Proof of Work Fallacy

Merchants purchase mining services that meet their rules for a satisfactory fee. There is a theory that mining services are subservient in this trade. This subservience is sometimes described as "asymmetry" or "users rule". This theory leads people to believe that mining can be strongly pooled as long as merchants are not centralized, as the economy can control the behavior of mining, rendering the system secure. The consequence of this invalid theory is complacency regarding the insecurity caused by pooling.

Miners control transaction selection, while merchants control property offered in exchange. If some part of the economy is unsatisfied with the selections of miners it can offer its property for sale in a split coin with a different work rule that obsoletes all grinding hardware. This is typically described as a proof-of-work hard fork.

According to this theory miners then suffer a catastrophic loss due to the unrecoverable capital investment in highly-specialized hardware. The hard fork may include a difficulty adjustment, allowing confirmation to continue despite a presumed significant drop in hash rate. Due to the lower difficulty and a presumed lack of specialized hardware, more individuals are able to mine. This introduces new miners to the business and reduces pooling.

It has been said that this ability of the economy to foist a capital loss on its trading partners is an asymmetry unique in comparison to other markets. For example, an apple buying community cannot simply "destroy" the orchards of all of its suppliers. The theory fails to understand that there is no asymmetry in trade. If all apple buyers decide that they will buy no apples from existing orchards they certainly have that power. Similarly, the orchards have the option to not sell. Price is the continual resolution of this tension. This is exactly the same relationship that exists in every market.

The theory also fails to account for lack of identity. It assumes that the capital loss will cause the exit of existing "bad" miners and the entry of new "good" miners. This is an unsupportable assumption. There is no reason to believe that existing miners will exit nor is there any reason to believe that new miners would not exercise the same decisions as previous miners given they are in the same business, assuming one could even tell the difference. At least in the apple scenario one knows from whom one is buying apples and can discriminate, this is not possible in Bitcoin.

The theory also fails to account for the economics of mining. There is an advantage to proximity that produces greater returns on capital for miners with greater hash power. Larger miners are therefore more profitable than small miners. Larger miners will therefore be better capitalized than their smaller competition. When the rule change occurs the miners that remain will be those who can afford to retool, which will be the largest.

It is irrational to assume that all miners will simply exit. Would we expect all apple growers to be replaced by new apple growers? In mining are not expertise, facilities, energy contracts, process, and non-specialized machinery important advantages over newcomers? Existing miners have an inherent advantage over their supposed replacements. This means they have greater access to capital. So not only do larger miners end up with less competition, all existing miners that remain have an advantage over any new miners.

The theory also fails to consider that merchants need mining. The process is not replaced by splitting, and this process retains complete control over transaction selection. So for example if the "bad" miners happen to be states that are attacking the coin, the state itself and co-opted miners will continue contributing to the same disruption. Mining services that are "good" for merchants cannot be produced by splitting.

Finally, the theory fails to recognize actual consequences. Based on the previous capital loss experienced by all miners for a given coin, all future miners of its replacement will insure against the likelihood of a similar event. They may self insure, but the increased cost is unavoidable. This will reduce hash rate for the same fee until the possibility of such an event is deemed negligible. So the economy raises its own fees and ends up with the same miners and greater pooling. This is a reduction of security on two levels, with no benefit.