• What a 51% Attack Actually Costs
• Why Hash Rate ≠ Governance
• The Revenue Problem
• The Censorship Trap
• The Real Security Assumption
• What Keeps Miners Honest

Academic papers warn that miners with >51% hash power can destroy Bitcoin. Here’s the math on why that attack is economically suicidal — and why nodes are what actually keep the network honest.

The Threat Model Nobody Likes to Talk About

Every few years, someone publishes a paper claiming Bitcoin is one mining cartel away from collapse. “Selfish mining,” they call it. The scenario: a large miner withholds blocks, builds a private chain, then releases it suddenly to orphan everyone else’s work.

The papers sound scary. The math is wrong.

Here’s why.

§What a 51% Attack Actually Costs

Let’s say you control 51% of Bitcoin’s hash rate. You want to:

1. Double-spend your own transaction (rewrite history)
2. Block everyone else’s transactions (censorship)

The double-spend:

You buy something with 100 BTC. You wait 6 confirmations (~60 minutes). The merchant accepts. Now you use your 51% hash power to mine a private chain starting from the block before your purchase — one that excludes that transaction.

The key question: will the rest of the network adopt your chain?

The answer is no — if nodes are doing their job.

Here’s the mechanism. When your secret chain broadcasts, every node on the network runs the same validation rules it always has. It checks: does this block follow Bitcoin’s rules?

If your block contains a transaction spending coins that were already spent, or violates the coin issuance schedule, or breaks any of Bitcoin’s 200+ consensus rules — nodes reject it. Instantly. Permanently. No matter how much hash rate it has.

The mining cartel’s chain isn’t more valid because it has more work. It’s less valid because it has invalid data. Work is not the same as truth.

§Why Hash Rate ≠ Governance

This is the part most people get backwards.

Bitcoin’s security model is often described as “the longest chain wins.” This is true — among valid chains. But validity comes first. Always.

Think of it like this: imagine a factory that produces cars. A group of workers controls 51% of the factory’s output. They decide to start building cars without safety belts.

Do the other workers adopt those cars? Do customers buy them?

No. The cars are invalid — not because fewer were built, but because they’re unsafe.

Bitcoin nodes are the safety inspectors. Miners are the factory workers. The workers can produce any block they want. The inspectors only accept the ones that meet spec.

§The Revenue Problem

Here’s the economic incentive that makes the attack suicidal.

When you control 51% of hash rate, you earn ~51% of all mining revenue. Bitcoin mining is competitive. Margins are thin. A miner earning 51% of $50M/day in total revenue is making ~$25.5M/day.

Now consider: what happens when you use that hash rate to attack the network?

1. Bitcoin’s price drops — every market participant knows an attack is happening. Confidence evaporates. BTC drops 30-50% in hours.

2. Your mining revenue collapses — you’re now earning in a depreciated currency. Your 51% of a crashed market is worth a fraction of what it was.

3. Your hardware becomes worthless — ASIC miners are Bitcoin-specific. If Bitcoin fails, they’re doorstops. You’ve invested hundreds of millions in hardware that depends on Bitcoin’s survival.

4. The community hard-forks you out — even if your chain technically survives, the social layer of Bitcoin (nodes, developers, exchanges) will coordinate a response. Fork the attacker out. It has happened before — Ethereum Classic survived the DAO hack by forking, but the attacking hash rate eventually gave up because the non-attacking chain had more economic weight.

The mathematical certainty: attacking Bitcoin is more expensive than simply mining honestly and getting paid.

§The Censorship Trap

What if the goal is censorship, not double-spending?

A 51% miner could refuse to include certain transactions. They could try to censor addresses, entities, or entire categories of spending.

Again: this only works if nodes go along with it.

In practice, censorship attempts trigger a coordinated response. The censored parties route around the attack. New mining pools form. The censoring miner burns enormous amounts of electricity earning nothing while their competitor’s blocks get accepted.

Censorship is expensive. Permanently censoring even one entity on Bitcoin would cost more than the entity is worth.

§The Real Security Assumption

Bitcoin’s security model doesn’t assume miners are benevolent. It assumes:

1. Most hash rate follows the longest valid chain — enforced by nodes
2. Attacking Bitcoin is more expensive than defending it — enforced by economics
3. Economic actors run nodes and validate rules — enforced by incentive

Point 3 is the load-bearing one. If nobody runs nodes, miners can do whatever they want with the chain. If enough people run nodes, the longest chain must also be the valid one.

This is why “not your keys, not your Bitcoin” has a companion principle: “not your node, not your verification.”

§What Keeps Miners Honest

The honest answer: nodes, through fee markets.

When transaction demand is high, fees rise. Miners are incentivized to include transactions. The competition for block space means everyone plays by the rules — because breaking rules means your block gets rejected and you earn nothing.

This is the fee market’s real function. It’s not just about security budget (though that’s real). It’s about miner discipline. High fees = miners follow the rules. Low fees = miners still follow the rules, because the cost of cheating exceeds the benefit.

The bottom line: The selfish miner scenario is a theoretical curiosity, not a practical threat. The economics, the math, and the incentive structures all point the same direction: attacking Bitcoin is irrational for any miner with significant capital at stake.

The nodes are what keep Bitcoin honest. Keep running them.

If this was useful, a zap is always welcome. ⚡

@tomford https://primal.net/p/npub10sq0nytnh22gfmcefe03v3jua3qjvtpmf2zp9tgtsw86n238f02q6v8n9y