Understanding the “Alchemy” of Hash Rate: Characteristics and Challenges of Bitcoin Computing Assets

Explain the key features of different types of Bitcoin computing power assets, their nature as financial instruments, and the challenges currently faced by each market carrier.

Written by: Leo Zhang and Karthik Venkatesh, founder and data analyst of Anicca Research, a computing power and derivative product research organization

The scientific method seeks to understand the nature of things, while alchemy seeks to make things the way people want them to be. In other words, the main goal of science is truth, while the goal of alchemy is to achieve manipulated success.

George Soros, The Alchemy of Finance

The economic value of computing power

Bitcoin is a new computing paradigm . It abstracts storage, communication, and computation from specific hardware. All nodes independently verify each transaction and new block, and choose the chain with the greatest computing power. Bitcoin can achieve state replication without central coordination. But its design is also characterized by trade-offs: in order for all participants to have a complete copy of the database, messages and data storage must be redundant . No matter how much optimization is done on the chain, this redundancy is inevitable. The purpose of the system is not to achieve higher efficiency than traditional distributed computing, but to formalize and publicize the way the computing process is completed.

Why do we need a highly redundant, inefficient, yet completely transparent computing paradigm?

In traditional distributed systems, all nodes and their coordination rules are channels that provide data storage and access to their services based on distributed systems. In contrast, the consensus and transaction rules of the Bitcoin network are uniformly followed, but there is heterogeneity between nodes. Its transactions are sporadically initiated by anonymous participants with different network speeds around the world. In traditional asynchronous distributed systems, there is no single source of time or authoritative coordinator, and it is basically impossible to determine at what point in time a transaction occurred.

Bitcoin solves the thorny problem of time coordination through proof of work (PoW), a mechanism that proves that a certain amount of computing resources have been used for a period of time. As Gregory Trubeskoy explains, "The difficulty of finding a hash that matches the answer plays the role of a clock ." Through the ticking of this decentralized clock, transaction blocks are time-stamped so that a P2P network can work together effectively. Hashpower condenses the order of multiple transactions on the ledger, allowing Bitcoin to automate trust and independently facilitate the transfer and storage of value with strong guarantees under the premise of achieving strong settlement capabilities.

Settlement guarantees are a key foundation for a settlement network to be adopted and operate over the long term. This settlement guarantee, or economic finality , is considered strong enough when the order of multiple transactions is difficult to tamper with. As blockchains become more popular and the network's computing power continues to accumulate, the "energy attached" to proof of work increases the economic weight of the block. The cost of acquiring 51%+ of the global computing power to launch an attack will become too expensive (compared to the potential benefits).

All economic activities flow into the settlement network. Saifedean argues in Economics of Bitcoin as a Settlement Network that Bitcoin ’s ability to handle large settlements is superior to the settlement system between central banks and financial institutions due to lower costs, greater verifiability, and no counterparty risk.

The Fedwire system processes more than $2 trillion per day. But only a few chartered banks can benefit from the value accumulated in this top-level economic activity. In the Bitcoin blockchain, anyone can contribute computing to keep the heartbeat of Bitcoin, and in return, people will receive a certain share of the economic value. Computing power ensures the security of the Bitcoin network, and in turn, the economic value of computing power is fundamentally driven by economic behavior on the network. This is the bridge that the Bitcoin paradigm connects energy with digitized information. This is the yin and yang of computing power.

Computing power asset class

We define hashing power assets as all assets that generate hashing power in exchange for cryptocurrencies, as well as synthetic contracts/financial instruments that simulate mining returns. Despite its short history, investing in this asset class has become increasingly popular. To meet this demand, many special purpose entities (SPVs), verticalized mining institutions, mining infrastructure/service providers, and financial contracts related to mining have emerged. In this chapter, we will describe the key characteristics of different types of hashing power assets, their nature as financial instruments , and the challenges currently faced by each market carrier.

Mining machine market

Unlike the purely digital goods that are mined, the operation of mining machines is greatly affected by physical properties such as quality and location. The production of computing power involves many external factors, such as chip design, tape-out, supply chain, electricity, and mining machine maintenance.

New products are usually released before the flood season in southwest China to attract the interest of miners, and new machines are generally sold out immediately after they are released. Mining machine manufacturers collect the willingness of large mining farms and distributors to determine pre-orders, and only a small number of machines are used for online retail. The latest generation of mining machines is usually shipped about 6 months after the product is announced. Buying a new mining machine from a mining machine manufacturer is similar to buying a regular supply contract for oil before the 1980s. The oil seller agrees to provide a specific amount of oil to a buyer on a specific date in the future, and the price is unilaterally determined by the oil seller.

After 2018, mining machine manufacturers have become increasingly cautious in inventory management. Mining machine manufacturers only start assembling mining machines after orders are confirmed and aggregated, and buyers usually have to wait 2-3 months, which is the logistics time. Since manufacturers only release a small number of machines to the public, retail buyers can usually only buy mining machines through distributors and need to pay a premium. If the distributor has inventory, the benefit is that the buyer can get the mining machine as soon as possible. Based on the distributor's geographical location and inventory status, the price of the same mining machine can vary greatly:

Source: asicminervalue.com

There is also a sizeable secondary market for used mining machines. Trading used mining machines requires extensive experience. The information in the market is highly asymmetric . Transactions are usually peer-to-peer, and sellers often know the quality of the mining machines better than buyers. Used mining machines are usually out of warranty, and it is not uncommon for them to not run the expected hash rate. Not to mention that some sellers are outright scammers. When purchasing mining machine products from distributors or the secondary market, it is important to choose reliable distributors and channels with a good reputation, and to clearly promise in the contract how to compensate if the mining machine is delayed or does not meet the expected performance.

The mining machine market is notoriously illiquid. Some models are more easily available on the secondary market because they have been active for a long time or have higher production. Mining machines are a commodity . Mining machines of similar power from different manufacturers may have roughly the same price per unit on the manufacturer's website, but on the secondary market, everything is determined by supply and demand . So even though Shenma and Canaan Creative have rapidly eroded Bitmain's market share over the past two years, Bitmain's mining machines have dominated transactions on the secondary market so far in 2020.

Source: Luxor Mining

The Dimension of Computing Power

There are many factors that affect the price of mining machines. In this section, we will build a simple approximate model for the estimation of hashrate and deconstruct how it responds to changes in the underlying variables.

The value of computing power fluctuates with changes in Bitcoin price, network computing power (essentially difficulty, but considering that most mining pools calculate payments based on expected value, hash rate is a sufficient proxy indicator) and transaction fees. Using the mining machine data on Hashrateindex , we can compare the relationship between the historical price changes of different mining machines and these variables:

Data source: Hashrateindex, Coinmetrics

Bitmain's S19 and Shenma M30 mining machines are not included due to their short time on the market.

Not surprisingly, mining machine prices track the unit mining revenue most closely. But this does not help us understand the expenses associated with running computing power. In today's market, the most popular metric for evaluating computing power is the static days-to-breakeven (Chain News Note: Generally referred to as payback days in China, static days-to-breakeven), which is particularly valued by the Chinese mining industry. It is easy to calculate and can be understood intuitively. Similar to the " implied volatility " (implied volatility) is an approximation method for option traders to make valuations, the payback days have become a popular approximation method for mining machine valuations:

In the above formula:

• D is the static payback period

• d is the capital expenditure paid in advance

• p is the current price of BTC

• S is the hash rate of the purchased device

• H is the total network hash rate

• m is the coinbase reward, currently 6.25 BTC.

• n is the current average transaction fee per block

• k is the efficiency of the mining machine (J/T)

• r is the total electricity cost ($/KwH)

r (total electricity cost) Backtest the static payback days of an S17 mining machine under different values:

This is a static calculation that does not capture the option value of a mining machine that changes with changes in the underlying variables. This metric has two components: revenue and cost . Before investing in mining, most miners understand their cost structure (hopefully). The unit cost is fixed over the life of the mining machine (hopefully!). Revenue, on the other hand, is determined by three random walk variables:

We can examine the sensitivity of the payback period by looking at its partial derivatives with respect to each variable. This is similar to options delta, which measures how the value of an option changes when the price of the underlying asset changes. The higher the absolute value of the result, the faster the payback period responds to price changes. The formula for sensitivity to price changes is as follows:

Using the above formula, we can estimate the sensitivity of this indicator to BTC price:

The graph shows that as the price of BTC rises, the payback days become more sensitive to price changes . The S9 miner's sensitivity changes faster than the other two machines. It is important to note that price, network hashrate, and fees are interrelated. But if there is no explicit function of dH (p) / dp or df (p) / dp, we assume that the variables are independent of each other. The same analysis can be applied to changes in the total network hashrate:

Sensitivity to changes in network computing power:

Next, let’s look at the sensitivity of average fees per block at different coinbase rewards. As the block reward for Bitcoin mining is scheduled to decrease, transaction fees play a more important role in the source of mining revenue. The sensitivity of the S19 Pro miner is as follows:

Although days to payback is a simple and intuitive static measure of mining machine performance , it removes the inherent option value of mining machines, so the metric itself is highly volatile. To create a more rigorous valuation model for hashrate, we need to treat it as a multivariate option to capture future uncertainty, or to synthesize it using a basket of different instruments. Each of these valuation methods has advantages and disadvantages. We plan to discuss these trade-offs in a future article on the " Architecture of Hashrate Markets " and explain the mathematical principles of each method. In addition, we will introduce other "hashrate Greeks" such as gamma (second derivative, sensitivity of sensitivity) and theta (time decay).

Synthetic computing power

In addition to the complexity of financial assessment, there are many operational challenges associated with buying and running mining machines. For retail buyers, this process may be too arduous and difficult to navigate. A simpler way to participate in mining is to purchase a cloud mining contract . Cloud mining is a primitive form of hash power derivatives that decouples the future production of mining machines from their current physical location. In this section, we will introduce several different variations of synthetic hash power.

Over the past few years, countless cloud mining projects have sprung up and disappeared. The dilemma of cloud mining is that the market is limited to retail buyers , as it is more profitable for large miners to mine on their own. However, evaluating these cloud mining contracts requires a lot of insider knowledge of the mining industry and expertise in complex option pricing. Although in theory, the concept of cloud mining represents the next natural step in the development of capital markets, most current cloud mining projects are considered " scams " (and many of them are scams) for the reasons mentioned above.

As an immature and still small space, the cloud mining market suffers from a complete lack of market standards. HoneyLemon Market is an excellent cloud mining data collector. Using its dashboard, we can see how much the different cloud mining platforms vary in terms of contract terms and prices:

Most cloud mining contracts are currently unprofitable. Due to the built -in premium based on production costs , cloud mining is only profitable in very specific markets. Generally speaking, when the price trend of cryptocurrencies begins to reverse into a natural bull market after a long period of low prices (such as April-May 2019), the conditions are right. The demand for computing power will suddenly increase. Since it takes a long time to purchase and install mining machines, buying cloud mining contracts becomes a way to quickly establish a position. Through cloud mining, investors can also get exposure to new startup projects that are not yet listed on exchanges. In addition, hackers can rent hash rate to conduct 51% attacks on small networks. This is called computing power Darwinism , which is to eliminate poorly designed PoW projects.

Another type of synthetic hashrate asset is mining machine tokens . They are liquid tokens that represent a portion of the equity of a mining machine. Traders bet on the volatility of the mining machine secondary market rather than investing in the coins produced by the mining machine. Although the concept has been around for a while, its trading volume has not grown much. This is mainly because the multivariate nature of mining income makes it difficult for speculators to reach a consensus on the pricing of " liquid mining machines ."

Experienced traders can build synthetic hashrate portfolios. For example, one could go long a cloud mining contract while simultaneously going long an implied hash rate futures on FTX and adding a short position in BTC spot futures. There are a variety of creative ways to use financial instruments to construct a mining portfolio. For funds and trading institutions that don’t want to own and operate hardware, this is an easier way to gain mining exposure.

Here is how the asset mix has changed:

Based on the magnitude of the fluctuations, when price and difficulty move in opposite directions, "long cloud mining contracts" are most likely to generate profits and losses

In practice, building such a portfolio is a very complex task. As the sensitivity analysis of the payback period shows, the sensitivity of mining indicators can change rapidly as the market changes. The hedge ratio of each instrument must be constantly updated. Management also needs to be closely tracked and adjusted frequently. Moreover, there are many limitations in reality, such as poor liquidity and pricing ambiguity . When making risk calculations, traders need to understand exactly where they will have friction and what kind of discount factor calculations should be made. By building a virtualized computing power portfolio instead of actually operating mining, investors' mining operational risks are converted into additional financial risks.

There is also a new way for miners to hedge their exposure, namely hashpower forwards . Similar to renting hashpower on a cloud mining platform, forward contracts allow miners to sell a fixed amount of hashpower at a fixed price over a certain period of time. Unlike cloud mining, they are usually structured over-the-counter transactions and have a high degree of customization flexibility. However, since there is no public benchmark, the forward contract market does not have any established pricing framework. Every transaction becomes a negotiation, and the department executing the transaction inevitably bears some market risk. A recent high-profile success example is the transaction between digital asset financial services company BitOoda and CoinMint , a large mining company in New York.

As more exchanges/financial services vertically integrate with mining pools , these types of transactions will occur more frequently. Mining pools are excellent aggregators of miner traffic , but over the years they have transformed into commoditized software. All want to capture miners' trading volume. The strong liquidity reserves of exchanges enable them to offer creative and potentially risky types of transactions to win miners' trading flow.

Miners can set parameters, such as always selling a certain percentage of computing power, to lock in operating expenses, while their counterparties take the currency flow produced by this computing power from the mining pool. For example, miners can pre-sell 100Th/s of computing power for 30 days at the beginning of each month. Based on the estimated difficulty growth and fee assumptions, the mining pool proposes to pay a prepayment of 0.02 BTC. Miners will lock in the production of 100Th/s computing power for the entire month, which actually transfers the production risk to the mining pool.

Figures are for demonstration purposes only

Binance, OK and Huobi are actively expanding their mining pool businesses to grab market share. We expect other large exchanges to follow suit or to form alliances with existing mining pool operators soon. In addition to exchanges, some lenders and trading institutions are also exploring this aspect: PayPal Finance has launched its Ethereum mining pool, and Three Arrows Capital will launch structured financial products through the Poolin mining pool. In the past, different layers of the cryptocurrency industry were highly isolated and fragmented. After years of trial and error, key infrastructure layers have become increasingly standardized and commoditized. The business models of some of these independent infrastructures are not competitive. Then, integration or verticalization is inevitable.

Hash power investment tool

Investing in mining is a full-time job. Most traditional investors and venture capital firms do not have enough time, resources or expertise to run mining machines or build complex virtualized computing power portfolios. They generally gain mining exposure through mining SPVs or mining enterprises.

In early 2019, we saw a wave of overhyped GPU-powered projects. A rumor spread like wildfire that deep-pocketed VCs were investing hundreds of millions of dollars in Grin. GPU mining farms, which had a dismal fourth quarter of 2018, were resurrected and riding the wave. Investors who wanted to build large positions but didn’t have the manpower or technical expertise to run the rigs ended up pooling their funds together to invest in mining operations. The managers of the SPVs used the investors’ initial investments to procure and run the rigs, and they also took a share of the mining revenue.

Among large BTC operations, it is also common to invest in mining companies. Some of them have become public companies . Mining companies produce a large number of coins every day. So a mining company is both a hardware operator and a manager of a liquidity fund. Many well-capitalized mining projects have failed due to poor management of trading positions. A notorious example is the 2018 bankruptcy filing of GigaWatt , a pioneer in US Bitcoin mining, in Washington. According to court documents, when the company declared bankruptcy, "its assets were estimated to be worth less than $50,000, and its estimated liabilities were between $10 and $50 million."

How operators manage cash flow is a crucial issue. Developing a reasonable selling strategy to cope with changes in market conditions is crucial to the financial success of the fund/company. Below we list the backtest results of four typical strategies:

Assumptions:

• Start Date: 1/1/2019

• Valuation date: 7/1/2020

• Initial investment: $1,000,000

• Number of mining machines: 4840 S9s, each purchased at $206.6. Linear depreciation is calculated over 18 months.

• Total hashrate: 67,761 Th/s

• Total power consumption: 6,389KW

• Average electricity price: $0.04 / KwH

The four strategies demonstrated above represent different risk preferences for BTC and USD positions. Note that the strategy that produces the highest returns may become the least profitable in another market environment . Some miners prefer to hold on to their coins no matter what the market does until they are forced to sell. Based on the manager's goal (accumulating BTC or chasing USD returns), the strategy should be adjusted accordingly. Mining companies equipped with good traders can also sell at high prices and buy them back when the price drops below the cost of production, or use a combination of financial instruments such as mortgages, Bitcoin futures , etc. to protect against downside risks. We will discuss this topic in more depth in a future article called hashpower financialization and risk management.

A brief summary of computing power assets

In the second part, we will explore the internal logic of the various variables that drive mining trends and introduce how reflexivity works in the hashrate market. We will unpack the macro-cyclical patterns that emerge from these complex interactions and use real market examples to demonstrate their ebbs and flows.

Source link: mp.weixin.qq.com