Interstellar Game Theory
Before Bitcoin, if you wanted to prove you owned money, you needed someone to say so.
Not a document. Not math. A person, or an institution staffed by people, whose job was to tell you and everyone else what your balance was. The bank said what you had. The exchange said what your trade was worth. The clearinghouse said what settled. You didn’t own a fact. You owned a relationship with an institution that could, in principle, change its mind.
This wasn’t corruption. It was the only possible answer to a real problem. Financial commitments exist in time. Proving that you had the money before you spent it — that the trade happened in this order and not that one — required a witness. And whoever holds the clock holds authority over what happened.
Bitcoin replaced the witness with arithmetic. When you send bitcoin, thousands of computers around the world independently run the same simple calculation: did this money exist, and has it been spent before? No computer asks another. No computer asks a bank. If the math checks out, the transaction is valid. Not because someone said so. Because anyone can check.
That’s the whole trick. And it’s not limited to money.
Why the Trick Keeps Breaking
Every system built after Bitcoin tried to extend this trick to more complex things — trading, lending, derivatives. And almost every one of them accidentally rebuilt the institution they were trying to replace.
Here’s why: Bitcoin’s trick only works if checking is easy. Bitcoin transactions are deliberately simple — did this money exist, was it unspent? Any cheap laptop can verify that in milliseconds.
The moment you add complexity — smart contracts, automated trading pools, sophisticated financial products — checking becomes expensive. And the moment checking becomes expensive, most people stop checking. The moment most people stop checking, whoever did run the computation gets to define what happened. The institutional witness comes back, wearing different clothes.
This is not a flaw in the people building these systems. It’s a law of economics. If verification costs more than most people can afford, they delegate it. And whoever they delegate to holds authority over their account.
The rule this implies is simple: verification must remain asymmetrically cheaper than execution — cheaper in computer power, cheaper in time, both simultaneously.
If it costs a dollar to do something and a penny to check that it was done correctly, you have an honest system. If it costs a dollar to do something and fifty cents to check it — most people won’t check. You have a trust system again, regardless of what the whitepaper says.
What “Paying Someone to Lie” Actually Looks Like
The clearest proof that bundled payments recreate institutional authority came from MEV — Maximal Extractable Value. It’s a jargon term for something simple.
When you swapped tokens on a popular decentralized exchange, your transaction sat in a public waiting room before it was processed. Automated bots could see it. They would buy the same token a fraction of a second before your transaction went through, then sell right after — pocketing the price difference your trade created.
This wasn’t hacking. It wasn’t even cheating by the rules of the system. It was the inevitable result of one group of people being paid to both order transactions and execute them. When the same payment stream covers both jobs, the person doing both will naturally use their ordering power to maximize execution profit. Anyone rational would.
The industry spent years trying to fix this. Every fix was a bandage. The wound was the bundle.
When payment streams are bundled, authority recombines. The building you were trying to tear down gets rebuilt from the inside.
The solution is not cleverer code. It is separated payments.
The Jobs, and What Makes Each One Honest
A verification-first network has several distinct jobs. Each must be paid for separately, or they collapse back together.
The Record-Keepers — in Zenon, called Pillars — maintain the official sequence of events. Think of them as the notary who stamps “this happened before that” without reading what that is. Their entire income comes from being trusted to keep an accurate, neutral record. The moment a Pillar starts playing favorites — processing one application’s transactions ahead of another’s, or inserting their own execution logic — they’ve stopped being a record-keeper and started being something else entirely.
The Checkers — called Sentinels — are paid to find problems. This is economically unusual, worth dwelling on. In almost every system you can think of, agreement is rewarded. Go along with the crowd, collect your pay. A Sentinel’s pay structure is the opposite: finding a genuine error earns a reward. Staying quiet when there’s an error earns nothing — and costs future credibility. Raising a false alarm costs a bond you put up beforehand.
Think of them as bounty hunters, not security guards. A security guard gets paid whether or not crime happens. A bounty hunter gets paid when they catch something real — and faces consequences for false accusations.
The Carriers — Sentries — are paid to move evidence from where it’s created to where it’s needed. Sentries move evidence but cannot define validity, so their risk surface is operational rather than epistemic. A checker who never receives the evidence can’t check anything. Without explicit payment for carrying, only a few well-resourced carriers will bother, and they become the gatekeepers for what gets checked.
The Users benefit indirectly, but really. When you can verify your own account state without asking anyone’s permission, every service provider who wants your business has to compete on quality rather than lock-in. You may never run the verification yourself. But the fact that you could means the people you deal with can’t take you for granted.
Why Being Paid Isn’t Enough
Here is the piece most incentive systems miss — and the piece that separates a reward system from something genuinely robust.
Paying someone to behave honestly hopes they will. Requiring them to risk capital first ensures they have something to lose before they ever earn a cent.
These are different systems. The first trusts actors until they misbehave. The second makes misbehavior expensive before it begins.
There is one invariant that explains every staking and burning requirement in Zenon:
Anyone allowed to influence truth must first become more expensive to corrupt than the value of corrupting them.
This is not a design preference. It is a control condition. If corrupting an actor is cheaper than the profit from corrupting them, rational attackers will corrupt them. It’s arithmetic. The network’s job is to keep that equation permanently unfavorable for attackers — and the way it does that is by requiring actors to lock and burn real capital before they are allowed to participate.
Think of it this way. A judge who owns nothing can be bribed cheaply. A judge who has put their house, their career, and their pension on the line to sit on the bench is expensive to corrupt — not because they’re more virtuous, but because the price of their cooperation has become very high. The stake is what makes the judge credible, not the robe.
A Pillar must lock significant ZNN and permanently destroy a substantial amount of QSR to operate. Notice that two different mechanisms are at work here, and both are necessary.
The locked ZNN is a stake — capital at risk that can be lost if the Pillar misbehaves over time. This makes long-horizon corruption expensive: a Pillar that degrades the ordering layer watches its own locked value collapse along with everything else that depends on the network.
The burned QSR is different. It is gone immediately, regardless of how the Pillar behaves afterward. This prevents a different attack: join the network, behave honestly just long enough to build credibility, extract value quickly, then exit and recover your stake. Burning forecloses that strategy. There is nothing to recover.
Staking makes dishonesty expensive later. Burning makes short-term extraction expensive immediately. Together they prevent both exit scams and long-horizon corruption.
A Sentinel must lock meaningful capital to operate. This is even more interesting. Sentinels are allowed to accuse the system — to say “this record is wrong.” That is extraordinary power. It must be expensive to wield carelessly. Requiring capital at risk before a Sentinel can dispute anything means every challenge they raise comes from someone already financially exposed to being wrong. They are not just paid skeptics. They are bonded skeptics. Their credibility is collateralized.
Sentries require lighter commitment — intentionally. Their job is infrastructure, not judgment. They move evidence; they don’t rule on it. Requiring heavy staking for Sentries would immediately concentrate them in well-capitalized data centers. Lighter requirements allow broader participation, which is exactly what evidence propagation needs.
This is not inconsistency. It is risk-weighted collateralization. The closer a role sits to defining truth, the more capital must be locked before it can participate. The architecture of required stakes follows directly from the architecture of influence over truth.
Bitcoin did the same thing differently. Miners burn electricity first — irreversibly, whether the block was honest or not. Once that energy is spent, it’s gone. Zenon bonds its actors through locked and burned capital — the same economic law, adapted for a system where the job is verification rather than chain extension.
Irreversible cost precedes authority. In both systems. Every time.
Payment rewards behavior. Stake makes behavior credible. Rewards create incentives. Bonds create consequences. A system that only pays actors hopes they behave. A system that bonds them first ensures they cannot afford not to.
The Inversion That Makes It Work
Here is the shift that most people miss, and that most systems don’t make.
Bitcoin pays people to build the record. Anyone can check the record for free.
This network pays people to challenge the record. Building it is the cheap part.
Those are different businesses. Different incentives. Different equilibria.
In a system that pays for building, security comes from making it expensive to build a competing version. In a system that pays for challenging, security comes from making it expensive to get all the challengers to look the other way simultaneously — and each challenger has already posted real capital that suffers if they’re caught ignoring valid problems.
The question “how do you attack this?” has a specific answer: you must bribe or neutralize enough independently-bonded watchers — who come from different jurisdictions, different infrastructure providers, different network paths, who have no shared single point of failure, and who each have locked capital that evaporates if they’re caught missing real fraud.
The network’s security does not come from execution fees. It comes from paid, bonded skepticism.
Bitcoin solved: who pays to write history?
Zenon solves: who pays to doubt history?
And the answer, in both cases, is the same: whoever is paid to do the job must first make it expensive to do it badly.
Two Tokens, One Sentence Each
Zenon has two tokens. This confuses people until you understand that they’re funding two fundamentally different sides of the same equation.
Recall the invariant: anyone allowed to influence truth must be more expensive to corrupt than the value of corrupting them. Maintaining that condition requires two different kinds of ongoing work — and they cannot share a payment stream without breaking each other.
ZNN secures the past until dependency secures it instead.
Early in any network’s life, no one depends on it yet. Corrupting the ordering layer costs little because the attacker loses little — their stake isn’t worth much, and nothing else depends on the record being intact. You have to pay record-keepers to show up, and you have to inflate the value of their stake to make corruption expensive before organic dependency does it naturally.
Over time, something changes. Real systems start relying on the record — financial products, custody arrangements, settlement protocols. Once that dependency accumulates, an attacker who corrupts the ordering layer destroys their own stake in everything that depends on it. The record-keeping becomes self-defending. ZNN emission can gradually decline because it’s paying for a transition: from “I’m paid to stay honest” to “I can’t afford to be dishonest.” One gradually replaces the other. Emission bootstraps alignment. Dependency sustains it.
QSR secures the present forever.
Checking never stops. Evidence must always move. Challenges must always be possible. This is not a bootstrapping problem — it is a permanent operating cost, like electricity or bandwidth. There is no end state where the immune system can shut off.
QSR is not inflation for security. It is the protocol’s standing budget for verification bandwidth — a perpetual procurement contract that buys low LDL, evidence availability, and independent skepticism. Think of how internet providers buy transit capacity, or how hospitals maintain reserve staff. The work is never done; the budget is never zero; the cost-per-unit may fall as the market matures, but the spending continues.
If QSR emissions decline too much, verification participation drops, LDL rises, users stop independently checking, and authority quietly recentralizes. The system doesn’t crash loudly. It slowly becomes custodial again. That is the real failure mode — not a hack, but a gradual drift back to the world that came before.
One token buys inertia — the accumulated credibility of an unbroken record, maintained until dependency makes corruption self-defeating. The other buys motion — the continuous work of keeping that record honest in real time, forever.
Same goal. Incompatible mechanisms. This is why they must be separate tokens.
The Metric Nobody Is Tracking (But Should Be)
The blockchain industry measures transactions per second. This is the wrong thing to measure for a verification-first system.
Imagine a security camera network for a city. Would you evaluate it by how many frames per second it records? Or by how quickly it detects and flags suspicious activity?
Transactions per second is frames per second. What actually matters is how quickly lies get caught.
The right metric is LDL — Latency to Detect Lies. The time between a false claim entering the official record and the network’s independent, bonded observers recognizing and refusing it.
A fast network that catches lies slowly is a fast network for processing fraud. A somewhat slower network that catches lies within seconds is a network that actually protects you.
This reframes what “scaling” means entirely. The question isn’t how many transactions per second. It’s: how many independently-bonded watchers are watching, how quickly can they detect something wrong, and what happens when they do?
Optimize throughput and you build a fast machine with an expensive fraud problem. Optimize LDL and you build a market where lying is quickly and cheaply challenged by actors who are individually paid — and bonded — to do exactly that.
Why Paying the Fastest Watcher Is a Mistake
The obvious response to wanting fast lie detection: pay the fastest checkers most.
It creates one big vulnerability disguised as many small protections.
When you pay for speed above all else, the fastest nodes cluster in major cloud data centers. They use the same backbone networks. They exist in the same regulatory jurisdictions. Their failure modes are correlated.
One cloud outage, one legal order, one infrastructure attack — and a large fraction of your “independent” verification layer disappears simultaneously. What looked like hundreds of independent watchers was one infrastructure provider with hundreds of accounts.
Real independence means different jurisdictions, different internet providers, different hardware, different network paths. A checker running on a modest server in a small country with a slow connection might be more valuable to the network than a fast node in a major data center — because when the data center’s region has a problem, the modest server still sees.
The payment structure must price this correctly. Pay for independence, not speed, and the network’s security actually improves with diversity rather than quietly centralizing behind a performance veneer.
The Self-Healing Market and the People Who Try to Game It
If independent, bonded checking is what makes the network secure, a smart system pays more when it’s scarce.
The network monitors its own health. When a region lacks sufficient independent checkers, it raises the reward for providing them. Gaps attract participants. Participants fill gaps. No central authority required.
Two obvious ways to game this, and why they fail.
Fake a gap. Coordinate with other checkers to pretend coverage is worse than it is, trigger higher rewards, then “restore” coverage to collect. The defense: the network pays for marginal contribution, not presence. A hundred nodes that always agree with each other look like one node to the payment system. And each node in that cluster has bonded capital at risk if the coordination is detected — making the scheme expensive to sustain.
Fake independence. Set up many nodes that look geographically diverse but actually coordinate. The defense: behavior over time is hard to fake. A cluster of nodes that always reaches the same conclusions in the same order reveals correlation in timing and agreement patterns. Sustaining convincing fake independence across months of behavioral history costs more than it earns — especially with real capital locked in each node.
One timing note that matters: the network detects lies quickly — that speed doesn’t change. What settles slowly is the payment. Truth arrives fast. The invoice waits for verification that the work was genuine. Detection is immediate. Payment is careful.
Where Theory Meets Reality
Six places where this gets genuinely hard. Worth being honest about each.
Checkers who don’t check. If the reward only comes when you catch something, and nothing is wrong for a long time, rational checkers might minimize work and hope someone else catches the rare problem. The fix: pay for demonstrated checking effort — verifiable proof that you sampled random evidence — separately from the bonus for catching an actual problem. Presence is not work.
False alarms as an attack. Without any cost for raising a false alarm, a bad actor can flood the system with fake disputes. The bonded stake structure handles this directly: to raise a dispute, you risk capital. Legitimate disputes return the bond plus a reward. Fabricated disputes lose the bond. False alarms become expensive. Honest alarms become profitable.
Paying people to look the other way. An attacker could try to pay checkers directly to ignore specific fraud. The bonded structure makes this expensive in two ways. First, every checker already has capital locked that loses value if the system’s integrity breaks — they’re being asked to accept a bribe that partially destroys the thing making their own stake worth something. Second, a checker who stays quiet sees their historical track record suffer when others catch what they missed, costing future earnings. The attacker must pay enough bonded, independent actors — across enough jurisdictions — to all look away simultaneously. The cost is superlinear: twice as much fraud to hide costs more than twice as much to suppress.
Record-keepers with subtle power. Even a “neutral” record-keeper has discretion over what order transactions enter the record. Sequencing has economic value. The solution is mechanical: ordering rules enforced on-chain, visible to Sentinels, with no discretion gap that earns money outside the sanctioned payment stream.
Who sets the parameters. Correlation penalties, reward rates, detection thresholds — someone has to set these, and that someone becomes a target for capture. The best defense is automating as much as possible based on measured network behavior, with slow and difficult processes for any manual overrides.
Starting from zero. All these protections work better with more participants. Early on, when participation is thin, the system is most vulnerable. Early rewards must be calibrated to guarantee enough coverage before organic demand takes over. The bootstrap phase is the weakest point — honest design and adequate early rewards can minimize the risk, but not eliminate it.
The Honest Scorecard
The design is genuinely strong. The question — which rational actor deviates, and why is it profitable? — has specific answers for each role. In each case, deviation is visible, costly, and self-defeating over time.
Pillars that capture execution signal it in their payment behavior and risk both their locked stake and their burned commitment. Sentinels that miss real problems lose future earnings and damage track records. Carriers that withhold evidence earn nothing. Fake-diverse clusters earn degraded returns as correlation penalties apply. Bad-faith challengers lose their bonds.
The system doesn’t rely on people being good. It requires that they be expensive to corrupt before they’re allowed to participate, and makes honesty the profitable option once they’re in. It doesn’t ask actors to trust each other. It makes trusting each other unnecessary.
What isn’t solved: data availability under adversarial conditions at scale remains an active engineering problem. Independence measurement will always be heuristic rather than perfect. Bootstrap fragility is real. These are ongoing maintenance problems the model correctly names — not flaws in the design, but the work the design makes explicit and visible.
The system’s virtue is not that it guarantees truth. It is that it makes untruth measurably self-revealing — and makes the people paid and bonded to reveal it individually better off for doing so.
The Short Version
For a long time, financial systems required institutions to say what happened because there was no other way to prove it. You needed someone to hold the clock.
Bitcoin proved the clock could be mathematical instead of institutional. Anyone can read it. No one owns it. The institutional witness became unnecessary for money because verifying the math was cheaper than trusting the institution.
The same logic, extended one level up: who is being paid — and bonded — to make sure the math stays honest?
The answer follows from one constraint: anyone allowed to influence truth must first become more expensive to corrupt than the value of corrupting them.
That single condition explains every staking requirement, every burning mechanism, every dual-token structure. Stake makes long-horizon corruption expensive. Burn makes short-term extraction impossible. Together they close every exit. Pay record-keepers to record neutrally, require them to lock and burn enough that corruption destroys their own position. Pay checkers to challenge honestly, require them to stake against false alarms, pay them for demonstrated work rather than mere presence. Pay carriers to move evidence reliably, with friction calibrated to their actual influence over truth. Let users verify independently whenever they choose.
When those payments stay separated, those bonds stay in place, and the corruption-cost equation stays unfavorable for attackers, the profitable thing to do at each role is the honest thing. Not because of values. Because the math makes everything else too expensive.
The institutional witness was never a feature. It was a cost. A cost that gets cheaper to replace every year.
What replaces it isn’t a better institution. It’s a market for paid, bonded skepticism — maintained by people who profit from keeping it honest and who have already paid to prove they mean it.
Not because they trust each other.
Because they don’t have to.
Zenon’s Network of Momentum is fully open-source and community-run. More formal community documentation and ongoing community research can be found at: https://github.com/TminusZ/zenon-developer-commons