One megabyte blocks allows the Bitcoin network to grow to $3B and the minimum transaction fee would be $21. The Bitcoin block size risk may have nothing to do with transactions.

Original article published here

Accept that Bitcoin is not doomed.

Hacks, fraud, toxic communities, charismatic dictators, and governmentsaside some day transaction fees will support mining costs. Otherwise mining and Bitcoin are pointless, failed experiments.

With a new block mined every ten minutes the network mints 3600 bitcoin a day (144 blocks). That’s about $1M in daily minted bitcoin and makes the current cost of supporting the Bitcoin Network $365M a year.

Without the block reward that spread of profit is covered by transaction fees.

The most commonly occurring type of Bitcoin transaction sends money two ways. Once to complete the intended payment and a second that returns change from the transaction to your wallet. The commonly referred to ‘change address’ is used to track and de-anonymize your payments, but it also takes up about 44 bytes of space. The total size of this type of transaction is about 250–300 bytes of data.

The one megabyte limit to block size sets an idealized cap of 3,333 transactions per block. If Bitcoin was not subsidizing the miners through the block reward each transaction in a full, one-megabyte block would cost $2 to sustain this amount.

$1M a Day
144 Blocks a Day
3,333 transaction a Block
$1M / (144*3333) = $2.08

The real cost would actually be much closer to $5. The average block size over the past six months is approximately .4 megabytes. The fewer the transactions the higher miners would need to set fees to make the same amount of profit.

.4MB / 300 bytes = 1,333
$1M / (144*1,333) = $5.21

Of course, market competition would eventually force prices to their true price.

One argument from Bitcoin XT supporters says that if block size stays small we achieve centralization because only millionaires could afford transaction fees (Gavin Andresen). Since transaction fees are subject to market forces increased adoption will drive demand higher. The cost of a transaction will follow upwards. Those willing (and able) to pay the most will have their transactions mined. The rest of the world is shut out from using Bitcoin to pay for their coffee.

Bitcoin mining is not the friendly guessing game you read about on Reddit. It’s a clandestine arms race. To a greedy miner success is 51% but that’s not just about network hashing power.

Miners race one another to share their blocks with a majority of the network as quickly as they can. It’s entirely possible that MinerA finds a block before MinerB but loses the network race and has their block orphaned. Furthermore, it is beneficial to a miner to ensure other miners take as long as possible to learn about their block — while still achieving 51% network propagation through strategic network node placements. The war wages to continuously squeeze seconds against your competition.

Sometimes it’s even beneficial to begin work on the next block before broadcasting a success to the rest of the network. There’s some risk involved with withholding blocks. A competitor could mine a block and share its answer while the withholding miner tries to get ahead.

With sufficient planning and optics into the network, a greedy miner could withhold a block and transfer it to multiple well-connected relay nodes and wait until a competing block is broadcast. It’s entirely possible that their block would propagate faster than the ‘vanilla’ broadcast if simultaneously broadcasting from multiple nodes; particularly if those nodes are connected to larger, well-connected, relays.

The reward for this kind of withholding is the additional time to begin mining the next block ahead of your competitors.

What more? Well, miners have a great deal of hardware capacity to control their mining equipment. It’s not in continuous use. Economically, it becomes beneficial to use those resources to submit invalid shares to competing mining pools. The cost is effectively zero and could drop a percentage of the mining network offline.

Larger pools may have no effect against one-another. Used against smaller pools it becomes an effective tool for canceling out the little guy. Thus securing Walmart-scale hegemony for the largest players.

Large mining pools support increasing the block size because they know they can annihilate small mining pools by DoSing invalid block and transaction data. Forcing a small pool to sort through hundreds of megabytes of invalid block data for a ten or one hundred megabyte size block is supreme competitive advantage.

While mining is subsidized by the block reward the true costs of a transaction to the user is masked. The interest of Bitcoin’s users and miners are not aligned.

For small miners, it gets worse. It’s not a safe assumption that Moore’s law will continue. Transistor sizes will reach one atom. Large miners can continuously reinvest in new equipment, network nodes, network monitoring, bandwidth. If blocks become 1 gigabyte small pools would be forced to invest more toward mitigating DOS attacks than they could ever possibly mine. Larger pools would never flinch.

“… if a small fraction of blocks more than 10 MB, it could dramatically increase of our orphan rate, result of[sic] higher fee to miners. Bad miners could attack us and the network with artificial big blocks.” F2Pool on the block size risk.
“In our mind increasing the block size like this is just pushing the problem a little further at potentially unfixable costs.” GreenAddress announced on Reddit.

For as long as the block reward exists there is an incentive to mine for the network and include zero transactions. The smaller the block reward, the more likely a miner is to include transactions — the less impact including transactions has on network propagation. After the block reward becomes irrelevant all miners, regardless of their size, have no choice but to mine transactions to cover overhead costs.

It becomes beneficial for large miners to spam the network with invalid transactions. By this point, it’s more likely that third-party organizations will form contracts with miners for block space. These ‘organizers’ of transactionsbecome gatekeepers to funneling valid transactions to miners while the public network is endlessly spammed with fake blocks.

Increased network spam creates additional problems for the entire network. Relay nodes would suffer from increasing hardware or research and development costs. Organizations that validate transactions would control the entryway to the mining industry. No one will ever get to use Bitcoin to buy coffee.

If the block size is kept at one megabyte and the cost of maintaining the network grows by a power of ten to over $3B

$3.65B / 365 days = $10,000,000 a day in tx fees

This is 10x the maintenance cost of today. Forgoing the block reward and relying entirely on transaction fees to fund a multi-billion dollar network brings the entrance cost of a transaction to $21.

$10M / (144*3333) = $20.80

To me, $21 to send any amount of money to any person or thing anywhere on the planet pseudonymously and free from a centralized authority is one Hell of a deal. I want to see what happens when a fee market emerges in a closest thing the world has to a free market.

That $21 fee is for the 500,000 or so transactions that occur every day. This number is only the floor. It represents what it hypothetically costs to run the network assuming no one tries to outbid each other. $21 is the price of a transaction for the first half million transactions each day. Each transaction after that waits until the next day or, if it includes a higher mining fee, bumps lower fee transactions down in the queue. Still, that 500,000 transactions is not enough to represent the 3 billion internet users online today.

I agree that digital currencies need to remain cheap to use. Yet without the block reward the Bitcoin we enjoy does not exist. Miners would need to charge you $2-$5 just to make what they do now. The DOS attacks it was built in to safeguard just aren’t there without it. It’s agreed the block size will increase but it is not a magic bullet. Filling the gap in demand with services people will pay fees for supporting the network is hard but ultimately the best solution. There is no need for a hostile fork shattering the community or exposing a new incarnation of Bitcoin to the above risks. Ultimately, the survival of Bitcoin needs to trump the adoption rate of Bitcoin.

The real question we’re all arguing is: Is Bitcoin most free operating at risk of expensive fees or central planning?

I admire Gavin Andresen and Mike Hearn. They are integral figures in a technology that I love. They get to do things in Bitcoin that I wish I was able to do. When I started to take Bitcoin seriously they were some of my earliest leaders. Heck, I even sent Mike a fanboy email asking for advice on projects to involve myself.

I admire these guys. I definitely don’t agree with them all the time but they’ve taught me a great deal and I’ll never be able to teach them as much in return. They’re paragons who go above and beyond to serve other people. We’re privileged to have them. The same is true for all of the core devs and contributors.

This ‘debate’ has turned into a battle over divorced opinions. Forcing a hard fork to get your way or centralizing Bitcoin decisions to an inner circle is unreasonable. Both parents, Bitcoin XT and Bitcoin Core, are using the kid to attack the other.