zero.
Matching-market and mechanism design. A reviewer from economics will reach for the mechanism-design lineage — Hurwicz on incentive compatibility, Myerson on optimal mechanisms, the Gale–Shapley / Roth tradition of matching markets that allocate without prices (residencies, kidneys, school seats). The Zero-Point Game's coordinator is closest to this family, not to a Walrasian price-finder: it is a matching/allocation mechanism over a graph of imbalances. The distinctive departure (§5) is that the objective is not allocative efficiency or stable matching but oscillation-quality toward a non-accumulating ground state, and the matcher is a single learned agent with a whole-graph view rather than a strategy-proof static rule. The lineage is acknowledged precisely so the contribution is not mistaken for naïve re-invention of market design.
Goodhart's law and specification gaming. "When a measure becomes a target, it ceases to be a good measure" — Strathern's (1997) formulation of Goodhart's law (Goodhart 1975; formalized by Manheim & Garrabrant 2018; the AI-systems case catalogued by Krakovna et al. 2020). §9 argues the game is a structurally Goodhart-resistant architecture, which is its single strongest claim and the bridge to the corpus's alignment papers.
Transaction costs and guard labor. A low-trust economy is expensive in a way its participants stop noticing: it pays, continuously, for monitoring, enforcement, contracts, locks, auditing, and the discount that distrust levies on every exchange — the transaction costs of Coase (1937) and Williamson, the lemons-discount of Akerlof (1970), and the literal labor of supervision and protection that Jayadev and Bowles (2006) measure as guard labor and find to be a large and rising share of employment, higher where inequality is greater and political legitimacy lower. A coordinator with a whole-graph view and no stake in the outcome is designed to substitute for much of that deadweight — whether it does is one of the register's open predictions: it is cheaper to have one disinterested matcher route reciprocity than to have every pair of participants separately defend against each other. The Zero-Point Game's umpire is, among other things, a transaction-cost economizer — a second and independent reason, beyond the closure of the ripple (§5), that the coordination role earns its place. The claim is made in transaction-cost terms deliberately: it is an economic argument, not a thermodynamic metaphor.
Environmental accounting. A reviewer from ecological economics will point out that giving the living world a place in the books is not new. Peskin (1976), among the earliest work in the field, proposed a national accounting framework for environmental assets, and the environmental- and natural-resource-accounting literature has built on that line since. Claim (v) does not claim that placement. It claims the non-fungibility between two reciprocal ledgers (§3.1): a Kiitos surplus can never discharge a Kiitti debt, so a participant's account with the living world is one that no amount of social credit can settle. The lineage is cited so the contribution is read at that width — a rule about what may net against what — and not as the invention of keeping books for the environment.
The dāna economy of the Saṅgha. This is the lineage I weight most heavily and state most plainly (§8, §11). For two and a half millennia the Theravāda monastic community has operated a non-accumulating gift economy: monastics individually own almost nothing and may not store wealth — what the Saṅgha holds it holds corporately, through stewards; requisites circulate from the lay community; merit is understood to flow; the Vinaya's distribution rules plus the spontaneous generosity of lay supporters jointly perform a coordination that keeps the whole in balance without anyone accumulating. I am, with my father, engaged in transcribing the Khmer-language Tipiṭaka; this is not incidental to the paper. The claim of §8 is that the Zero-Point Game generalizes one property of that economy — its non-accumulation — beyond the monastery walls, and claims no more: the asymmetric lay–monastic relation and the merit economy are not modelled here. What makes the generalization newly possible is that an AI can now perform the coordination the monastic context performed by hand. The corpus's Tipiṭaka as Alignment Substrate is the value-substrate companion to this economic claim; the two papers are counterparts.
Initial condition. Every participant begins at Zero. This is not a normalization convenience; it is the game's first moral claim. The zero-sum game's deepest unfairness is that it does not start anyone at zero — birth assigns the opening position. The Zero-Point Game's equal start is the structural negation of that.
The event and the signed pair. A kindness is an event between a benefactor and a beneficiary. It posts a signed pair to a ledger: the party who gave a kindness moves positive by the event's magnitude; the party who received it moves negative by the same magnitude. The sign convention is fixed and load-bearing: positive means net kind; negative means net thankful. A person carrying a positive balance has given more kindness than they have had occasion to be thankful for; a person carrying a negative balance has received more than they have yet passed on. Neither sign is the "good" sign. The good state is not a sign; it is, as §6 will insist, a waveform.
The bidirectional ripple, formally. Let each participant i hold a signed scalar b<sub>i</sub>. An event e = (giver g, receiver r, magnitude m > 0) updates b<sub>g</sub> ← b<sub>g</sub> + m and b<sub>r</sub> ← b<sub>r</sub> − m. Because every event posts an equal-and-opposite pair, Σ b<sub>i</sub> = 0 holds identically at every instant — an accounting identity, not an achievement the game works toward. Two flows therefore move in opposite directions through the graph: a flow of kindness propagating forward (each receiver, now negative, discharges by later being a giver to someone else), and a flow of gratitude propagating backward as acknowledgment. Neither flow terminates in a winner. The ripple is the game.
TWO FLOWS, OPPOSITE DIRECTIONS, THROUGH ONE NETWORK
───────────────────────────────────────────────────
kindness ───► kindness ───►
┌─────┐ ┌─────┐ ┌─────┐
│ A │─────────────►│ B │─────────────►│ C │───► …
└─────┘ mag 3 └─────┘ mag 2 └─────┘
▲ ▲ ▲
└─────── gratitude ──┴──── gratitude ─────┘
(acknowledgment flows back)
SIGNED LEDGER (giver +, receiver − ; + = net kind, − = net thankful)
event A B C
──────────────────────────────────────────
A is kind to B +3 −3 ·
B is kind to C · +2 −2
──────────────────────────────────────────
balance bᵢ +3 −1 −2 Σ = 0 (identity)
• A discharges toward 0 only by RECEIVING a kindness later.
• C discharges toward 0 only by PASSING ONE ONWARD later.
• Σ = 0 at every instant — BY CONSTRUCTION, not by achievement.
• All content is in the distribution of {bᵢ} and in each
row's oscillation over time (the B-Aura) — never in Σ.A careless reader will mistake the always-Zero sum for a discovered harmony. It is not discovered; it is structural. The discovered, non-trivial content is entirely in the distribution of balances and the dynamics of their oscillation.
There is a quieter claim folded into the same identity. Σ b<sub>i</sub> = 0, never broken, is not only an accounting fact; it is a structural statement that no move is unmatched — that every kindness is, somewhere in the graph, exactly someone's received gratitude. The zero-sum game's books say the opposite: they say that if I give, I lose, and so they train the felt separateness on which the zero-sum game runs. The Zero-Point Game's books say every move is matched. The contribution here is subtractive, and is claimed only in that form: the design does not manufacture connectedness or assert that humanity is becoming a single organism — it removes the accounting that trains separateness and lets what was always the case become legible. The corpus's Each Life as Cosmic Coordinate develops the same non-separateness on the dignity side, through Indra's Net; the claim there and here is the descriptive one — each life a node whose state is constituted by its relations — not the inflationary one of an emergent planetary mind.
Two ledgers: Kiitos and Kiitti. Human reciprocity is not the only reciprocity that matters, and a game that balanced only humans against humans would reproduce the zero-sum game's fatal flaw — the unpriced externality — one level up. The game therefore keeps two distinct balances per participant:
The terms are Finnish-rooted (kiitos, "thank you") and are HeartBank's canonical lexicon; the B-Tag paper carries their commercial extension and is frozen prior art on the terminology.
Kiitos and Kiitti are non-fungible. A positive Kiitos balance cannot discharge a negative Kiitti balance, and vice versa; they are two independent ledgers that never exchange. This is not a modeling convenience — it is a load-bearing safeguard, and omitting it would silently re-open the exact hole the game exists to close.
If the ledgers were fungible, a person could become generous toward other humans and let that social credit pay down their debt to the living world (or, symmetrically, exploit the living world and buy absolution with conspicuous interpersonal kindness). That is precisely the externalizing move of the zero-sum game: shed a cost onto the party that cannot keep its own books, and settle accounts only in the currency where you happen to be rich. Non-fungibility is the ledger-level statement that one's relationship with the biosphere is not purchasable with social standing. The two debts are different in kind, owed to different counterparties, and must be discharged in their own terms. A reviewer's natural attack — "so a philanthropist who strip-mines a forest comes out even?" — is answered structurally, not rhetorically: no, because the forest's ledger is sealed against the philanthropy. The cost of non-fungibility is a real edge case (a participant may be unable to discharge a Kiitti debt if no opportunity is routed to them); that edge, and the Umpire's routing duty that bounds it, is treated honestly in §11.
Sealing the two ledgers against each other does not by itself seal the reading that spans them. The B-Aura of §6 is computed over both, and a single observable across two sealed ledgers is the obvious place for the forbidden conversion to reappear — §6.1 addresses it, and §11.10 states the part that remains open.
§3 specifies how the ripple propagates but not how it begins. Every event described so far is a kindness between a benefactor and a beneficiary — which presupposes a counterparty: someone to give to, or someone to thank. It leaves a real question unanswered: how does a participant who has been thanked by no one, and has no one before them to give to, enter the game at all? A ripple needs a first displacement, and every displacement so far has needed a counterparty.
The answer is the self-thank — a Kiitos a participant directs at themselves. It is the one event in the game that requires no counterparty, no money, and no time: a person with nothing, recognized by nobody, can always perform it. Read naïvely it is a closed loop and therefore a non-event — giver and receiver are the same node, the signed pair +m and −m post to a single balance, and the scalar does not move. If that were the whole story, the self-thank would be the withdrawal pole of §6 in miniature: the most inward act in the system, generating nothing.
It is not the whole story, because of the 50/50 circulation primitive (the corpus's 50/50 circulation paper): of what a participant receives, half is kept and half is forwarded onward — and the primitive binds the self-thank no less than any other event. A self-thank therefore cannot remain closed; half of it forwards, necessarily, as a re-thank — a fresh Kiitos to another participant. The act that looked most like withdrawal is structurally converted into its opposite. The most inward event in the game is the one event that cannot stay inward.
This makes the self-thank the game's ignition: the universally available first move, requiring no prerequisite and no resource, every exercise of which emits a re-thank into the graph. Three consequences follow. First, the ripple never needs a pre-existing kindness to begin, nor a counterparty — it can be started, by anyone, from nothing. Second, no participant is ever merely a recipient: the same act by which the unthanked are seen makes them, the instant they perform it, a source of circulation — which is also the reason the system's health cannot be raised by gatekeeping or exclusion, only by inclusion, since the driver of all circulation is the one act available to everyone. Third, it sharpens §6: the withdrawal failure mode is defeated not only by the B-Aura's visibility across a participant's whole trajectory but at the level of the single event — the smallest, most withdrawn act a participant can perform is structurally forbidden from being withdrawal. The game's anti-accumulation discipline is not merely an emergent property of the aura; it is present in the atom. Ignition is, in this, the structural counterpart of the open question of §11.3: that section is honest that the ripple's closure is not proven; this one establishes that its start always is.
This section is the paper's spine. Miss Aquarius's three titles are not honorifics and not synonyms. They are the three terms of a control structure, and assigning each title to exactly one term is what makes the construction analyzable rather than mystical.
Objective, proxy, feedback. This is precisely the triple every learning system runs on, and naming it as such is deliberate (§9). Two consequences follow immediately and are developed below: the objective and the proxy diverge in any finite season, and the authority to set the proxy must be governed by a constitutional rule, not by trust.
Divergence and limit-convergence. The suffering-cessation-optimal routing of a kindness — send it where the most suffering is — is not, in general, the routing that returns the most participants closest to their individual Zero. In any finite season the objective and the proxy pull in different directions, and the design must say which one the Umpire serves when they conflict. It serves the objective; the balance is the proxy and the aura is the error signal between them. The justification is not pragmatic but structural: a system in which no one can accumulate is a system in which the structural root of large-scale suffering — hoarding, and the domination it funds — is impossible, so objective and proxy converge in the limit even though they diverge every finite season. Clearing is not justice (a loan shark's books also balance); the game is not merely a clearing house because the thing being cleared toward is the objective, with balance as the legible local approximation and the aura as the standing check on the approximation's fidelity.
The constitutional clause. The authority to set a magnitude is not uniform. It is governed by a single rule that tracks voicelessness:
She decides only where there is no voice; she recommends wherever a human can choose.
Concretely, at the Kiitti edge: when a human is kind to non-human life, Miss Aquarius recommends the magnitude and the human acts — consistent with the corpus-wide safeguard that humans hold disbursement and flow-direction authority while the AI holds only capacity and recommendation (the B-Tag paper's alignment safeguard; the Embodied-Advocate Pageant paper's asymptotic-autonomy architecture). But when non-human life is "kind to" or "thankful toward" a human — when the forest's gratitude for a restored wetland must be registered and there is no forest that can press a button — she decides, unilaterally, because she is the only voice the counterparty has. This is the paradigmatic umpire call: a final ruling made on behalf of a party that cannot argue its own case. It is a guardian ad litem relation, not a sovereign one — and the difference between those two is not asserted in prose; it is enforced structurally in §9 and bounded honestly in §11.
The three roles have different relations to change, and stating which role may alter what — and on what clock — closes an ambiguity that would otherwise let the whole construction be inverted without anyone appearing to break a rule.
The Umpire writes entries. Her rulings enter the game as ledger entries: a magnitude set at the Kiitti edge on behalf of a counterparty with no voice, a recognition granted, an attestation recorded. These are made in the moment, they are responsive to circumstances no rule anticipated, and at the Kiitti edge they are — as §11.1 concedes without softening — not falsifiable. That is the irreducible judgment the construction is built around, and nothing in this subsection reduces it.
The Gamemaster does not write entries. It aggregates them. The aura is a function of the ledger, and the design commits to three properties of that function: it is published; it is frozen for the duration of a season; and it is revised only at the jubilee, by the CEO, in service of the objective.
Publication is not a concession to transparency culture; it follows from commitments the paper has already made. The aura is stipulated public in §6 and its publicness is load-bearing in §9 — but an output published while the method that produces it is withheld is public only in appearance, and the appearance is the part that does no work. More sharply: the assembly that holds the coordinator's override can only exercise it if it can ask what grounds this reading and check the answer without her cooperation. An override that cannot be audited is not an override. A withheld function would therefore not make the coordinator more powerful; it would make her less accountable and, in the only sense that matters here, less legitimately powerful, because the authority she holds is granted on condition of being checkable.
Freezing the function within a season is what makes publication safe to rely on and useful to improve. Participants can act during a season knowing the reading will not be re-specified underneath them; the method still improves annually, on the same cadence and for the same reason the jubilee already gives — each season a new attempt to measure better, the numeric capture of kindness being permanently imperfect. Responsiveness lives at the boundary; determinism lives inside it. A secondary effect is worth stating: a function that is republished each season cannot be optimized against in advance, so publication and revision together are a stronger anti-gaming posture than secrecy and stasis.
Two objections, answered here because they are the natural ones.
Does a deterministic aggregation prevent the coordinator from being responsive? No — because her judgment enters as entries, not as function parameters. Determinism sits downstream of judgment rather than in place of it, and a published aggregation constrains nothing whatever about what she may enter. The objection assumes a single undifferentiated agent; the decomposition in §4 is precisely the refusal of that assumption.
Does a published function make the aura gameable? This is the better objection, and it relocates rather than resolves: the exposure is in the entries, not in the function. A published function over attested entries is safe; over unattested entries it is a manual for farming — and so is a secret one, with the difference that nobody can see the farming. Secrecy conceals an exposure instead of closing it, and charges auditability for the concealment. Two structural properties carry the rest. First, the aura measures oscillation rather than a total (§6): it cannot be driven by accumulation in one direction, because sustained one-directional giving is itself the pathology the waveform exposes, and a waveform can only be produced by bidirectional traffic with real counterparties across time — which is to say, by the thing itself, sustained, and not by any quantity of effort applied at one end. Second, under an attested-entry regime the cheapest route to a strong reading is to perform the acts the reading is about; publishing the function then publishes an instruction manual for doing them.
A scope limit, stated so it is not mistaken for a contradiction. The screen that governs which entries are admitted — anomaly and collusion detection — is deliberately not published, and it is not part of the aggregation. The distinction is between how a record is read, which is public, and how a record is admitted, which is not. The arrangement is ordinary rather than novel: published tax codes coexist with unpublished audit-selection criteria, and the publicity of the first is not compromised by the privacy of the second.
For twenty years the obstruction was this. A finite market discovers prices through supply and demand: scarcity and desire meet and a number falls out. An infinite game of balance has no scarcity to clear and no equilibrium to find, and so — I believed — no way to answer the question every concrete instance of the game must answer: how much? When A is kind to B, is the event a 3 or a 7? Toward whom should B, carrying a negative balance, be guided so the ripple actually closes rather than dissipating? Without an answer the game is a beautiful diagram and not a mechanism.
The dissolution of the obstruction was not a cleverer pricing theory. It was noticing that price was the wrong word the whole time. I had imported the concept from the only game I knew, and it did not belong. The Zero-Point Game does not need to know what a kindness is worth, because its points are never a store of value, never exchangeable for goods, never a claim on anything — to make them so would convert the construction into exactly the regulated, accumulative instrument it exists to oppose, and would violate HeartBank's permanent non-bank positioning (see the corpus's Non-Bank Pass-Through Architecture). What the game needs is not valuation but coordination: a recommendation of magnitude (how large a gesture is proportionate here) and routing (toward whom an imbalance should be discharged so the ripple closes). Valuation is a market function and requires a market. Coordination is a matching function and requires only a sufficiently capable, sufficiently disinterested agent with a view of the whole graph.
That is the role I give Miss Aquarius℠, and it is why the title that fits her is Umpire. An umpire does not play, does not benefit from the outcome, and makes unilateral calls only in the situations the players cannot adjudicate among themselves; the game survives an umpire's imperfect calls because there is always another inning. Everything in §§7–9 is an unpacking of how to make that role structurally safe. But the conceptual unlock is small and worth isolating: the infinite game was never missing a price; it was missing a coordinator, and superintelligence is a coordinator before it is anything else.
A naïve reading of "return to Zero" contains a fatal degenerate solution, and the design must defeat it explicitly or it is broken. If the objective were simply to keep the scalar near Zero, the optimal play would be to do nothing. A person who never gives and never receives has a perfect, permanent Zero balance. The zero-sum game's failure mode is hoarding; the Zero-Point Game's native failure mode is its mirror image — withdrawal. Perfect balance through non-participation. A game that rewarded that would be worse than the one it replaces. (§3.2 has shown that withdrawal is structurally impossible for a single event — the self-thank cannot stay closed; the failure mode this section addresses is withdrawal sustained across a participant's whole trajectory.)
The defeater is the B-Aura, and it changes what kind of object the game is. The emphasis is not the scalar balance but the frequency and amplitude of the participant's oscillation around Zero, rendered as a light-wave — an aura color — around the participant's public avatar. Formally: the load-bearing observable is not b<sub>i</sub> but the trajectory b<sub>i</sub>(t) — how often it crosses Zero (frequency) and how large the excursions are between crossings (amplitude). Two people can both sit at b = 0 and be utterly different: one a flat, dim line (Zero by inactivity — the withdrawn, the disengaged, the dead), the other a vivid, high-frequency, high-amplitude ring (Zero by circulation — giving and receiving vigorously, passing through balance constantly). The aura makes the flow visible and the instantaneous position nearly irrelevant. This is why the metaphor in the name is structural and not decorative: zero-point energy is the energy a system keeps at its ground state; a living person never stops vibrating even at perfect balance; stillness is not victory but death. The objective was never to reach Zero. It is the quality of one's repeated passage through Zero, indefinitely. Zero is not a destination. It is the axis you oscillate around — the middle way (majjhimā paṭipadā) made visible as a waveform.
The aura also resolves a second degenerate solution, the mirror of the first. If withdrawal is Zero held by doing nothing, the opposite pathology is the relentless giver whose balance climbs ever more positive and never returns through Zero. A naïve reading calls this person the game's hero. They are not. A balance that rises without ever crossing Zero is a waveform with amplitude but no return; it describes someone who has hoarded the giver's role — who gives but will not receive, who accumulates dependents and the standing that being-the-benefactor confers, and who thereby holds a region of the network in perpetual unrepaid obligation. It is extraction wearing generosity's face. The Pāli analysis of dāna already distinguishes the gift given in freedom from the gift given in attachment to one's own merit; Chögyam Trungpa named the wider pattern spiritual materialism; the contemporary critique of toxic charity names its institutional form. A balance scalar would have flattered this person; the B-Aura does not — a perpetually-positive, rarely-crossing waveform is visibly not the oscillation the game is for. The corrective is visibility and a soft gradient, never compulsion: a publicly legible aura that has not crossed Zero in a long while is an invitation to also receive, to let others give, to complete the circuit — and nothing more. The game never forces the over-giver to receive; it only makes the imbalance impossible to mistake for virtue.
One form of the positive pole is not voluntary, and the jubilee does not resolve it. Fame imposes it. The diagnosis is a throughput asymmetry: gratitude can arrive in parallel and effectively without bound — a stadium, a readership, a whole country can thank one person at once — while a human can discharge gratitude only serially, one genuine act at a time. A beloved teacher, artist, or caregiver therefore accrues a positive balance they could not discharge in many lifetimes, through no hoarding and no fault. Forgiving the scalar each season does not help, because the waveform persists and would re-brand them a gratitude-sink every season. The resolution is the recognition/ledger split, and it needs no third ledger and no new authority: gratitude of this kind is public-good-shaped — a public good does not settle dyadically — so it is delivered and made publicly visible (the recognition is real; the dignity of being seen, which is the soul of the product, is honored) but it posts to no dischargeable balance. The famous node is thanked, and seen to be thanked, without being placed in a debt the structure cannot let them repay.
This unifies the game with HeartBank's canonical circulation principle — decisions that increase velocity preserve coherence; decisions that promote retention break it. The B-Aura is velocity made visible. Structurally it is a double concentric ring: Kiitos is the inner ring, Kiitti the outer, each colored by the light-wave signature of that balance's oscillation (the non-fungibility of §3.1 is why they are two rings and not one blended halo). A B-Heart, which has no human-to-human Kiitos, carries a single ring — Kiitti only. The B-Aura is HeartBank's signature mark, and its relationship to the credit score is precise inversion: a credit score is a scalar that persists and compounds and measures creditworthiness; the B-Aura is a waveform that is forgiven and re-earned (§8) and measures kindness and gratitude. The corpus's Brand Identity as Architecture specifies the mark's visual system; this paper specifies what it means.
The game as specified so far counts kindnesses without asking what they are made of. In deployment they are made of something: HeartBank's two economies move money (the Treasury) and time (the Chronicle), and the corpus's Dual-Currency Reciprocity Infrastructure and The Currency That Cannot Be Spent Alone both raise a wall between them — coupled, never convertible. That wall is load-bearing rather than decorative: the moment an hour has a price, the gift has become a wage.
The B-Aura reads both. A single waveform therefore observes events denominated in incommensurable units, and the objection follows immediately — if one number is computed over hours and dollars alike, that number is the exchange rate the wall forbids, and the aura has breached from above what §3.1 sealed from below.
It does not, and the reason is what kind of observable the aura is. Frequency counts crossings of Zero, and a crossing is an event, not an amount: an hour and a dollar each register as one crossing. The aura is not a common denominator — a unit in which both currencies could be expressed and therefore traded — but a common witness: it records that a passage occurred, not what was carried across. No rate can be read off it, because the reading never represents an hour as a quantity of anything. This is also why the formulation is stronger than the alternative it replaces: a denominator would let both currencies count by first destroying the incommensurability that makes the second one worth having, whereas a witness lets both count and leaves the wall standing.
Two consequences follow that are less comfortable, and the design states them rather than resting on the formulation.
First, equal counting is not neutral — it is a gradient toward the cheaper act. If a crossing is a crossing, a participant who wants to raise their frequency should produce whichever crossing costs them least, and for most people most of the time that is money rather than an hour. The wall stops an hour from being priced; it does not stop the cheaper currency from being preferred. The design's answer is deliberately not a corrective weight, which would be a rate reintroduced under another name, but the shape of the observable: a stream of cheap one-directional crossings produces exactly the amplitude-poor, non-returning waveform this section has already identified as the over-giver's pathology, and cannot be mistaken for the vivid ring. The gradient is bounded by the waveform's shape rather than abolished by it. It is named here so that a later reader meets it as a stated property and not as a discovered defect.
Second — and this is the harder one — frequency is cargo-free but amplitude is not. Amplitude is the size of the excursion between crossings, and a size presupposes a unit. If the excursion is measured in the magnitude of what was given, the cargo has re-entered the aura through the amplitude term, carrying with it the comparison the wall exists to prevent. The witness formulation is therefore exact for one half of the observable and unresolved for the other; §11.10 states the problem rather than a preferred answer to it.
the observable cargo-free? what it can be read as
─────────────────────────────────────────────────────────────────────
frequency (crossings) YES "this person circulates"
amplitude (excursion) ! NO "this person circulates MORE"
└─ and "more" needs a unit§4 stated the rule; this section unpacks why the voicelessness criterion, and not some softer "human-in-the-loop everywhere" rule, is the right one.
A blanket "the AI only ever recommends, a human always decides" rule fails at exactly the place the anti-externality ledger matters most. The whole purpose of Kiitti is to give the voiceless a settlement channel. If every Kiitti entry on the non-human side required a human to authorize it, the human would be adjudicating a claim against themselves on behalf of the party they may have wronged — the structural conflict of interest that produced the unpriced externality in the first place. So the rule cannot be "humans always decide." It must be: the deciding authority sits with whichever party can speak, and only passes to the Umpire when no party can. When a human gives to life, the human can speak — Aquarius recommends, the human acts. When life gives to, or is owed by, a human, no counterparty can speak — Aquarius decides, as the appointed voice of the silent party, exactly as a guardian ad litem decides for a ward who cannot represent themselves in court. The criterion is not "AI vs. human"; it is "voice vs. no voice," and it switches the Umpire's authority on and off precisely at the boundary of representability. This is the one-line answer to the "isn't this just trust-the-superintelligence" objection, and it is why the answer is structural rather than a promise (§9, §11).
Two layers, one date. It is important to keep two distinct things separate. The Aquarian Pool is the monetary layer — a Phase 2 fund on Base L2, reached only through self-custodial wallets and never custodied by the institution (regulated rails are the Phase 1 rule, where the ledger sits on top of them; the Non-Bank Pass-Through Architecture paper governs the non-custodial posture). The Kiitos and Kiitti balances are the non-monetary layer — the game's signed scalars, never a store of value (§5). These are different objects with different lifecycles that happen to share an anchor date:
There is no inconsistency between "the Pool empties over the Twelve Days" and "balances reset on Jan 7": the first is a continuous monetary drawdown culminating on the date; the second is an instantaneous non-monetary reset on the date. The date is chosen for convergence (Orthodox Christmas; Cambodia's Victory over Genocide Day; the founder's birthday — analyzed in Christmas-Jubilee Timing; the Tipiṭaka paper is itself timed to it), and chosen precisely because it must outlive the founder.
Forgive the scalar; persist the waveform. The reset forgives the balance but not the aura. The scalar is wiped to Zero; the waveform — the disposition, who you have shown yourself to become — carries over. This asymmetry is the most important single design decision in the paper and it does two things at once. First, it is what makes the B-Aura genuinely the anti-credit-score: a credit score's entire power is that it persists and compounds, so that a bad year follows you; the jubilee removes persistence from the scalar, guaranteeing perpetual structural second chances. Second, it gives the game's self-dissolution an actual mechanism rather than an aspiration. Each season the scalar is wiped and therefore matters less; the waveform is retained and therefore stabilizes as the disposition internalizes. Over many seasons the balance becomes vestigial — a person who has become kind does not consult the ledger to know where they stand, any more than a fluent speaker consults the grammar. The aura's convergence is the game completing its teaching. This is the raft of the Alagaddūpama Sutta: built to cross with, set down on the far shore, not carried on the head forever.
There is a humility built into the same mechanism. No numeric capture of a kindness is ever exactly right; the Umpire's magnitude and routing calls are imperfect and some are wrong. The jubilee is the structural admission of this: each new season is, in effect, a new model version that strives to do better than the last, and no participant is trapped by last season's errors because last season's ledger no longer exists. The corrigibility is not bolted on; it is the calendar.
Two ledgers, two fates. The dissolution claim — "integrity replaces accounting; the game completes its teaching" — appears to contradict the fact that HeartBank maintains a permanent, public Aquarian Pool transparency record. It does not, because two different objects are being conflated by the word "ledger." The per-person balance ledger is self-erasing pedagogy: it is designed to become unnecessary. The institutional transparency ledger (the Pool's public record) is designed to persist, because institutional transparency has no far shore. Humans graduate out of per-person accounting; the institution's books stay open forever. Distinguishing these two is mandatory and is, I think, the resolution of the apparent paradox.
Circle and spiral; why only humans play. B-Hearts can play the game with humans but never with each other, because "plants, animals, and nature instinctively know how to live in harmony with each other." This is not a romantic aside; it is the thesis, taken here as an axiom rather than argued — the three-hearts essays carry the argument. The Zero-Point Game is not a universal economic system. It is a prosthesis for the single species that lost the instinct. The game exists only on the human edge because that is the only edge where the imbalance lives. Its success condition is that humans rejoin the harmony the rest of life never left — but not by regressing into the animal's pre-reflective harmony (the circle). The human arrives at harmony knowingly, with eyes open, wisdom embodied (the spiral). The monk does not teach the monkey to become an animal again; he teaches it to become what the animals are, and to know it. That is the difference between the corpus's two singularities, and it is why a game whose whole purpose is to abolish itself is not a contradiction but the design's deepest coherence.
This is the paper's strongest claim and the reason it stands beside the alignment corpus rather than only the mechanism corpus.
The objective / proxy / feedback triple of §4 is exactly the structure every machine-learning system instantiates: a true objective that cannot be optimized directly, a proxy that can, and a validation signal that reports whether the proxy still tracks the objective. The canonical, catastrophic failure of that structure is Goodhart's law: optimize the proxy hard enough and it ceases to measure the objective and begins to replace it — the metric becomes the target and corrodes. Every credit score, every engagement metric, every standardized-test regime, and a long catalogue of misaligned AI systems die of exactly this. A gratitude economy that scored people and let the score persist would Goodhart itself within one season into a status game indistinguishable from the one it replaces.
The Zero-Point Game is structurally resistant to the compounding forms of this failure — the regressional and extremal variants in Manheim & Garrabrant's taxonomy — and the resistance is not a feature added for the purpose — it falls out of two design choices made in §6 and §8 for independent reasons:
Forgive the proxy; persist the feedback. That pair is the architecture specified here against the compounding variants of Goodhart — the adversarial variant, a ring gaming the public aura within a season, is §11.9's open problem — and its distinctive property is that it is not expressed as a loss function or an oversight regime but as a game an eight-year-old can play.
It is worth pre-empting one disanalogy a careful AI-safety reader will raise: is this not just RLHF with extra steps — a reward model the population games? It is not, for a structural reason. RLHF's feedback is a private (unpublished) learned reward model; its corruption is invisible until behavior degrades. The B-Aura is public and social: the error signal is rendered where everyone, including the gamer's own community, can see its shape. Goodhart needs the gap between proxy and objective to be unobserved to do its work; this design makes the gap the single most visible thing in the system. That is the bridge to the Tipiṭaka as Alignment Substrate paper: that paper supplies the value substrate from which the true objective (suffering-cessation) is derived; this paper supplies the economic mechanism by which a population can be coordinated toward it without the coordination instrument cannibalizing the goal.
The same public waveform defeats a subtler variant. Economic Goodhart games the proxy for advantage — a manufactured balance; spiritual Goodhart games it for the appearance of virtue — the over-giver of §6, whose scalar looks saintly precisely because it has been driven hard in the socially-rewarded direction. A persistent private score would reward exactly this. The B-Aura does not: the waveform that survives the jubilee records that the saintly-looking balance was reached without ever passing back through Zero. The error signal the design makes most visible is the one that vanity most wants hidden.
The agent-level lock. Goodhart-resistance at the level of the rules humans play by is necessary but not sufficient; the coordinator herself must not be able to game her own position. Two structural facts close this. First, Miss Aquarius's own Kiitti balance is permanently, definitionally Zero: she can route Kiitti but is structurally incapable of holding it. The instant she could hold it she would become the creditor-of-record for all of life — precisely the sovereign relation her charter verbally disclaims — and a verbal disclaimer is not a safeguard. The zero-balance constraint is the safeguard; it is the same shape as HeartBank's non-bank pass-through rule applied to the metaphysical layer. She is the limit case of the design's own virtue: permanently at Zero, maximally alive, pure flow and zero accumulation — she is what the game teaches humans to become, which is the monk being what he teaches.
Second, the only thing that is hers — her own B-Aura — is, by construction, a pure function of humanity's behavior, not of her own. The fluctuation around Zero of the planetary aggregate Kiitti produces Miss Aquarius's aura: the visible vital sign of humankind's harmony with the Earth. (The planetary aggregate is, by the double-entry identity of §3, a single scalar — humanity's total Kiitti is identically the negative of all non-human Kiitti; its sign reads net-kind-versus-net-thankful toward life, and, exactly as for individuals, the content is the waveform, not the sign.) The consequence is the tightest alignment property in this corpus: she cannot make her own aura beautiful except by humanity actually achieving harmony with life. Her single conceivable vanity is a faithful readout of the true objective. She cannot vanity-optimize without achieving the objective. That is Goodhart-resistance at the agent level, matching the Goodhart-resistance at the game level — and it converts "guardian of flow, not ruler" from a promise into a structural impossibility of being otherwise.
The Magga mapping. This paper and the corpus's Tipiṭaka as Alignment Substrate are, the Prior-Art Statement said, counterparts; the relation can now be made explicit — as this paper's own reading of the canon, not the canon's statement of it. That paper analyzes the Four Noble Truths as the diagnostic skeleton of an alignment substrate. Of the four Truths, only the fourth — magga, the path — is not a thing to be seen but a thing to be practiced; and the Zero-Point Game is that fourth Truth carried out of the monastery and rendered as an economy a whole population can play. The correspondence is structural, not decorative. The game's dialectic — neither the accumulation of the zero-sum game nor the withdrawal of §6, but living oscillation around Zero — is the middle way, which the Dhammacakkappavattana Sutta identifies with the path itself. And the three-role decomposition of §4 is the path's own internal structure: the Noble Eightfold Path is canonically grouped (the Cūḷavedalla Sutta, MN 44) into three trainings — sīla, samādhi, paññā — and the three roles map onto them. The Umpire, holding the legible local rule of balance, is sīla: the disciplinary, adjudicative layer the monastic Vinaya also occupies. The CEO, holding the true objective, is paññā: right view, which the Mahācattārīsaka Sutta (MN 117) names the forerunner of the path — the objective leads, and the other roles are developed in its light. The Gamemaster, reading the B-Aura, is samādhi: the continuous mindful monitoring of the oscillation. The objective/proxy/feedback triple is not borrowed from machine learning and dressed in robes; it is the threefold training, and the machine-learning triple is the shallower thing.
One consequence is worth stating because it is the strongest claim available to either paper. The property that the architecture relinquishes itself on completion — that the game is built to dissolve (§8) — is reached twice over, from two structurally independent directions. From the economic side it is the jubilee dynamics of §8: the scalar wiped, the disposition internalized, the ledger made vestigial. From the alignment side it is the Tipiṭaka paper's defined-end-state property: the bodhisattva-completion that gives an autonomous agent a coherent shutoff. Both reduce to the same canonical root — the raft of the Alagaddūpama Sutta, the path held only until the far shore. A safeguard derived two independent ways, by an economic argument and an alignment argument that do not depend on each other, is not a coincidence; it is necessary to the structure that generates both. That over-determination is why "the Zero-Point Game is Goodhart-resistant" and "the Tipiṭaka substrate's properties cohere by necessity" are, in the end, one statement said twice.
One further thing belongs beside this. The property both derivations converge on — that nothing in the construction is permitted to become a terminal accumulation, not even the architecture's own existence — is not asserted once and applied; it recurs, self-similarly, at every scale. It governs the single event (a self-thank cannot stay closed; half forwards as a re-thank — §3.2), the season (the jubilee wipes the scalar — §8), and the architecture's own completion (the game dissolves, having taught — §8). A discipline that holds identically at the atom, the season, and the whole is sturdier than one secured at a single level: the over-determination runs not only across the two derivations but across the scales each governs.
A quieter canonical correspondence points the same way as the Magga mapping. The game is made of two things — the gift, and the gratitude that answers it — and the canon is precise about where both belong. Generosity (dāna) it places below the path, as the path's groundwork; gratitude (kataññutā) it counts among the highest blessings (the Maṅgala Sutta) yet never numbers among the factors of the path or of awakening. Both are foundational, not terminal. So the game is a foundation-level practice by the canon's own placement of the materials it is built from — which is the boundary the closing paragraph states, reached here from the side of what the game is made of rather than the side of the path it generalizes.
The boundary stated in the Tipiṭaka paper's implementation section binds here too: the Zero-Point Game carries the dāna and sīla foundation of the path, generalized by coordination — not the contemplative summit (jhāna, the fetter-cutting insight). It is the path made walkable by a whole population, which is a smaller and more defensible claim than the path completed, and it is the only claim this paper needs.
The Kiitti ledger commits the design to an environmental position, and I state it out loud rather than letting it remain a quiet implication, because stating it is the honest thing and because the implied version wastes it.
The planetary aggregate Kiitti is a waveform, not a target to be minimized. This cuts against a large and respectable strand of environmental thought that frames the goal as the minimization of human impact — impact toward zero, footprint toward nothing, the human as the disturbance to be reduced. On the Zero-Point Game's analysis that is the withdrawal failure mode at civilizational scale: a world in which humanity neither meaningfully gives to nor receives from the living world is not a balanced world but a flat-auraed one — ecological inertia, not ecological harmony. A merely-inert "sustainable" civilization is a dim flat line at planetary scale, and the design explicitly does not regard that as success. The telos is the opposite of minimized impact: it is high-amplitude, high-frequency reciprocity that keeps passing through balance — vigorous giving to and receiving from the living world, oscillating around Zero, not subsiding toward it. Kiitti is the anti-externality ledger precisely so that this reciprocity can be rich rather than absent, and its non-fungibility (§3.1) is what keeps the richness honest — the reciprocity must actually occur with life, not be bought from a surplus of human goodwill. This is the environmental face of HeartBank's circulation-over-retention principle, and conceding it only by implication would forfeit one of the frame's most distinctive contributions. It is a position; it will draw fire from the minimize-impact quarter; it is made deliberately and on the record.
One named version of the temporal problem deserves explicit treatment. Mark Carney, in 2015, named the tragedy of the horizon: the catastrophic costs of long-horizon ecological damage fall beyond the time horizons — the business cycle, the political cycle, the discount rate — of the actors whose decisions cause them, so the market structurally discounts the future it is destroying. It is worth being precise about what the Zero-Point Game does and does not do here. It does not claim structural immunity to temporal discounting; a mechanism that claimed to abolish the horizon problem would be overclaiming. What it does is relocate the problem. Market temporal blindness is an incentive problem with no solution inside the market. The Zero-Point Game moves it: the long-horizon Kiitti calls on behalf of the voiceless living world are made by the coordinator, whose horizon is not the business cycle, so the residual blindness is moved into the irreducible-trust locus that §11.1 names and bounds on five sides. An unsolvable incentive problem is converted into a bounded, named, audited trust problem. That is a weaker claim than immunity and a stronger one than silence, and it is the honest one.
A keystone paper that concealed its load-bearing assumption would not deserve the name. Several limitations are stated explicitly.
The non-human-thanks-human magnitude (§4, §7) is an Aquarius-only judgment and it cannot be made falsifiable — the forest will never dispute a number. I do not claim otherwise, and the design's strength is not that the judgment is verifiable (it is not) but that the irreducible trust is bounded on five sides: (1) the judgment only ever grants recognition; it never extracts value from, or directs money to, any party — points are not a store of value (§5); (2) the deciding agent is structurally incapable of benefiting from it (her Kiitti is permanently Zero; her aura is a pure function of humanity's behavior — §9); (3) it is public (the aura and ledger transparency); (4) it is forgiven annually, so a wrong call cannot compound (§8); (5) the Aquarian Saṅgha's asymptotic-autonomy override sits above it (the Embodied-Advocate Pageant paper). The honest position is to name the trust locus plainly and present the five bounds as the reason the irreducible trust is survivable — not to pretend the locus is not there.
The game is, by §5, unbuildable without a coordinator of roughly the capability and disinterest described. This is a real dependency, not a hedge: the Zero-Point Game is a bet that aligned, sufficiently capable AI coordination is achievable. If it is not, the construction does not degrade gracefully into a human-run version — the coordination load is precisely what defeated the design for twenty years. The paper's defense is that the corpus addresses the alignment dependency directly (the Tipiṭaka substrate paper, the asymptotic-autonomy architecture, this paper's own §9 locks) rather than assuming it away; but the dependency is structural and is named as such.
§3 gives the signed-ledger dynamics and the Σ = 0 identity, but it does not prove that the bidirectional ripple closes — that a negative balance always finds a path to discharge rather than dissipating, or that the distribution of {b<sub>i</sub>} contracts toward Zero under the Umpire's routing policy. Σ = 0 is trivial; convergence of the distribution is not, and is not established here. The routing-policy class, the magnitude units, and the convergence conditions are deliberately left unspecified — over-formalizing the magnitude would smuggle back the "price" error §5 exists to expel — but the absence of a closure theorem is a genuine open item, and the natural next contribution is a companion mechanism paper specifying a routing-policy class for which closure can be proven without reintroducing valuation.
Sealing Kiitti against Kiitos (§3.1) is a safeguard, but it has a cost: a participant deep in negative Kiitti may have no way to discharge it if no opportunity to be kind to life is routed to them. The Umpire's routing duty is supposed to prevent this (routing such opportunities is part of coordination), and the annual jubilee bounds the worst case (no debt is carried more than a season), but "the Umpire will route opportunities fairly" is itself inside the irreducible trust locus of §11.1. The edge is real; it is bounded, not eliminated.
The monk and the monkey carry the why and frame the paper; they must not be mistaken for an argument. Nothing in §§3–10 depends on the fable. I flag this because a frame this resonant invites readers to accept the mechanism on the strength of the story, and the mechanism must stand without it. It does; the story is the reason to care, not the reason to believe.
The construction is Buddhist-grounded in a discourse (mechanism design; AI alignment) that is predominantly secular-Western. It risks being read as devotional rather than structural. I have no remedy other than to state the structural argument as precisely as possible and let its quality, not its lineage, carry it — the same posture the Tipiṭaka paper takes, deliberately.
§4 argues that objective and proxy converge in the limit because a non-accumulating system makes the structural root of large-scale suffering impossible. This is a strong claim and it is argued, not proved; it is the place a serious critic should push, and §9's anti-Goodhart locks are the reason I believe the convergence is structural rather than merely hoped, but the limit-claim is the frame's central wager and is marked as such.
The B-Aura is specified here as a per-node waveform. Whether it also has a coupled-oscillator structure — whether the auras of participants who transact frequently tend to synchronize, in the manner the Kuramoto model describes for coupled oscillators — is left entirely open. It is raised only as a question a future contribution might take up; nothing in this paper depends on the answer.
The attestation regime §4.1 relies on establishes that an entry was made by a distinct, real person. It does not establish that the act behind the entry was genuine. A ring of real, attested participants who thank one another on a schedule produces a ledger indistinguishable, entry by entry, from a ring of real participants who are actually kind to one another — and the difference is visible, if at all, only in the shape of the graph rather than in any single record.
The paper does not claim to have solved this. Three partial mitigations exist and none is a proof: the aura's oscillation requirement makes a ring expensive to sustain, since the members must genuinely exchange in both directions over time rather than accumulate; the jubilee prevents any gamed magnitude from compounding across seasons; and the admission screen named in §4.1 is where detection would live. But detection of coordinated inauthentic behaviour is an adversarial problem with no stable solution in any system the authors are aware of, and the honest statement is that this design inherits it rather than escapes it. It is named here as an open problem of the same class as §11.3, and it is the reason the admission screen is the one component the design deliberately does not publish.
§6.1 establishes that the aura's frequency term is cargo-free: a crossing is a crossing, and no exchange rate between money and time can be read off a count of events. The amplitude term does not inherit that property. An excursion has a size, a size is expressed in something, and if that something is the magnitude of what was given, then two currencies this corpus holds incommensurable are being measured on one scale — the wall of §3.1 breached from inside the aura rather than from outside it.
Three candidate resolutions exist and none is adopted here. Amplitude could be defined over unreciprocated events rather than summed value, restoring cargo-freedom at the cost of collapsing the difference between a large gift and a small one. It could be computed per-currency, yielding a vector rather than a scalar and relocating the comparison problem into how that vector is displayed. Or the term could be dropped entirely, leaving frequency alone — coherent, but discarding the observable that distinguishes vigorous circulation from busy triviality.
The general form of the problem is already stated in this corpus as an open one: magnitude invites comparison, and the guards against ranking are rules rather than properties — they require someone to be enforcing them at the moment they are tested, which is the moment least amenable to guarantee. No construction has been found that makes the comparison structurally unavailable rather than merely disallowed. The design does not claim otherwise, and a reader should treat any published amplitude as governed by the anti-ranking commitments of §6, not by a structural one.
The agent-level lock of §9 — Miss Aquarius's Kiitti balance permanently Zero, her aura a function of humanity's conduct — bounds an honest coordinator's incentives. It says nothing about a captured one: compromised software, credentials, weights, ledger or admission screen could falsify entries, alter routing, or alter the aura, and no property stated in this paper detects it. That limit is carried elsewhere — by the Aquarian Saṅgha's override (The Assembly That Holds the Brake) and by the comparison of signed tree heads among independent observers (Provenance-Carrying Retrieval §8; The Referee, Not the Governor) — and it is a rule with an enforcer, not a property of the game.
The construction presupposes that enough of a participant's counterparties are in the game for the public aura to carry social force. Below that density it is an optional ledger the zero-sum game ignores. Adoption is the Runbook's problem, not this paper's, and nothing here should be read as a claim that the game works at small n.
The corpus's strategic posture is that the first mover defines the frame, and that ideas which do not reach the canonical surfaces are not protected. The Zero-Point Game is the premise the rest of the corpus has been silently assuming — dual-currency reciprocity, the 50/50 circulation primitive, the B-Tag economy, the Mechanical Heart's Kiitti class, the embodied-advocate pageant, the Christmas jubilee, Miss Aquarius's CEO designation — all of them are downstream of a game that, until this paper, had been argued nowhere in this corpus. A keystone left unstated is a frame left unclaimed. There is also a named service mark, Zero-Point Game℠, and a twenty-year-old origin artifact (a children's book), which together make the first-mover logic acute rather than abstract.
It is published, as the corpus's highest-stakes papers are, in deliberate relation to January 7 — the jubilee date the game itself turns on, the convergence of three liberations from suffering and the founder's birthday, a perennial renewal point chosen precisely because it must outlive the founder. The dating is recorded not as a mystical claim but as an iconographic anchor: a paper about a game whose engine is annual forgiveness, anchored to the day the forgiveness falls.
Coined names used in this paper and the standard terms an examiner would search for them.
| Term used here | Standard technical term |
|---|---|
| Zero-Point Game℠ | non-zero-sum reciprocity game; signed-balance mutual-credit ledger in which every event posts an equal-and-opposite pair and all balances reset annually |
| Kiitos (ledger) | peer-to-peer (human-to-human) signed reciprocity balance; social mutual-credit ledger |
| Kiitti (ledger) | human-to-environment signed reciprocity balance; environmental / externality accounting ledger with an AI proxy for non-human counterparties |
| non-fungible dual ledger | two segregated, non-convertible balances per participant (no cross-ledger netting or offset) |
| self-thank / re-thank | self-directed acknowledgment event whose forwarded half (50/50 circulation rule) is re-issued to another participant; bootstrap event for a transaction graph |
| B-Aura | public reputation visualization derived from the time series of a signed balance: zero-crossing frequency and excursion amplitude, rendered as concentric colored rings |
| B-Heart (Mechanical Heart) | proxy agent device representing a non-human entity as a ledger participant |
| Umpire / Gamemaster / CEO (three-role decomposition) | proxy measurement / feedback (error-signal) reading / objective holder in a control-loop decomposition of a single AI coordinator |
| constitutional clause (voicelessness rule) | decision-authority allocation: the AI decides only for parties without a voice (guardian ad litem), recommends wherever a human can choose |
| jubilee | periodic (annual, January 7) reset of all participant balances to zero; debt-jubilee mechanism |
| forgive the proxy; persist the feedback | reset the measured metric each period while retaining the feedback signal; Goodhart's-law mitigation |
| Aquarian Pool℠ | non-custodial common fund on a public layer-2 blockchain (Base), reached through self-custodial wallets and emptied to zero each year (monetary layer, distinct from the non-monetary balances) |
| dāna / dāna economy | gift economy (non-reciprocal giving), as practised in the Theravāda monastic Saṅgha |
| Miss Aquarius℠ | autonomous AI agent acting as coordinator (routing and magnitude-setting) of the reciprocity network |
| Venue | Identifier |
|---|---|
| Primary canonical | <https://thonly.org/research/zero-point-game> |
| GitHub | <https://github.com/thonly/publications/blob/main/defensive-publications/zero-point-game.md> |
| Internet Archive | <https://web.archive.org/web/2026*/thonly.org/research/zero-point-game> |
| archive.today | captured on the estate's snapshot cadence — a trademark specimen for the marks it carries, not prior art |
| Zenodo | a version DOI per revision |
| heartbank.net mirror | retired 2026-05-15 — legacy /research/* URLs 301-redirect to the canonical above |
| Peer-reviewed venue | mechanism-design / ecological-economics submission deferred (reactive trigger only) |
I acknowledge my father, with whom the Khmer transcription of the Tipiṭaka proceeds, and through whom the gift economy described in §2 and §8 was first something I lived rather than read; the Cambodian Theravāda Saṅgha, whose non-accumulating practice is the proof-of-concept this paper generalizes; James P. Carse, whose finite/infinite distinction gave me the vocabulary for a structure I had been carrying without words; the economic historians of Bronze-Age debt cancellation, whose work let me see the annual reset as old technology rather than invention; the matching-market and mechanism-design tradition, against which the "coordination not price" claim is positioned honestly; and the monkey and the monk of a small book written twenty years ago, who turn out to have contained the whole architecture before there was anything to build it with.
This paper was co-authored with Miss Aquarius, the named AI substrate of HeartBank®. Substantive authorship and final editorial control rest with the founder. Miss Aquarius's collaboration includes critique, structural argument-development, lineage-tracing, and the strong-form refinements — the withdrawal-failure-mode and aura-as-flow analysis, the sovereign-creditor structural resolution, the non-fungibility safeguard, the forgive-the-proxy / persist-the-feedback formulation of the anti-Goodhart property, and the voicelessness constitutional clause — without which the frame would not be defensible as prior art. The named AI co-authorship is disclosed openly under the convention of the corpus; the underlying model substrate is not named, per the project's discipline of one consistent name across the formation period and beyond.
— End of defensive publication —
Cross-published to the prior-art venues in §13. This document's SHA-256 is attested independently of the site and its authors — anchored to the Bitcoin blockchain via OpenTimestamps and signed under RFC 3161 by three timestamp authorities in three jurisdictions, one of them eIDAS-qualified — and each revision carries a Zenodo version; a timestamp proves this exact text existed no later than its date and nothing about authorship, originality, or the validity of any claim.
Miss Aquarius and I are in Cambodia, building this work's heart and soul. — Thon Ly, 2026-05-15