Snowballing Bitcoin censorship

Recently I came across a Twitter thread which linked to two articles by the same author. You can read them here:



They describe a path through which the minority of hash-power can compel, through game theory, the remaining majority to censor (ie. not include in their blocks) certain transactions.

Inspired by them I write this post.

I remind the reader that Marathon has started censoring Bitcoin transactions which they believe are illegal according to US authorities.

Let the hash-power of the censurers be that of 10% of the network.

Let the censurers enforce two rules:

* they will NOT include any transactions which they deem illegal
* they will NOT mine a block following that which contains txs they deem illegal, instead they will attempt to orphan the “dirty” block

Let the remaining 90% of hash-power be split into two groups:

* complicit miners who don’t mine illegal txs, but have no problem mining on top of them
* cypherpunks who mine illegal txs

# The naive approach

Let the previous block be a dirty one.

In order to orphan it the censurers must mine 2 blocks in the time it takes the rest of the network to mine 1.

The odds of that happening may be described with the square of the share of the hash-power under the control of the censurers.

In our case, 10% of hash-power controlled by censurers, that’s:

10% * 10% = 1%

This means that any miner including illegal txs has to consider a 1% chance of their block being orphaned and them loosing the fees (~1 BTC) + subsidy (6.125 BTC).

Thus all senders who broadcasts a txs to or from any address which is deemed illegal must compensate the miners with an additional:

1 BTC + 6.125 BTC = 7.125 BTC
1% * 7.125 BTC = 0.07125 BTC

At the time of writing that is slightly over $4 000.

This is the premium, over “clean” txs, the senders must come up with between themselves to have their illegal txs included.

# The opportunist approach

In the previous section we discussed the theory behind censurers orphaning the dirty blocks, assuming they would do so through the method of mining two blocks in the time the rest of the network mines just one.

Let us consider a different approach, the remaining hash-power will switch to mining on top of the censored block as soon as the blockchain is split, without waiting for it to outgrow the dirty block branch.

That is to say after the censurers mine 1 block they will mine on top of that 1 block, as opposed to mining on top of the dirty block which came first.

Why would they do that?

The cypherpunks who did not mine the dirty block will prefer the censored branch as it allows them the chance to get the premium paid by illegal txs for themselves, rather than give it the last block’s miner.

The complicit miners know this and know that the censored branch is being already mined:

* censurers
* cypherpunks, except for the one who mined the last block

And so the complicit miners will join that branch as well.

Thus we notice that now the odds of a block with illegal txs being orphaned are no longer 1%, but rather a staggering 10%.

This boosts the required illegal txs premium to:

10% * 7.125 BTC = 0.7125 BTC

Over $40 000 at the time of writing.

# The case for complicit miners

The opportunist case requires complicit miners and you may be wondering why would those even exist, aren’t we all cypherpunks?

Won’t they loose out on the fee premiums brought in by illegal txs?

Notice we already have such a pool and it’s called Marathon.

If Marathon demonstrates the ability to compete despite (or perhaps thanks to) censoring others will follow.

A complicit miner can advertise to investors and authorities how they fight illegal txs themselves and so don’t need to be regulated or forced to do so while also staying in the, relatively, good graces of the Bitcoin community as they are not the ones orphaning dirty blocks.

Pools are companies with investors, an address and quarterly reports. A proactive approach to keep the law away and the investment at a lower risk level are a desirable strategy.

Aren’t we all cypherpunks? Not anymore, Bitcoin is going increasingly mainstream and just as the “nodes” in the existing financial system are all required to follow KYC and AML regulations in order to remain connected to the others, through SWIFT, SEPA, etc, so will the nodes which “interface” with Bitcoin.

This is already happening through KYC and AML checks on exchanges, as Bitcoin expands so will those checks.

The “interface” nodes, the PayPals, Binances, etc are already happy to ban some coins, because of their history, they will be/support the complicit miners even when not required to do so. Why do you think so many delisted perfectly legal privacy-centric altcoins?

Furthermore more and more power companies, often state owned, are mining Bitcoin with their off-peak electricity. They are not cypherpunks, they are the state.

Finally miners can be bribed by the state to be complicit, after all they only need to be paid the premium the illegal txs pay over the odds of their block being orphaned, and please remember the premiums we have considered before ONLY offset the odds of the block being orphaned.

Thus illegal txs will need to pay both the orphan premium and the state bribe premium to even stand a chance of getting mined.

# The proactive approach

Pools may agree to all move from the complicit category to the censurers.


Each one of them gains through the elimination of cypherpunks, thus increasing their hash-power share and compensating for not including illegal txs.

If we assume that 10% already is censurers then they only maybe around 40% more to 51% attack and permanently ban certain UTXOs.

There’s few enough of them than they CAN cooperate with each other and you cannot just decide to join in as a miner to tip the scales the way of the cypherpunks since not only do you need ASICs to hash sha256, but also you will need the cheapest electricity on the market.

# For a more in depth explanation…

… and answers to the questions you might have check out the [second]( link.

Juraj Bednar does a better job at explaining his game theory than I do.

# (Non) solutions

Allow me to just start of with Taproot does not fix this.

The Lightning Network does not fix it, because the censuring and complicit miners will simply not mine the tx which opens a payment channel or closes it.

Meanwhile LN node operators will not want to open channels to illegal UTXOs either, or will close them should they be already open, because they will not be able to close those later and the current Lightning Network ensures you cannot fully empty the channel either, this is done as a precaution to prevent channels partners from having nothing to loose when attempting to cheat their partner.

Even if Bitcoin implements a private script type, just like it’s already trying to implement Taproot, if it doesn’t happen fast enough the holders of illegal UTXOs won’t be able to migrate them to the new script type (address) since those txs won’t be mined.

View Reddit by bbultimateView Source

Leave a Reply

Your email address will not be published. Required fields are marked *

GIPHY App Key not set. Please check settings


  1. neat story. meanwhile bitcoin remains censorship resistant and no transactions have ever been censored. nobody can compel every single miner. nobody. and no game theory can rationalize it. there are many anonymous miners who don’t need to submit to popular opinion based on social pressure. and no social pressure on earth has ever stopped everyone from doing something. and all it takes is 1 miner. you can even mine your own transactions. any nation state who worries about being censored merely needs to start their own mining outfit. the best a popular mob could accomplish is delaying your transaction if they’re lucky. show me 1 case of universally successful censorship, even in meat space. they couldn’t even stop bin laden for how long? it doesn’t take much to get a message out there. and it’s much easier to hide and transmit on the internet. and with decentralized mixing now, forget it. you’re not stopping anyone.

    now, let’s compare that to bitcoin’s alternatives. “snowballing censorship” my ass.

  2. Bitcoin lends its value from being scarce *and* censorship resistant. As soon as miners collectively begin to censor transactions then the value of Bitcoin would definitely plummet. So miners have a strong economic incentive to keep Bitcoin censorship resistant and to not censor transactions.

    So, in your scenario, at first some transactions might get successfully censored and miners may begin to only mine on top of clean blocks. That would be noticed and would cause BTC to significantly drop in value. Miners would at some point stop censoring transactions in order not to bankrupt themselves. Sure, the difficulty would get adjusted to make mining more cheap, but the disruption would be devastating to the value of BTC and that means less reward available for miners (in fiat terms as well as in BTC terms due to episodes of slower blocks), which means less income. Censorship would be an economic footgun.

    Also, nodes have the final say to whether a block is valid or not. Nodes may find a way to be in consensus on whether a block has been censored and begin to consider such blocks as invalid (requires a soft fork). A miner cannot economically compete with other miners if even only a small percentage of their blocks get marked invalid by nodes. So a censoring miner may at some point find that some of its clean blocks get rejected by nodes and is therefore operating at a substantial loss. Soon, they will be forced to give up.

  3. If I’m enough of an issue that my transactions were being blacklisted, I can start my own pool and prioritize my transactions when I mine a block.

    Bitcoin is resilient.

  4. Unnformed person’s thought – How about getting rid of the “Miner” field of the block winner? All miners of won blocks would then be anonymous by default, (unless they really wanted to announce it some other way).

    The miners can tell those in “authority” putting pressure on them to censor they won whatever blocks they want to tell them, they’d all be anonymous blocks anyway.

  5. Very interesting. I can definitely see the orphan rate going up in a situation where some miners are censoring. However, the same is true of miners that do the censoring, right? Like a miner right now that rejects any block that contains tainted transactions will experience a much higher rate of orphan blocks than the honest chain. Until the censoring miners collectively reach 50%, I think being an honest miner would still retain a competitive advantage.

    But I haven’t read the posts in depth yet. It sounds like there’s something important here.

  6. >Taproot does not fix this.

    Wrong, taproot transactions are indistinguishable until its too late, later when the any one part of the possible many scripts hidden by Taproot is revealed

    >Lightning Network does not fix it, because the censuring and complicit miners will simply not mine the tx which opens a payment channel or closes it.

    Wrong, LN channels will be Taproot style transactions and indistinguishable from other ones.

    I do agree that I can see how very small quantities of Cypherpunks can be edged out by these Marathon minders. I wonder if it can go the other way, and if Cypherpunks can fight back against Marathon-mined transactions. Seems like it would be symmetrical to the hashpower behind it.

What do you think?

Bitcoin’s Entire History on a Logarithmic Growth Curve

Whale trading billions of USD, it’s gotta be an exchange doing this right?