Sydney Bitcoin Meetup (Dec '14) - Hashing Out Hashrates

Hashing Out HashratesUnderstanding Bitcoin’s Network Hashrate as a Market In Its Own

Right

All data as at December 1 2014

Introduction• This presentation will focus on the economics of Bitcoin’s Network Hashrate – glossing over the technological complexities

• Give viewers an idea on the make-up of the network, and how this may evolve in the future due to underlying economic and business fundamentals

Understanding Hashrate• What is Network Hashrate, and what does it tell us?

• Nature & Drivers of Growth • Nature of Competition• The Profit Motive: To mine or not to mine

What is “Network Hashrate”?• In the simplest terms, it is the amount of incentivised computing power currently administering and securing the Bitcoin Blockchain

• Hashrate change is proportionately tied to difficulty changes every 2016 generated blocks (around two-weeks), to ensure a consistent and predictable block generation cycle of around 10 minutes.

• E.g. if Hashrate rises 10% during a 2016 block cycle, difficulty will rise 10%. If Hashrate decreases by 5%, difficulty will decrease by 5%, etc.

Recent Growth History• Since the November 2012 Reward-Halving, the network has seen 62 difficulty changes, with an average difficulty change of +16.95%, with a standard deviation of +11.51%.

• Average time between difficulty changes has been 11.98 days, with a standard deviation of 1.26 days

• The difficulty has decreased 3 times since the halving, with the most recent time being in January 2013

Recent Growth History

Nature & Drivers of Growth

• Main Inter-dependent Drivers– Prevailing Market Price of BTC (Supply/Demand-Driven)– Prevailing Cost to Mine BTC– Competition in the Mining Industry– State of Mining Technology ($/GH, J/GH)

• This does not differ meaningfully from traditional physical commodities markets, like Gold and Iron Ore Mining

Supply & Demand / Cost & Price• In Short-term Equilibrium…

– If Cost to Mine < Cost to Buy, Mine -------> Hashrate Increase– If Cost to Mine > Cost to Buy, Buy --------> Hashrate Decrease

• In Medium-to-Long-term Equilibrium…– If Demand Exceeds Supply -> Price Increases -> Hashrate Increases

– If Supply Exceeds Demand -> Price Decreases -> Hashrate Decreases

• Sustained periods of increases or decreases in hashrate may indicate the strength of supply and demand forces / health of the market

Perfect Competition & Bitcoin Mining

Market Characterist

ic

Application to Bitcoin (short-to-medium-term: 0 – 3 years)

Application to Bitcoin (long-term: 3 years+)

All market participan

ts are “price takers”

“Temporary price makers” dump/buy vast amounts of coins on an exchange, causing dramatic instantaneous negative/positive price movement, respectively. Once done however, market power and future effects are proportionately permanently reduced.

As bitcoins become less concentrated due to inherent scarcity, the gross majority of

all market participants will

become price takers

Homogeneous Products

All bitcoins are homogenous and identical for the gross majority of practical intents and purposes, and will

always be.

No barriers of entry and exit

No onerous barriers to entry or exit can by created by incumbents to restrict competition due to Bitcoin’s open-source and global nature, and impracticality of unified global regulation or licencing requirements,

and this will always be the case

Property Rights

The Blockchain ensures that there is no doubt about ownership of Bitcoins and their owner’s rights, and

this will always be the case

Market Characteristic

Application to Bitcoin (short-to-medium-term: 0 – 3 years)

Application to Bitcoin (long-term: 3 years+)

A Large number of buyers and sellers

There is currently only a relatively small number of buyers and sellers compared to traditional markets,

however, this number is increasing exponentially in an analogous way to other network-effect based disruptive

technologies

Large number of different types of buyers and sellers

(investors, merchants, exchanges,

remittance, etc.)

Zero transaction

costs

Transactions are theoretically free – but free transactions are subject to the possibility of delays. Fees are

not set by the market, and are voluntary based on desired transaction

speed.

Transaction costs will be near zero

Perfect Factor

Mobility

Factors of production (Location, Labour & Capital) are almost perfectly mobile, allowing for adjustments to

changing market conditions

Factors of production are

perfectly mobile in the long term

Non-increasing returns to

scale

Non-increasing returns to scale when an individual miner or pool of miners approach 50% of network power. Huge

disincentives to exceed 50% of network hashing power.

Non-increasing returns to scale

Market Characteristic

Application to Bitcoin (short-to-medium-term: 0 – 3 years)

Application to Bitcoin (long-term: 3 years+)

Profit Maximisation

Miners will sell at the intersection of Marginal Cost and Marginal Revenue, except during positive/negative hype cycles, where sales strategy differs wildly across the

industry

Miners will sell at the intersection of Marginal Cost

and Marginal Revenue

Perfect Information

In the short term, “Price Makers” prevent the overall market from having access to perfect information, as they can individually influence

market price.

Due to the open-source nature of Bitcoin, in the long term, all consumers and producers are assumed to have perfect knowledge of price, utility, quality and mining methods.

No externalitie

s

The only externalities are emissions due to proportion of network using fossil-fuel to provide electricity for

mining, and waste produced by obsolete mining equipment.

Externalities trending to zero due to decentralised low-emission electricity (Solar, Fuel Cell), and improvements in recycling

Trends & Heuristics• It suggested that in a competitive market place, there is a natural tendency for the market to be dominated by three or four players – known as “The Rule of Three” (Henderson, 1976).

• This hypothesis was tested by observing the evolution of roughly 200 competitive markets (Sheth & Sisodia, 2002).

Trends & Heuristics• Typical S&P500 Tech-Sector profit margin > 20%– Semiconductor Businesses = 18-22%

(source: http://www.businessinsider.com.au/sector-profit-margins-sp-500-2012-8)

• Profits are typically high in new markets– Apple’s Original iPhone (2007) helped Apple achieve profits >16% for long periods of time until smartphone competition intensified.

The Cutting Edge – The Difference 6 Months Can Make

June 2014 December 2014 % Change

Hashrate 100,000,000 GH/s 300,000,000 GH/s + 300%Retail-Best

MinerCointerra

Terraminer IVBitmain Antminer

S4*$/GH $2.99 $0.41 - 87.3%

W/GH 1.1 0.69 - 32.5%* - Spondoolies-Tech SP-20 and ASICMiner Prisma have very similar $/GH and W/GH performance

Trends & Heuristics• Moore’s Law (and it’s several variations) -> $/GH will become cheaper at a non-linear rate for the foreseeable future

• Koomey’s Law: GH/J has been doubling approx. every 1.5 years since the 1950s, and has been faster than Moore’s law.

Trends & Heuristics• “Our goal is to get to 0.05 W/GHs, 0.03 $/GHs

miners by mid 2015 and power more than 30% of the bitcoin network," Corem explained, adding that he believes these figures will help the company match its rival firms in the US and China”- Guy Corem, CEO, Spondoolies-Tech

The Mining Mix

* Cloud miners sampled include PBMining (4.2PH/s), Cloudhashing.com ( 3.5PH/s @ $0.59/GH), Hashnest (3PH/s @ $0.44), Cex.io (2PH/s @ $0.63/GH)

Miner Type % of Network Hashrate Typical $/GH Typical W/GH

Cloud-Miners* ~5% $0.63 ~ 0.7Retail-Miners 30% (+/- 10%) $0.4 0.69Manufacturer-

Miners 65% (+/- 10%) ~$0.30 ~0.50

The Mining Mix• The “Weighted Network-Average Miner” would mine for between $0.3365 to $0.3565/GH and 0.548 to 0.586 W/GH

• This heuristic allows us to estimate the following, within a range of +/- 20%:– Approximate cost to attempt sustained 51% attack– Approximate cost to mine a Bitcoin

Conclusion• It’s not rocket science! Understanding it is just a mix of understanding disruptive evolutionary technologies and managerial economics.

• Perfectly Competitive Markets tend to stay that way in the long-term – so expect more groundbreaking technological improvements, more competition, and more of the same old managerial economics at work!