The SegWit2x (B2X) Hardfork – Protecting Yourself and Your Coins, Part 2: Investment Strategies

Abstract: The upcoming SegWit2x hardfork is likely to lead to price volatility.  In this piece we look at some potential investment strategies which could allow you to capitalize on the event.


Strategy 1: Do nothing

The most popular investment strategy following the split is likely to be to take no action and remain a holder of both BTC and B2X. This is probably the most prudent approach, as your assets may be protected whichever coin becomes more valuable. Most people may pursue this strategy out of laziness rather than choice.  However, even if this is your preferred investment strategy, it may still be sensible to try and split your coins anyway; to increase the flexibility of your investment strategy and protect your funds in case you need to make a transaction.


Strategy 2: Invest in your favored coin

Many investors may support either the BTC or B2X coin for ideological reasons or because they feel their chosen coin has the best characteristics.

  • Supporters of BTC typically prefer the consensus rules to be robust, as they feel this results in superior or more unique monetary characteristics. Also they typically value the cautious and meticulous approach of the current development team. BTC supporters may want flexibility and innovation to come from other layers in the system above the consensus rule layer.
  • In contrast to this B2X supporters may value a more flexible consensus ruleset to ensure the system is dynamic and able to cater to user requirements more quickly. B2X supporters typically value the user experience over the monetary characteristics of the system. Typically they draw less distinction between changes in the consensus layer and other types of changes to the system.

If you agree with one of these visions more than the other, it might be a good idea to invest in the coin that matches your vision, this may not only help you obtain larger returns if your vision is correct, but may also help ensure your favoured token is the “winner”, as it may contribute to the value of the coin.

Alternatively you may not care which vision “wins”, but want to back the winner. In this case it’s important not just to gage opinion, but also the level of conviction those on each side of the debate have. Somebody slightly favouring one of the visions but also wanting to hedge their bets may have less of an impact than a die hard supporter of one of the visions, who is willing to sell all their coins on one side of the split no matter what. This factor could favor the BTC side, since many die hard “large blockers” may have already sold some BTC to invest in Bitcoin Cash.

The most common investment strategy after the fork may be to do nothing. However, of the tiny minority of people that do act, many of those people may be sellers of B2X. This is because the section of the minority that took action in favor of the large block chain in August, will not be allocated either BTC or B2X for the coins they sold for Bitcoin Cash. In contrast, the minority that quickly sold Bitcoin Cash in August, will be allocated BTC and B2X this November. Although, obviously this is highly speculative and nobody really knows what will happen.


Strategy 3: Invest in whichever coin is the cheapest/the “bad” coin – The Joel Greenblatt strategy

The top investment tip in one of our favorite books on investing, Joel Greenblatt’s “You can be a stock market genius” (bad title but a great read), appears to be that if a stock split occurs, one should buy the less favored company, the spin-off.  In this case B2X is the spin-off token.

Joel Greenblatt’s Gotham Capital achieved annualized investment returns of 50% from 1985 to 1994. One of the core strategies of the fund in this period was to invest in “bad” spin-off companies. As Greenblatt explains:

There are plenty of reasons why a company might choose to unload or otherwise separate itself from the fortunes of the business to be spun off. There is really only one reason to pay attention when they do: you can make a hole of money investing in spin-offs. The facts are overwhelming. Stock of spin-off companies, and even shares of the parent companies that do then spinning off, significantly and consistently outperform the market averages. One study completed at Penn State, covering a twenty five year period ending 1988, found that stocks of spin-off companies outperformed their industry peers and the S&P 500 by about 10% per year in their first three years of independence.

Greenblatt mentions that spin-offs tend to be driven by a “desire to separate out a bad business so that an unfettered good business can show through to investors”. The bad company or spin-off company is typically sold by investors, with the negative narrative around the bad company dominant at the time of the split, causing negative sentiment. As Greenblatt explains:

The spin-off process itself is a fundamentally inefficient method of distributing stock to the wrong people. Generally, the new spin-off stock isn’t sold, its given to shareholders who, for the most part, were investing in the parent company’s business. Therefore, once the spin-off’s shares are distributed to the parent company’s shareholders, they are typically sold immediately without regard to the price of fundamental value. The initial excess supply has a predictable effect on the spin-off stock’s price: it is usually depressed. Supposedly shrewd institutional investors also join in the selling. Most of the time spin-off companies are much smaller than the parent company. A spin-off may be only 10 or 20 percent the size of the parent. Even if a pension or mutual fund took the time the analyze the spin-off’s business, often the size of these companies is too small for an institutional portfolio.

Greenblatt goes on to cite four spin off case studies, Host Marriott/Marriott International, Strategic Security/Briggs & Stratton, American Express/Lehman Brothers and Sears/Dean Witter, where this thesis applied.

In many ways there are some analogies between the opportunities which may arise from Bitcoin spin-offs such as B2X and stock spin-offs. Perhaps B2X is being distributed to the “wrong people”. Bitcoin investors typically value robust rules and the resulting highly resilient monetary properties. Perhaps, some of the Bitcoin investors who value other characteristics such as flexibility and user experience may have already divested into Ethereum or Bitcoin Cash, therefore the remaining investors may “dump” B2X. The risk for B2X proponents is they allocate their new coin to the “wrong people” and the price becomes “depressed”.  However, this could provide contrarian investors an opportunity.

The price of B2X could fall to cheap levels and there could be significant amounts of negative sentiment with some people writing the coin off. This could then be a good time for contrarians to invest in B2X. This investment philosophy seems to go against a common narrative in the Bitcoin space that “network effect is king” or “the most work chain wins”, meaning a minority chain has limited prospects. However, there may be little investment basis for this view.

However, whether the Greenblatt spin-off philosophy really applies to Bitcoin spin-offs such as B2X is not clear. Greenblatt still does fundamental analysis on the bad spin-off company, and whether one can take this type of fundamental approach to Bitcoin or its spin-offs, is not obvious. It’s certainly more risky. Although, in my view, after the hardfork, if B2X is trading at c5% or less of the price of Bitcoin and the prevailing narrative is that B2X is dead, then the “bad” coin may be worth a small punt.


Strategy 4: Take advantage of different policies on different exchanges

During the Bitcoin Cash hardfork, different financial platforms had different policies. For example BitMEX essentially ignored Bitcoin Cash, and the futures price just followed Bitcoin. However, Kraken for example, supported Bitcoin Cash, in such a way that those with long margin positions on Bitcoin were also given Bitcoin Cash. Critically on Kraken if you were short Bitcoin at the time of the fork, you were then automatically short Bitcoin Cash. These different policies between exchanges provide asymmetry, which in theory can be used to earn free money.

For example going into the Bitcoin Cash hardfork, an interesting strategy was to open a margin long position of 1 BTC on Kraken and then hedge the position by taking a margin short position of 1 BTC on BitMEX. Thereby after the hardfork, you receive one Bitcoin Cash token on Kraken, essentially for free, since there was no corresponding Bitcoin Cash liability on BitMEX associated with the short.

When it comes to the upcoming fork, there are four relevant potential exchange policies one needs to consider, when trying to engage in this type of arbitrage.


Potential financial platform policies regarding the B2X spin-off token

Policy A Policy B Policy C Policy D
Split user Bitcoin deposit balances into BTC & B2X
Split user Bitcoin margin long positions into BTC & B2X long positions
Split user Bitcoin margin short positions into BTC & B2X short positions
Bitcoin lenders are due back BTC & B2X
Bitcoin borrowers owe BTC & B2X  

Note: It is also possible to have a different policy with respect to Bitcoin lending and Bitcoin margin positions, which is not illustrated in the above chart.

An interesting investment strategy to engage in before the B2X fork could be to open long Bitcoin positions with exchanges with policy C and D, and potentially open short Bitcoin positions on exchanges with policies A, B and C.  In theory, this should allow you to get B2X tokens for free.

One may think that policy C may seem a slightly inappropriate choice, as it results in an asymmetry. However some exchanges did have a policy similar to this with respect to Bitcoin Cash. The rational for this was that the burden on customers who were short Bitcoin, to go out into the market and buy Bitcoin Cash may have been too high, particularly if the liquidity of Bitcoin Cash was low.

As the B2X fork approaches, we may write a piece summarizing the policies of the main exchanges and how one could engage in this type strategy. Although, if you wait for it to be clearly explained, it could be too late and spreads could have already opened up, reflecting the opportunity. On Bitfinex, B2X is already trading at over $1,000, therefore there could be money to be made by engaging in these types of strategies. Perhaps a good idea, if you really like taking risks, may be to review the policies exchanges took with respect to Bitcoin Cash, to get an idea of what their policies might be with respect to B2X and then open your positions before the policies are officially announced.

ICOs – Part 1 – An Interactive Visualization of the ICO Space

Abstract: In this week’s piece, we look at the world of ICOs.  We present an interactive visualization, that illustrates the entire ICO landscape and the interconnections between people officially affiliated with each ICO.


Interactive ICO team member network map

Source: BitMEX Research, ICO websites, icoHunt own and provided the raw data (Do not use the data without their permission)

Network map notes

  • Blue circles represent ICOs
  • Yellow circles represent ICO team members, advisers or backers
  • Circle size is weighted by number of links to other ICOs
  • Data includes around 500 ICOs &  around 4,900 team members
  • We apologies for any errors, omissions or spelling mistakes

ICO network map viewing tips

  • Modify the number of nodes to around 1,000
  • Alternatively, start with 5 nodes, then gradually press the up arrow to add 10 nodes at a time
  • Displaying all 5,388 nodes may use up significant amounts of system resources
  • An illustration of the type of visualization which can be produced is provided here


The negative implications of the extent of ICO team member interconnections

The above network visualization illustrates the high degree of interconnections between the ICOs, with a few dominant individuals involved in many projects.  Skeptics may therefore argue that it is not feasible for any one individual to be involved in this many ICOs, and this is an indication of misleading marketing, bad practice or inadequate disclosure.


Comparing ICOs & IPOs

Many of the individuals listed in multiple ICOs are often listed as a backer or adviser, rather than as an individual involved in day to day management.  In some ways they may be more of an ICO “sponsor” than someone actively involved in project development.  For example, as the chart below illustrates, the concentration of individuals involved in multiple ICOs compares reasonably well to financial advisers or book-runners involved in IPOs.

Vitalik Buterin is the most prolific person when it comes to ICO involvement.  Our analysis shows he has involvement in at least 13 projects, around 2.5% of the total.  Although if you think of Vitalik’s role as more of a sponsor, then Vitalik has a lower market share than Citibank when it comes to IPOs, which has been the lead bookrunner in 3.7% of the IPOs in the last 12 months.


Market share of advisers (Top 20) – ICO/IPO comparison

Source: BitMEX Research, icoHunt, ICO websites, Bloomberg IPO database
Notes:  The above dataset is from around 520 ICOs and 3,266 IPOs which have occurred in the last 12 months.  The market share is based on the number of offerings, not value of the offerings.


The role of a lead bookrunner in an equity IPO, may be to do the following:

  • Underwrite the offering,
  • Ensure the offering complies with regulations,
  • Ensure price stability and liquidity after the stock lists,
  • Prepare marketing materials for the IPO,
  • Bring management of the company on a tour to see potential investors in a roadshow,
  • Use its investment sales staff to promote the offering to its clients,
  • Assist in the allocation of the shares to the offering participants.

In many ways these responsibilities are similar to the roles of some of the most prominent individuals in our ICO analysis above, certainly with respect to marketing and promotion. However, in the world of ICOs the role of the “sponsor” is a lot more murky.  The individual in question is sometimes said to be the person who originated the idea, is involved in management or even the development of the product.  The ICO market therefore lacks clarity with respect to the roles and responsibilities of the individuals involved.


The initial allocation

The initial allocation of shares during an IPO is often an opaque process.  Typically the bookrunner/adviser decides the allocation on a discretionary basis, with the largest institutional and hedge fund clients favored first, rather than a fair random process or auction driven mechanism.  It is somewhat surprising that financial regulators have not clamped down on this practice. Although a counter argument to regulation is that listing companies should have the freedom to decide who their new shareholders are, if their are multiple bidders.  Unfortunately the bookrunner/adviser seems to have a significant influence on the allocation process.

Despite the opportunity to improve on what IPOs do in this area, the initial allocation process when it comes to ICOs does not appear to be much better.  The initial allocation of ICOs is also a murky process, but in some ways even worse than for IPOs, since earlier or better connected investors sometimes get lower prices. The allocation process can be based on who is friends with the people in the team or other industry connections.  Therefore on balance, despite poor practice, IPOs appear to have a somewhat fairer process when it comes to allocation than ICOs.

Although one could cynically say there is a key difference in the offering process between ICOs and IPOs, this time we are the ones in control.

ICOs – Part 2 – Investment Returns

Abstract: Although the reliability of the data is limited, in this piece we look at the investment returns one would have achieved  by investing in ICOs.  We compare the returns against the amount each ICO raised and look at the historic investment performance of the most prolific ICO advisers.


ICO investor returns (log scale)

Source: BitMEX Research, Coinmarketcap, ICO websites, Reddit comments
Notes: A return of 10x means the investor made 10x their money.  Prices up to 25th September 2017


The integrity of the data

  • In general the quality of pricing data for ICOs is poor, we are likely to have made significant errors in many cases.  We apologies for any inaccuracies.
  • ICOs often have particular circumstances and conditions, for example multiple ICO rounds or offerings in different currencies.  Unfortunately due to a lack of transparency and a high degree of complexity, in some cases, we have not been able to account for all of these particularities in a consistent basis.
  • In some circumstances we may have been unable to determine the ICO price and therefore may have used the price on the first day of trading.
  • There may be some selection bias in our analysis, for example we may be more likely to have investigated the investment performance of ICOs which subsequently turned out to be successful.
  • Unfortunately, in a small number of cases we obtained the ICO price from Reddit comments, due to the lack an alternative source.


ICO investor returns compared to the amount raised in the offering (log scale)

Source: BitMEX Research, Coinmarketcap, ICO websites, Reddit comments, Coindesk
Notes: A return of 10x means the investor made 10x their money


In the above chart we compare the amount raised in the offering with the investment returns of the ICOs.  From looking at the chart, it appears that there is a slight negative correlation, although the correlation is not strong enough to pass rigorous statistical requirements.


ICO investor returns of for the top 20 ICO advisers (log scale)

Source: BitMEX Research, Coinmarketcap, ICO websites, Reddit comments, ICO Hunt
Notes: The numbers in brackets represent the number of ICOs we were able to obtain return data for compared to the total number of ICOs the adviser is involved in


Finally we look at the investment returns of the top 20 most prolific individuals in the ICO space, by number of ICOs.  In many cases, the ICOs occurred very recently and therefore pricing data is not yet available.

The SegWit Transaction Capacity Increase – Part 2 – The First Month After Activation

Abstract: SegWit adoption has grown at a steady pace since activation one month ago and this has surprisingly coincided with a large reduction in transaction fees.  Although we think this drop in fees may be a coincidence, as the absolute level of SegWit adoption in the first month is reasonably low


The first month of SegWit data

As the chart below (figure 1) shows, in orange, SegWit adoption has increased steadily since activation, in a somewhat linear fashion.  Adoption has now reached around 4.5% of transactions.  This should represent a small theoretical c1.6% capacity increase for the network, after just one month.

As figure 1 illustrates, the activation of SegWit does appear to have coincided with a sharp reduction in average transaction fees.


Figure 1 – Transaction fees compared to SegWit adoption

Source: BitMEX Research
Note: Data based on daily averages, up to 20th September 2017


As figure 2 below shows, Bitcoin denominated transaction fees do somewhat correlate with the Bitcoin price and Bitcoin price volatility.  This relationship is probably due to higher Bitcoin prices causing increased levels of transaction demand. The first month since SegWit activation was a period of large Bitcoin price increases and high price volatility, which makes the large transaction fee drop appear somewhat unusual.  The drop is transaction fees appears to start during the period when the price was still rapidly increasing.


Figure 2 – Transaction fees compared to Bitcoin price

Source: BitMEX Research
Note: Data based on daily averages, up to 20th September 2017


As figure 3 below illustrates, SegWit usage is still low compared to non SegWit usage.  Although markets and the price elasticity of demand can be very difficult things to quantify and it is possible that the small 1.6% implied capacity increase caused a reduction in transaction fees, we believe this is unlikely.  Therefore the sharp reduction in transaction fees is likely to be an odd coincidence.

It is therefore too early to tell if SegWit was successful in increasing transaction throughput and lowering transaction fees.


Figure 3 – SegWit adoption compared to old format transactions

Source: BitMEX Research
Note: Data based on daily averages, up to 20th September 2017


An alternative explanation for the fee reduction is that not all of the historic demand was “genuine”.  A spammer could have been attempting to inflate transaction demand and increase transaction fees.  Perhaps the activation of SegWit or launch of Bitcoin Cash meant this spammer now has less reason to continue the expensive attack.  Although there is no direct evidence to support this theory.

The SegWit Transaction Capacity Increase – Part 1

Abstract: SegWit replaces the old 1MB blocksize limit, with a new more complicated 4 million unit blockweight limit.  This should double the transaction throughput of the network, but only after wallets upgrade to send new SegWit style transactions.  This piece analyzes how the network throughput could increase, as users gradually upgrade to SegWit


Overview of the capacity increase

Segregated Witness is an upgrade to the Bitcoin protocol, which may increase the level of transaction throughput.  In order to benefit from increased transaction throughput, users need to upgrade to the new transaction format.  This may make the throughput increase slow and gradual.  Although those that do upgrade, may experience the benefit of a reduction in transaction fees reasonably quickly.

Higher transaction throughput is achieved by replacing the old blocksize limit, with a new blockweight limit.  The old 1MB blocksize limit (shown below), has been removed from the code:


The new, more complex, blockweight limit is determined by the following formula:

4 * non witness data (in bytes) + witness data (in bytes) < 4 million units

This new limit should approximately double the transaction throughput of the network, once users and businesses upgrade to the new transaction format. Miners trying to maximize transaction fee revenue, will attempt to include as many transactions as they can, with blockweight being the new limiting factor.


What is the witness data?

The witness data is the digital signature, authorizing the spend of the payment.  The below diagrams attempt to illustrate the transaction structure before and after SegWit.  The witness data can take up around 54% of the data in a typical transaction and is represented in the blue rectangles in the 2nd diagram below.

The diagrams below illustrate two transactions, with transaction 2 spending the payment received in transaction 1.


Pre SegWit – Illustrative transaction structure (percentage data usage)


SegWit – Illustrative transaction structure (percentage data usage)


Fixing third party transaction malleability

It is often mentioned that SegWit “splits” the witness and non witness data up, into two separate places.  This is not really true, the witness and non witness data still exists together.  As the diagram above illustrates with the green arrows, the key difference with SegWit is that the witness data is no longer needed when calculating the transaction ID (TXID), which is later used to spend the funds in the next transaction.

This is the solution to the third party transaction malleability problem.  If the signature changes, from one valid signature to another valid signature, the transaction ID remains the same.  Therefore it is reasonably safe for transaction 2 to be created and broadcast to the network, before transaction 1 has been confirmed in the blockchain.  This should improve the user experience, as there is now less need for users to wait for confirmations, which can be slow.


The throughput benefits of upgrading to SegWit style transactions for individual users

Assuming the witness data makes up 54% of transaction space, when one upgrades to SegWit, transactions should take up around 41% less blockweight that old style transactions, as the following formula illustrates:



Therefore, assuming fee market conditions are unchanged (i.e. transaction fee levels are the same as before SegWit), by upgrading to SegWit, the user should benefit by paying 41% lower fees than before.  This benefit occurs before other users upgrade to SegWit.

Users who do not upgrade to SegWit, should pay the same fee levels as before.  However, there is a theoretical “side benefit” which they may experience, caused by other users freeing up space by using SegWit.  This may improve fee market conditions and result in lower fees, even for users that have not upgraded to SegWit.  Estimating the magnitude of this so called “side benefit” is almost impossible, since it depends on fee market conditions.


The throughput benefits of SegWit on the entire network

Below we analyse what the network throughput dynamics could look like, if users and businesses decide to opt in to SegWit style transactions, such that there are high adoption levels.

The chart below compares how the blockweight and blocksize could vary, as the amount of witness data in a block increases.  The left hand side of the chart illustrates low levels of SegWit adoption (where we are now), whilst as we move to the right, higher adoption rates are assumed.

If we achieve high adoption rates, the blocksize may be around 2MB, consisting of 1.3MB of witness data and 0.7MB of non witness data.  The block weight would be 4 million units.  This would mean the throughput of the network increases by around 100%.


Blockspace dynamics assuming “normal” SegWit usage


Theoretical blocksize problems

Skeptics of SegWit often mention some problems.  A common claim is that:

SegWit enables 4MB blocks, with only 1MB of transactions

Perhaps the people who are making this claim are visualizing something similar to the below chart. This chart is similar to the one above, except instead of assuming “normal” transaction patterns, we assume the block weight remains at 4 million units, while the witness data grows to 4MB.

On the right hand side, the chart shows that a block could theoretically be 4MB in size (containing only witness data), which a spammer could produce for the same fees as producing a 1MB block (containing only non-witness data). Since both blocks contain 4 million units of weight.


Blockspace dynamics when blocks always contain 4 million units of weight


It is true that a spammer could produce such a 4MB block, which would cost the same in fees as a 1MB block.  This is a potential problem.

However, this is not a change to the security properties of the system, because the 4MB block is not cheaper than the 1MB block, it merely costs the same. A spammer can always outbid legitimate users, regardless of SegWit. Indeed an attacker could simply produce 1MB of non-witness data, to compete with “legitimate” users, and this attack vector is no cheaper than before.  SegWit does not and cannot change the security dynamic, that if an attacker wants to outbid users with spam data, they can.

There is still a potential problem here.  There is now a “worst case” block of 4MB, even though the expected capacity benefit may only be slightly more than 2x the old 1MB limit, rather than 4x.

There is some merit to this criticism of SegWit.  However, due to the scaling properties of the witness data, the situation may not be as bad as one would think.  In some ways the witness data has better scalability characteristics than the non-witness data.


We discuss some of these characteristics below:


  1. Signature verification: SegWit mitigates the quadratic hashing of sighash operations problem, by changing the calculation of the transaction hash for signatures so that each byte of a transaction only needs to be hashed at most twice. Therefore a block containing a large amount of witness data, will have a large amount of data with this linear hashing characteristic, reducing potential signature verification times and improving scalability.
  2. The impact on the database on unspent transaction outputs (UTXO): The witness data relates to signatures for transaction inputs. If one wants to create new transaction outputs, this would not be included in the witness data. Therefore a theoretical 4MB block contains no new outputs. A large block block therefore has a neutral or even positive impact on the size of the UTXO set.
  3. Long term storage costs: The worst case 4MB blocks can have a negative impact on long term storage costs.  However, when segregating the witness, the witness is no longer used in the calculation of the the transaction ID (TXID). This means one can calculate the TXIDs, the merkle root hash and block hashes, without the witness data. All one needs is the witness commitment and the block hash can be calculated without the witness data. A wallet can therefore get “proof of work (PoW) assurance” over the security of payments, while discarding the witness data. Although if one wants to verify all the witness data on restart of the node, the witness data must be stored.
  4. Block propagation/latency: In the worst case scenario, these theoretical 4MB blocks will have inferior propagation characteristics. However, technologies such has compact blocks have mitigated some of these risks.  Although the recent apparent crackdown in China may illustrate that this is a potentially serious problem.
  5. Bandwidth: 4MB blocks could increase the bandwidth burden on nodes/wallets, compared to 1MB blocks of non witness data. This is therefore a potential disadvantage to these larger blocks. Although, perhaps some light wallets could only download the witness data for their own transactions, they would therefore never need to download the rest of the witness data, but they could still link the block hashes to the transactions and therefore obtain “PoW assurance” over the payments and wallet balances.

Mining incentives, part 2: Why is China dominant in Bitcoin mining?

Abstract: This piece explores why China appears to dominate the Bitcoin mining industry. Chinese Bitcoin mining could be the inadvertent beneficiary of a massive boom and over-investment in hydro power, linked to aluminum production, in remote regions in China over the last 10 years.

China globally dominates most energy related fields

One question often asked by some in the Bitcoin community is Why are so many mining pools and miners based in China?

The most basic response to this question is: why not? China is not just some random country. It is a global economic powerhouse, dominant in most industrial and energy sectors. For instance:

  • China accounts for roughly 68% of the world’s imports of iron ore.1
  • China produces roughly 54% of the world’s aluminium.2
  • China consumes roughly 50% of the world’s coal.3
  • Of the 60 nuclear plants under construction across the world, 22 are in China, and China has accounted for about 183% of the world’s growth in nuclear-power consumption in the last 10 years (this number exceeds 100% because Japan shut down its reactors following the Great East Japan earthquake of 2011).4
  • China consumes about 30% of the world’s hydro power and accounts for 76% of the world’s growth in hydro-power consumption in the last 10 years3

China is unquestionably already the dominant driver of the global energy industry, and perhaps it is no surprise that China is dominant in Bitcoin mining, an industry that is strongly related to energy generation. Indeed, new, efficient power plants, particularly in the field of hydro power, may partially explain China’s dominance in Bitcoin mining.

Reasons typically cited for Chinese dominance

One of the largest and most discussed advantages for China is the local infrastructure and expertise that enables mining farms to quickly set up as soon as the mining chips are available. This setup process is said to be much slower in other regions. Being able to quickly set up can be crucial, as new more efficient chips are developed all the time. China also has geographical proximity to where the specialist chips are manufactured.

Another alleged advantage is corruption in the power industry in China, with some Bitcoin miners perhaps able to pay power-plant owners for the right to place mining equipment inside or near to the power facilities.

Other factors may also contribute to China’s dominance, but these are not widely discussed within the Bitcoin community. Most notable is China’s surplus captive hydro-power capacity, constructed for the loss-making aluminium industry.

China’s hydro boom

Global hydro-power consumption by region in units of million tonnes of oil equivalent from 2006 to 2016. (Source: BP Statistical Review of World Energy June 2017)

As the chart above shows, China has been the dominant driver in world hydro-power growth in the last 10 years. Hydro power is a good energy source for Bitcoin mining, since the marginal cost of power generation is very low, with most of the costs related to the initial investment. In addition, hydro power is reliable and stable.

The largest of China’s hydro-power facilities under construction is the Baihetan project on the Jinsha River. The facility will have 16 turbines of 3.6 TWh each, producing a total of 57.6 TWh. The project began in 2008 and is expected to be complete in 2021. The Wudongde project on the Jinsha River in the southwest is the most recent large project and will provide 36.7 TWh of capacity. It began construction in 2014, and the first generator is scheduled to be complete in 2018. The whole Wudongde project is expected to be complete in 2020. It will be the sixth-largest hydro-power facility in the world.

As a point of comparison, the Three Gorges Dam, the world’s largest hydro-power project (with the possible exception of the Itaipu Dam in South America), produces around 100 TWh.

The Three Gorges Dam on the Yangtze River in 2009. (Source: Wikipedia)

The projects mentioned above are far too large for Bitcoin mining. Bitcoin’s total electricity consumption is believed to be around 15 TWh.

In January 2017, China’s ministry of water resources issued guidelines on promoting the development of small hydro-power plants. The guidelines outlined plans to develop and grow the small-scale hydro-power industry in China by 2030. Because smaller facilities are more likely to power Bitcoin mines, this could be an interesting development.

Aluminium production

Typically, hydro power is associated with aluminium production. The Bayer process and the Hall-Héroult process, the principal industrial means of refining bauxite to produce aluminium, involve electrolysis and require a massive amount of electrical power. Steel making does not require electrical energy to the same extent.

Energy costs make up a significant proportion of the cost of producing aluminium, perhaps around 30% to 65% of the total, depending on the producer and market conditions.

In the competitive aluminium industry, it’s vital to keep costs as low as possible to remain competitive throughout the economic cycle. Canada is the number-two hydro-power producer in the world, behind China, and much of that country’s hydro power supports Rio Tinto’s (previously known as Alcan) aluminium refining.5 Similar to Bitcoin mining, aluminium smelting will follow the cheapest, most reliable energy source, which is often hydro power.

According to the International Aluminium Institute (IAI), about 80% of Canadian aluminum smelting uses hydro power. In China, only 10% of smelting relies on hydro power, with the remaining 90% powered by coal.6 Yet due to China’s massive scale, China’s 10% consumes roughly the same 50 TWh of hydro power as Canada’s 80%. Power plants constructed specifically for aluminium smelting are often called “captive plants”.

Over-investment in hydro power and the aluminium market cycle

One could argue that China has over-invested in the development of hydro power. In recent years, this investment has slowed down, as most major sites have been identified and explored. Despite this, China still commissioned more new capacity in 2016 than any other country.

Although aluminium is less vital for China’s hydro power infrastructure than it is elsewhere, it is still significant. China invested considerably in aluminium-related hydro power in 2011 to 2013, a period that was followed by significant weakening in the price of aluminum in 2014 and 2015, as the chart below shows. As a result of this price drop, many aluminum smelters became uneconomical and stopped production — and the associated captive hydro-power facilities often suffered from limited local demand for their electricity. Many of the facilities  often had been constructed in remote rural regions whose infrastructure may have been unable to send the power to a major city, due to long distances and a lack of ultra-high-voltage (UHV) electricity transmission infrastructure in China.

Bitcoin mining may be one of the only suitable consumers of the energy in such circumstances. Without UHV lines, much energy is lost to resistance during electrical transmission. Bitcoin mining can, in theory, transfer the equivalent value of the energy to the main cities in the east of China, avoiding loss due to resistance in the lines. Bitcoin miners with access to this cheap electricity from captive plants may demand and receive lower prices of power than they could find in urban areas. Of course, this process only works if there is demand for Bitcoin in the cities, which in the last few years appears to have been true.

As the chart below shows, the price of aluminium has recently recovered, although this is partially due to China’s reduction of aluminium production by shutting down loss-making smelters.7 The Chinese government’s strategy is to reduce subsidies to loss-making heavy industries in order to shift the economy towards technology and consumer areas. This is also part of a wider push to protect the environment and improve the air quality in China.

Aluminum spot price. (Source: London Metal Exchange)

Capacity utilization of aluminium producers

The chart below indicates that China has undergone a massive boom in aluminium production, from almost none of the global market 20 years ago to being globally dominant. Although production growth has slowed in recent years as a result of massive over-investment and significant under-utilisation, as the second chart illustrates. The third chart indicates the scale of the overcapacity in China, with China having the least utilisation apart from the United States, which is also reducing production.

Global aluminium production by country from 1995 to 2016 in million tonnes of aluminium. (Source: BitMEX Research, US Geological Survey)


Surplus aluminium production capacity by region from 1995 to 2016 in million tonnes. (Source: BitMEX Research, US Geological Survey)


Aluminium production capacity utilisation. (Source: BitMEX Research, US Geological Survey)

Canada’s lower structural costs have allowed Rio Tinto to achieve strong aluminium-related net profit margins of around 10%8 in the last three years despite the lower aluminium price, which explains Canada’s high capacity utilisation, approaching 100% in 2016.

Rio Tinto’s modernised Kitimat hydro-powered aluminium smelter in Canada. (Source: Rio Tinto)

The implied hydro-power over-capacity

After analyzing the aluminium production overcapacity data, we multiplied this by the percentage of aluminum smelting powered by hydro power in each region to calculate an approximation of the surplus hydro-power capacity related to aluminium. This may be an overestimate, as the coal-powered smelters may have been disproportionately shut down.

The results, in the chart below, show that China has a huge surplus, despite the fact that only some 10% of its aluminium smelting is powered by hydro power.

Implied under-utilized hydro-power facilities from 2004 to 2016 in million tonnes of aluminium equivalent. (Source: BitMEX Research, US Geological Survey, IAI)

Although the chart shows China has by far the largest surplus of hydro-power capacity related to aluminium, this doesn’t paint the full picture. Much of the non-Chinese surplus capacity is concentrated in North America and Europe near regions of higher population density with stronger UHV transmission-line infrastructure.

A significant amount of China’s energy sources are located in the west of the country while most of China’s population, and resultant energy demand, resides in the east.

China has several UHV transmission-line projects underway, which will deliver energy from rural regions to cities in the east. The State Grid Corporation of China aims to invest US$88 billion on UHV lines between 2009 and 2020. Many lines are due to be complete in the coming years and it will be interesting to see if this has any impact on Bitcoin mining.

Global electricity consumption

Global aluminum production consumes about 804 TWh of electricity, around 3.2% of the total global consumption. Based on the latest estimates, Bitcoin mining consumes around 2.1% of the electricity used in aluminium production.

However, because of its greater flexibility with respect to where it needs the electricity, aluminium electricity is expected to cost far less than 3.2% of the global cost. Similarly, Bitcoin mining is expected to cost even less than 2.1% of what aluminium producers spend on electricity.

Global electricity consumption (TWh) 2016 %
Bitcoin mining* 16.7 0.07%
Aluminium production 804.0 3.22%
Other 24,179.0 96.72%
Total 25,000.0 100%

*Bitcoin annualised projection based on latest figures. (Source: BitMEX Research, BP, IAI, Digiconomist


The confluence of under-utilised hydro power due to over-investment in subsidized loss-making aluminium production, the construction of these facilities in the remote west of the country, and a lack of UHV transmission infrastructure have contributed to China’s Bitcoin-mining success.

We appreciate that the situation can be looked at from many angles and perhaps other explanations have more direct relevance to Bitcoin mining.


World’s Top Exports
US Geological Survey
BP Statistical Review of World Energy June 2017
Nuclear Energy Institute
Rio Tinto
International Aluminium Institute
Bloomberg: “China’s Metals Curb Plan Seen Risking Shortages in Biggest User”
Rio Tinto Annual Report (p. 199)
Chalco 2015 Annual Report (p. 11)
10 Reuters: “Chinalco plans shutdown of biggest aluminum smelter”

Mining Incentives – Part 1 – The Economics of the Difficulty Adjustment

Abstract: This piece contrasts mining economics between Bitcoin and traditional resource mining. We look at how the difficulty adjustment can impact profitability in the mining industry and some potentially perverse incentives.


Bitcoin’s Mining Difficulty Adjustment

Mining is the random process by which new Bitcoin blocks are found, such that transactions are confirmed. This is a necessarily competitive and energy intensive process.  In order to ensure a smooth and reliable network, every two weeks, based on how many blocks were mined in the period, the mining difficulty adjusts.  There is an average target interval between blocks of 10 minutes.

A two week period is split into 2016 sections of 10 minutes.

  • If more than 2016 blocks were mined in a two week period, mining becomes more difficult, such that if the hashrate remains constant, blocks are expected to be found every 10 minutes in the next two week period.
  • If fewer than 2016 blocks were mined in a two week period, mining becomes less difficult, such that if the hashrate remains constant, blocks are expected to be found every 10 minutes in the next two week period.
  • The maximum adjustment in any one period is a factor of 4 (i.e. a range of 25% to 400% of the old difficulty).

The below chart shows the difficulty adjustments (red line) and the calculated rolling two week hashrate estimate (green line), over the last 2 months.



Mining Equilibrium and the Misconception This Means Miners Cannot Make Profits

In theory, the difficulty adjustment keeps the system in check, in an equilibrium position, when external inputs change.  For example consider the following scenario of a sudden increase in the Bitcoin price:

  1. The Bitcoin price increases
  2. Mining profitability increases, since miners are rewarded in Bitcoin
  3. More miners join the network to take advantage of higher profit margins and there are limited mining entry barriers
  4. The network hashrate increases and the average block interval falls below 10 minutes
  5. After a few weeks, the mining difficulty increases and therefore mining profitability decreases
  6. The average block interval increases back up to 10 minutes

The same kind of logic can be applied to an increase in Bitcoin transaction fees or the release of new more efficient mining hardware.  The theory can also be used in reverse, for a sudden Bitcoin price decrease.

Contrary to a popular misconception, this short two week period combined with the theoretical ability of anyone to quickly enter the mining industry, does not mean mining industry profit margins tend to zero every few weeks.  What the difficulty adjustment may mean is that the profit margins of the most marginal miner quickly tend to zero.  However, not all miners are the same, for example some miners may benefit from structurally lower costs, such as lower electricity costs, more scale or an effective maintenance regime.  These more efficient miners may be able to produce stable profits throughout these cycles.

In addition to this, it is no longer true that there are limited barriers to enter the Bitcoin mining business, at a particular cost level. For example perhaps a large capital investment is required, there may be large time lags before the mining farm is active, cheap electricity may be difficult to source and the latest specialist mining chips may be hard to get hold of. Therefore the assumption that there are limited entry barriers may no longer hold anyway.


The Gold Mining Industry as an Analogy

The below chart illustrates total world gold mining production, by mining company.  The x-axis displays the proportion of global gold production by company, while the y-axis is a measure of the cost to produce one troy ounce of gold.  The cost measure is called “All In Sustaining Costs” (AISC), which includes both the marginal cost of production (variable costs) and annual depreciation of the capital investment incurred to construct the mine (fixed costs).  The straight red line represents the spot gold price.


Gold mining industry cost curve


As the chart illustrates, miners on the left hand side are able to generate profits, despite moderate reductions in the spot gold price.  In contrast miners on the far right hand side, can be classified as the most marginal miners.  The profit margins of these most marginal miners does tend to zero over the cycle.

If the gold price increases, over the long term, more gold miners may join the industry on the right hand side of the chart, increasing the supply of gold, eventually causing downward pressure on the price of gold.  This type of equilibrium cannot occur in Bitcoin since the supply is fixed, therefore the difficulty adjustment is an alternative equilibrium mechanism.

Starting up a new gold mine takes many years (currently much longer than in Bitcoin) and the miner may want a large price buffer before deciding to invest in a new mine, therefore Bitcoin’s equilibrium mechanism can be thought of as occurring much faster than in traditional mining.

The grey bar on the far right hand side of the chart shows a miner making a loss.  However, since the AISC includes depreciation, it may be rational for this loss making miner to remain open, to make a contribution to the capital investment initially made, although the original investment decision turned out to be a bad one.

The same logic could be applied to Bitcoin mining, the fixed costs are items such as acquiring the mining equipment and building a mining farm, while the variable costs are items such as electricity bills and maintenance.  Therefore, just like in traditional mining, it could be rational for a loss making Bitcoin miner to remain open, as long as it’s making a contribution to the initial fixed costs.  One could call this a free cash flow (FCF) positive miner. This could help ensure the mining hashrate and Bitcoin network conditions are more stable than they otherwise would be.


How the Mining Adjustment Impacts Industry Economics

When the mining difficulty increases, the cost for the miners to produce one Bitcoin should shift upwards.  Incidentally this is a not necessarily a parallel upward shift in costs.  An upward difficulty adjustment by 10%, should in theory only increase the variable costs needed to produce one Bitcoin by 10%, while leaving the fixed costs, such as depreciation, unchanged.

Each miner may have a different breakdown between fixed and variable costs.  The fixed costs should remain unaffected by a difficulty adjustment, such that miners with a higher proportion of fixed costs, perhaps due to very low electricity charges, may comparatively benefit from an upwards difficulty adjustment, compared to its peers.

This dynamic is different from traditional resource mining (e.g. gold mining). If the gold spot price changes, this neither impacts the variable or fixed costs, of producing one troy ounce of gold.  Therefore in gold mining, the actions of other miners cannot directly impact the costs in any one mine, whilst in Bitcoin difficulty adjustments, driven by the actions of other miners, can directly impact the cost per unit production. If the spot price of gold increases or decreases, all gold miners take a parallel hit with respect to  sales and therefore profits per troy ounce.  This dynamic is  different in Bitcoin mining.


In theory, Some Bitcoin Miners Could Even Profit in Absolute Terms from Falling Bitcoin Prices

We have explained how some Bitcoin miners can comparatively benefit from a lower Bitcoin price, however it may also be possible for some miners to get an absolute benefit from a falling Bitcoin price, in the short term.

If the spot price of Bitcoin falls, initially all miners take a parallel revenue hit per Bitcoin produced, just like in traditional mining. However, should miners on the right hand side of the cost curve leave, depending on the shape of the costs curve, some miners could actually see the absolute level of their profits increase.  In some ways this could make the Bitcoin mining industry earnings more resilient to price crashes, but in other ways it could produce perverse incentives.

As the contrived and perhaps unrealistic cost curve below demonstrates, if the Bitcoin price were to fall to $800 form $1,000, the large chunk of miners with costs just below $1,000 could shut down their machines.  The difficulty could then adjust downwards and miners on the left hand side of the curve could, in theory, increase the absolute level of their profits.  A similar analysis could be conducted, but instead of miners leaving, it could apply to the absence of miners entering the market.


Contrived mining industry cost curve (variable costs)


Therefore, perhaps some miners may benefit from falling Bitcoin prices, in some time periods.  Who knows, we may already be seeing this phenomenon, to some extent, with larger lower cost miners wanting the price to remain low such that new miners do not enter the market.  Although this is probably somewhat unlikely.

The SegWit2x Hardfork – The Supporters and Opponents

Abstract: This piece examines the political landscape around the SegWit2x hardfork. Although some see SegWit2x as a sensible middle ground compromise, others see this kind of bargaining as the worst possible outcome.

What Is SegWit2x?

SegWit2x is a plan to double Bitcoin’s capacity limit. SegWit2x should not be confused with SegWit itself, which was a capacity increase and upgrade, that has already been successfully activated on Bitcoin.

SegWit2x is an increase in the maximum amount of non witness data per block, to 2MB from 1MB (or an increase in the new block weight limit to 8M from 4M).  This should result in 2MB of effective capacity if users do not upgrade to the new SegWit style transactions or 4MB of effective capacity if SegWit is used.  

Unlike SegWit, the SegWit2x upgrade is an incompatible change to Bitcoin’s consensus rules, known as a hardfork, which means it may result in the launch of a new coin.  Bitcoin holders prior to the fork will receive both “original Bitcoin” and “SegWit2x coin” after the split.  In many ways, this is very similar to the recent Bitcoin Cash hardfork, both with respect to the capacity increase and the launch of a new coin.  

The hardfork is expected to occur on around Tuesday 21st November 2017.

The Supporters of SegWit2x

The SegWit2x proposal arose out of the New York Agreement (NYA), which took place in May 2017.  There are 56 signatories to the agreement, including large payment processing companies like BitPay and Coinbase. Almost all of the largest Bitcoin mining pools signed the agreement.  Miners can show their political support for the agreement by adding an “NYA” tag into the Bitcoin blocks they produce.  In the last three weeks, around 94% of blocks flagged support for the NYA.

However, critics of the NYA could point out that 33 of the 56 signatories, around 59%, are portfolio companies in the Digital Currency Group (DCG), an investment company run by the principal instigator of the NYA.  Therefore one may argue the signatories do not represent widespread support across industry.  Although some DCG portfolio companies did not sign, indirectly indicating that portfolio group companies were not forced to sign.

Please see the full list of signatories below:

# Full list of NYA signatories
Digital Currency Group Portfolio (DCG) Companies
1 Abra (United States)
2 bitFlyer (Japan)
3 BitPay (United States)
4 BitPesa (Kenya)
5 BitOasis (United Arab Emirates)
6 Bitso (Mexico)
7 Bitwala (Germany)
Abandoned NYA
8 Blockchain (UK)
9 Bloq (United States)
10 BTCC (China)
11 Circle (United States)
12 Civic (United States)
13 Coinbase (United States)
14 (Phillipines)
15 CryptoFacilities (UK) Abandoned NYA
16 Digital Currency Group (United States)

Principal agreement organiser

17 Filament (United States)
18 Genesis Global Trading (United States)

Wholly owned by DCG

19 Grayscale Investments (United States)

Wholly owned by DCG

20 Korbit (South Korea)
21 Luno (Singapore)
22 MONI (Finland)
23 Netki (United States)
24 OB1 (United States)
25 Purse (United States)
26 Safello (Sweden)
27 ShapeShift (Switzerland)
28 surBTC (Chile) Abandoned NYA
29 Unocoin (India)
30 Vaultoro (Germany) Abandoned NYA
31 Veem (United States)
32 Xapo (United States)
33 Yours (United States)
Bitcoin Mining Pools 6 month hashrate share
34 1 Hash (China) 2.8%
35 BitClub Network (Hong Kong) 3.5%
36 (St. Kitts & Nevis) 2.3%
37 Bitfury (United States) 7.1%
38 Bitmain (China) (Antpool) 17.5%
39 (China) 10.7%
10 BTCC (Also a DGC Company) 7.8%
40 BTC.TOP (China) 10.7%
41 F2Pool (China) 9.5%
Abandoned NYA
42 ViaBTC (China) 5.3%
Share of hashrate 77.2%67.7%
Other Companies
43 ANX (Hong Kong)
44 /Chandler Guo (China)
45 Bitex (Argentina)
46 (China)
47 (China)
48 Decentral (Canada)
49 Genesis Mining (Hong Kong)
50 GoCoin (Isle of Man)
51 Jaxx (Canada)
52 Ripio (Argentina)
53 SFOX (United States)
54 Wayniloans (Argentina)
Abandoned NYA
55 Guy Corem (Israel)
CEO of Genesis Mining (49)
56 Gavin Andresen (United States)

Signatories Reneging on Agreements

Signing letters indicating support for a hardfork, is not necessarily a new phenomenon, indeed five of the major signatories of the NYA have already backed out of a previous hardfork agreement.  BitPay, Coinbase, Circle, Xapo and, all signatories of the NYA, signed an earlier letter in August 2015, where the companies committed to a hardfork attempt, only to eventually renege on that agreement.  The infrastructure company BitGo also signed the August 2015 letter, but did not sign the NYA, although initially it was erroneously included as a signatory.

Three of the NYA signatories appear to have backed out of the agreement:

  • On 22nd August 2017, one of the signatories, Bitwala, announced that they would not follow the agreement.
  • On 31st August, one of the mining pools who signed the agreement, F2Pool, announced their intention not to support SegWit2x.  Although as it stands, this pool still has the “NYA” flag in its blocks, but says it plans to remove the flag next time it restarts its servers.
  • On 1st September, the CEO and founder of another company who apparently signed the NYA, Wayniloans, Tweeted that they never agreed to all of the NYA and that the agreement changed after Wayniloans signed it.

The Opponents of SegWit2x

Although many businesses agreed to support SegWit2x, it does not appear as if supporters of the agreement have sought the support of users or Bitcoin investors.  Many seem to feel that there was a lack of user engagement in the NYA process or even the pretense of requiring user support.  SegWit2x was therefore seen as more of an ultimatum, than a proposal, by some.  Many users appear to feel that SegWit2x is trying to force them to support the new coin, rather than taking a more friendly approach of inviting them to opt in.  Bitcoin users typically like to feel empowered, which is why many of them initially became involved in Bitcoin, rather than using the traditional financial system. Therefore SegWit2x may alienate some users and investors, who are unlikely to want to follow the new coin.

The majority of Bitcoin developers oppose the NYA.  When asked, almost all of the developers of the Bitcoin Core project  indicated they do not support it.  The Bitcoin Core project will also not implement SegWit2x, despite apparently needing to correct misleading statements indicating that Bitcoin Core supports SegWit2x.  

The new version of Bitcoin Core, 0.15, will attempt to ban SegWit2x peers.  Some people have misinterpreted this as an attempted political move against SegWit2x, by partitioning it from the network early, however this is not the case.  Core nodes banning SegWit2x nodes is beneficial for technical reasons.  Since Core will not be following the SegWit2x chain anyway, it is actually helpful not to connect to SegWit2x nodes, so that SegWit2x clients have better connectivity to other peers who do follow the SegWit2x chain.  Prior to the split, non upgraded Core nodes will act as a bridge between SegWit2x clients and Core 0.15 clients, such that no early network partition occurs.

As for businesses and exchanges who did not sign the NYA, most are yet to comment.  However, some major exchanges and trading platforms, including BitMEX, have published standards about minimum technical safety requirements necessary for a hardfork coin to be supported.  As it stands, the SegWit2x proposal does not seem to meet all of these requirements, therefore many exchanges may be unable to even list the new coin, let alone support it as an upgrade to Bitcoin.  However, there may be time for SegWit2x to fix these issues before the hardfork date.

The Competing Narrative Over Part 1 of the NYA

There were two parts to the NYA, part 1 was to activate SegWit, while part 2 was to do a hardfork within six months.  SegWit has recently been activated on Bitcoin.  Although there is no objectively true answer to this question, people on either side of the debate have been arguing which political circumstances caused the successful activation of SegWit.

Supporters of the NYA argue that since SegWit has activated, part 1 was successful and that therefore there is positive momentum behind the overall SegWit2x proposal.  Others argue that this is not the case and that the activation of SegWit was primarily caused by a User Activated Soft Fork (UASF) .

The SegWit activation client was originally released on 27th October 2016 and miners had not activated the proposal by failing to indicate they had upgraded, for almost 10 months.  A UASF client, called BIP148, was being promoted in some sections of the Bitcoin community.  This UASF client made it mandatory for miners to flag support for SegWit by 1st August 2017, or a new coin would be launched.  The SegWit2x client adopted BIP91, another proposal which also made flagging support for SegWit mandatory.  BIP91 activated on 21st July 2017, which then required all miners to flag support for SegWit on 26th July 2017, just 5 days before the 1st August deadline.  

Detractors of the NYA use these timings to push the narrative that the UASF brought about the successful activation of SegWit, not part 1 of the NYA.  

However, this issue only matters from a political point of view and has no technical significance.

The Impact of Bitcoin Cash

Many may also argue that the launch of Bitcoin Cash also fundamentally changed the politics of the situation.  In our view, Bitcoin Cash has made SegWit2x slightly less likely to succeed.

Firstly Bitcoin Cash delivered almost all of what the “larger blockers” wanted, namely it delivered a hardfork to larger blocks.  Many therefore may not see the need for yet another hardfork, which does almost exactly the same thing anyway.   

In addition to this, Bitcoin Cash has some  technical advantages over SegWit2x, with better safety features included in the hardfork and better scaling technology.  Bitcoin Cash delivers many of the technical benefits of SegWit, as we explained in our earlier piece on the subject.

The Political Landscape

Unfortunately, the Bitcoin community has become increasingly polarized into different camps, illustrated in the diagram below:

Illustrative diagram of the “blocksize debate” political landscape

Group 1 Group 2 Group 3
Group “Name” Robust consensus rules supporters. AKA “Small Blockers” The “large blockers” The free riders
Approximate percentage of the community by number of users (illustrative figures) 5% 5% 90%
Group characteristics Passionate, ideologically driven, tenacious, determined, committed, dedicated, patient and vocal Passionate, ideologically driven, tenacious, determined, committed, dedicated, impatient and vocal Passive, pragmatic, calm and quiet
Philosophy This group wants to ensure that any hardfork (with the new coin becoming known as Bitcoin) only occurs with widespread agreement across the entire community  and in a non disruptive way.  This group wants to prioritize user “financial sovereignty” This group supports a much larger blocksize limit and would like to prevent full blocks.  This group wants to prioritize the user experience and growth. This group doesn’t mind which strategy is used to increase the capacity of the system.  They want the coin to increase in value.  This group wants to financially benefit from the ingenuity, hard work and passion of the other two groups.
Favoured coin The current Bitcoin (e.g. not SegWit2x) Bitcoin Cash Whatever coin wins or SegWit2x
Typical members Developers, Early adopters, “Digital gold” advocates/investors, Cypherpunks, System engineers, Academics Early adopters, Merchant processors, People who want the system to compete with VISA, Front end developers Recent adopters, Exchanges, Speculators

For a coin to succeed a passionate grassroots movement, willing to promote and defend the system is vital.  Abandoning both of these passionate groups (group 1 and group 2), may not be a smart move for the free riders (group 3).  If the most enthusiastic and ingenious Bitcoin supporters are abandoned, the free riders may eventually find that they are not left with anything useful at all.

Bitcoin is an ambitious project, aiming to take on some of the most powerful financial systems in the world.  The Bitcoin industry should not underestimate the challenges and threats that will emerge.  It is naive to think that industry can make this project succeed alone.  Although they may not always be the easiest people to deal with, the passionate, tenacious and dedicated early adopters, investors and developers are a necessity.  A “forced compromise”, like SegWit2x, which undermines what many want from Bitcoin, may either make some of these people leave or at least lose the passion and fighting spirit, which may be needed in the long and bumpy road ahead.

SegWit2x as a moderate political compromise

At first glance, SegWit2x seems like a sensible middle ground political compromise, with both sides making some sacrifices.  One may think a middle ground compromise coin makes some sense and some large corporates appear to have rallied around this idea.

However, many are interested in Bitcoin precisely because they see it as being resilient against these kind of political compromises, which are already a common occurrence in the traditional financial system.  Rather than being a moderate compromise, these group 1 people see the NYA as the worst case scenario, precisely because it’s a forced political compromise.  However, being against this kind of compromise may be an unrealistic expectation and the real world is full of compromises.  The group 1 people may simply be too extreme or unreasonable.

It is certainly true that resilience against political compromise is an unusual and weird characteristic, not common in the business world.  However, to many this is what makes Bitcoin fundamentally very different and unique compared to other forms of electronic money and other systems.  And if we do not think Bitcoin has dynamics that are fundamentally different to typical business scenarios, then why bother with Bitcoin at all?  Afterall the US Dollar appears to work pretty well.  Perhaps what this comes down to is how ambitious one is for Bitcoin and how unique one thinks it can be.

As Satoshi said when responding to someone claiming that Bitcoin cannot solve political problems in cryptography:

Yes, but we can win a major battle in the arms race and gain a new territory of freedom for several years

Perhaps to some people, the success of SegWit2x could mean these several years are up. But we do not expect all of these people to give up without a fight.

Overview of the covert AsicBoost allegation

Diagram of the Merkle trees inside a Bitcoin block. (Source: BitMEX Research)

AsicBoost is a way to reduce the amount of work a mining ASIC must do in order to compute a hashing attempt for Bitcoin’s proof of work (PoW). SHA256, which is the hashing algorithm used for Bitcoin’s PoW, splits the data into 64-byte chunks before the computations occur. The Bitcoin block header is 80 bytes in size and so is split between two chunks — chunk 1 and chunk 2 in the image above. AsicBoost keeps the value of one of the chunks the same over multiple hashing attempts. This requires the miner to do only partial work for this chunk, for multiple hashing attempts.

The AsicBoost methodology was first formally described in a paper published in March 2016.

It is not known whether or not miners are actually using this method to mine Bitcoin. We do not have sufficient evidence to conclude one way or the other. The purpose of this piece is to outline the theory of covert AsicBoost and this should be considered merely an illustrative overview of covert AsicBoost, not an accurate or detailed analysis. We have overlooked some of the complexities.

”Normal” hashing

When a Bitcoin miner hashes, some data in the block header is changed for each attempt, to make the hash different. The string that changes is called the “nonce”.

The miner varies the 4-byte nonce, which is located in the block header (on the top right of the image above), for each hashing attempt. Once all the nonces have been exhausted, the extra nonce, which is located in the coinbase transaction (bottom left of the image) is altered, to provide extra entropy. Changing the coinbase transaction alters the Merkle root hash, which affects both chunk 1 and chunk 2, such that work is required on each chunk of the block header to calculate the PoW.

Satoshi’s original Bitcoin design contained the extra nonce, contrary to popular myth that it was added later to provide extra entropy to miners.

Overt AsicBoost

Instead of changing the extra nonce, a miner using AsicBoost alters the 4-byte version field in the block header (top left of the image). This implies that chunk 2 remains unchanged for multiple hashing attempts and work is therefore saved. If this methodology is used, changes in the version bits are visible to everyone, which makes this AsicBoost methodology easily detectable.

Covert AsicBoost

Covert AsicBoost aims to keep the last 4 bytes of the Merkle root hash the same while generating entropy by changing the first 28 bytes of the Merkle root hash. Therefore, chunk 2 remains unchanged for multiple hashing attempts and work is saved.

However, finding multiple Merkle root hashes where the last 4 bytes collide is not easy. The only way to do this is to calculate many hashes with brute force.

Miners must look for one collision of 4 bytes, which is 32 bits — so this collision is one of 232 hashing attempts. We take the square root this due to the birthday paradox to arrive at a  n approximate expected total of 65,536 attempts.

Each attempt may require changing the extra nonce to generate a new Merkle root hash.  However, due to the structure of the Merkle tree, this would require even more additional hashing. If we change the extra nonce, a new is hash is required for each row of the Merkle tree. For a large Bitcoin block with perhaps 10 or more rows, this is a lot of extra work and  is an inefficient process.

The following methodologies may make this process more efficient:

  • Option 1 is to produce empty or smaller blocks. This simply reduces the size of the Merkle tree and therefore requires fewer hashing operations to generate a different Merkle root hash. The extra nonce can therefore vary in the normal way to produce more Merkle root hashes.
  • Option 2  is shuffling the branches of the right side of the Merkle tree. The red circles in the explanatory image show this at work. Shuffling the top right side of the Merkle tree generates a new Merkle root hash with only two extra hashes, regardless of the size of the Merkle tree. This may be possible, but restrictions such as transaction ordering might prevent such shuffling. If one transaction spends the output of another transaction in the same block, the spender must be to the right side of it.
  • Option 3 is shuffling the position of transactions on the right side of the tree, or swapping transactions.

The above methodologies could be combined.

Impact of the Segregated Witness upgrade proposal

As shown in blue in the explanatory image, the Segregated Witness (SegWit) proposal gives the miner the option to use SegWit and add a second Merkle tree to the block, with the Merkle root hash of this tree inside the coinbase transaction.

This second Merkle tree must have the same structure as the main Merkle tree; the difference is that transaction inputs redeemed by segregating the witness have their signatures (as well as the other transaction data) in the second Merkle tree. Only the version, inputs, outputs, and locktime of the transactions remain in the main Merkle tree. Since the transaction structure must be the same, any alterations to the main Merkle tree, such as shuffling branches, must also be reflected in the second Merkle tree, which therefore impacts the witness commitment. Any efficiency gains from option 2 or option 3 above are lost, rendering covert AsicBoost inefficient.

After the SegWit upgrade, a miner could still use covert AsicBoost, but only by using option 1 or excluding the optional witness commitment. Doing either of these things too often may seem suspicious, yet the miner who is using AsicBoost might like to hide its usage to maintain a secret competitive advantage.

Alleged evidence of the usage of covert ASICBOOST

There have been accusations that a large mining pool may be using covert AsicBoost, making this pool more profitable than other miners.  Evidence for this claim is said to include the following:

    • The particular pool in question produces a much higher proportion of empty or smaller blocks than its mining peers (we will publish research on this in the coming weeks).
    • The ability to do AsicBoost is built inside the company’s hardware products. This ability has been available for over a year and may be costly to include. One could argue that this investment would be wasted if AsicBoost is not being used. However, this hardware-based evidence does not point in particular to covert AsicBoost and could also be used for overt AsicBoost.
    • The pool in question is said to own patents related to AsicBoost, which if true is  circumstantial evidence.
    • Further circumstantial evidence is the company’s desire to prevent SegWit being activated on Bitcoin. Even if this is true, this may however only be part of the reason for attempting to prevent the activation of SegWit.

Although it is possible that the allegations are true, the evidence is not conclusive.

Bitcoin Cash: Potential Price Implications of Investment Flow Data

The launch of Bitcoin Cash

On 1st August 2017, a new coin was launched, Bitcoin Cash (BCH).  This was a chain split from the Bitcoin (BTC) chain.  Therefore, every Bitcoin holder at the time of the split, received both Bitcoin Cash and got to keep their original Bitcoin holdings.  For example those holding 1 coin prior to the split, then received 1 BTC and 1 BCH after the split.


Bitcoin Cash has the following key features:


1. Blocksize limit increase

Bitcoin Cash increased the blocksize limit, to 8MB from 1MB.  This increases the potential transaction throughput of the network, by approximately 4x compared to Bitcoin with SegWit or 8x compared to Bitcoin before SegWit.  This higher capacity is a key advantage and should result in lower transaction fees.


2. New transaction digest algorithm

Bitcoin Cash uses the new transaction digest algorithm for signature verification in BIP143 (part of the SegWit upgrade to Bitcoin, while other parts of the SegWit upgrade were removed).  This upgrade fixes the quadratic scaling of sighash operations bug and improves scalability.  Using the old hashing digest, each time the transaction size doubles, the number of hashing operations required to verify the transaction increases by a factor of 4 (2 squared), however after this fix, hashing scales linearly with respect to transaction size.  The new digest algorithm is compulsory on Bitcoin Cash and is only optional on Bitcoin after SegWit.

The new transaction digest algorithm also ensures that Bitcoin Cash transactions are invalid on Bitcoin and Bitcoin transactions are invalid on Bitcoin Cash, such that issues with users accidentally losing funds are mostly avoided.  For example, if you send Bitcoin, your Bitcoin Cash remains where it is, and vica versa.


3. New mining difficulty adjustment

Bitcoin Cash has a new downward difficulty adjustment, this made the Bitcoin Cash block header invalid according to Bitcoin’s rules. Many mobile wallets therefore need to upgrade to follow the Bitcoin Cash chain, a key safety mechanism ensuring wallets follow the same chain as the one their transactions occur on. In addition to this, the new difficulty adjustment ensures the block time will not be too long, should only a small number of miners decide to switch to Bitcoin Cash.  

The new difficulty adjustment can cause the difficulty to fall too quickly if the gap between blocks is too long.  Unfortunately this can incentivize miners to leave the network so that they can return when the difficulty adjusts and profitability improves.  The impact of this appears to be that the capacity of the network oscillates in a volatile way and this appears to be a major weakness of the coin. In our view this requires a fix.


The coin launch was in many respects born out of a political dispute over the best way forward for Bitcoin, with opposing factions strongly disagreeing with each other on the best strategy.  Therefore, to some extent, there is an ideological competition between the two coins, with each side hoping their chosen coin will be the most successful.  The dispute is difficult to accurately characterize, however the cause of it appears to be that each side looks at the scaling issue with a different key focus.  One side appears to prioritise onchain capacity increases, while the other side appears keen to ensure that upgrades occur in a smooth, safe and voluntarist manner.  Therefore, although there is not necessarily any direct disagreement between the sides, both groups are determined advocates of their favoured solutions, such that their respective coins have the potential to remain viable in the long run, in our view.  It is this determination that can keep coins alive, more than anything else.

There has been some analysis comparing various metrics of the two coins, in particular focusing on mining profitability.  Mining profitability is an area of particular interest given the volatile difficulty adjustment on Bitcoin Cash, causing hashrate swings between the two coins.  This research note will instead focus on the apparent investor flows between the two coins.  Obtaining meaningful data here is more challenging than looking at mining profitability, however the implications of this analysis could prove to be more significant in the long run.  The fact that two potentially opposing groups were each given both tokens and can now trade them against each other, to reflect their ideological objectives, makes the trading and financial market dynamics of this situation somewhat unique and interesting, in our view.


Analysis of investment flows

As the following chart shows, since the chain split, as at 23rd August, 2.8 million coins on Bitcoin Cash have been spent at least once.  This compares to 3.4 million coins being spent at least once on the Bitcoin chain.  We calculated these figures by looking at both blockchains and subtracting the amount of coins which were unspent since the split, from the total coin supply. Bitcoin Cash has c82% of the spend, using this metric, compared to Bitcoin (2.8m/3.4m = c82%).  In contrast, according to data from, Bitcoin Cash only has c4.9% the transaction volume of Bitcoin.  Therefore, since Bitcoin Cash usage is low relative to Bitcoin, it may be somewhat reasonable to assume that most of spend on Bitcoin Cash is “investment flow related”.  One may think this low usage (4.9%) is a negative for Bitcoin Cash, however although transaction volume is an important metric, in our view, this has limited implications on financial markets compared to investment flows, which consider value.


Bitcoin (BTC) vs Bitcoin Cash (BCH) – Value of coins spent at least once since the chain split

Source: BitMEX research, Bitcoin blockchain, Bitcoin Cash blockchain


At the time of the split, there were approximately 16.5 million Bitcoins in existence.  Therefore, as the chart below shows, 17.2% of these have been spent at least once on Bitcoin Cash, while the remaining 82.8% of coins remain unspent.


Bitcoin Cash (BCH) – Proportion of spent vs unspent coins since the chain split

Source: BitMEX research, Bitcoin Cash blockchain


The above data can in theory be used to estimate the behavior of investors.  One could use this information to indirectly estimate the proportion of coin holders that have sold their Bitcoin Cash, which may have implications on the future price.  For example:

  • If the value of Bitcoin Cash coins spent since the split is large (perhaps 40%), many users may have already sold their Bitcoin Cash and therefore the future price outlook could be positive, as the future supply of Bitcoin Cash coins could be limited.
  • If the value of Bitcoin Cash coins spent since the split is low (perhaps 4%), many users may still be yet to sell their Bitcoin Cash and therefore the future price outlook could be negative, as the future supply of Bitcoin Cash could be high.

The below chart may illustrate this to some extent.  On the 2nd day after the split, there was significant spend on the Bitcoin Cash chain (around 520,000 coins spent for the first time since the split), which may have increased the supply of coins on exchanges and then had a negative impact on the price in the following days.  Since around 10th August, the daily spend (for the first time since the split) of Bitcoin Cash has been lower at around the 60,000 level, indicating supply entering the market may be lower.  This could be supporting the Bitcoin Cash price to some extent.


Bitcoin Cash price (% vs Bitcoin) compared to daily Bitcoin Cash spend

Source: BitMEX research, Bitcoin Cash blockchain, Bittrex (for market prices)


In our view, the 17.2% figure could be considered reasonably high, given that perhaps millions of coins may be lost and that the many investors are likely to be lazy and do nothing.  The supply of new Bitcoin Cash coins entering the market may therefore continue this gradual decline.  Therefore, in our view, this analysis indicates the price outlook for Bitcoin Cash is neutral to positive.

Of course the above analysis is an oversimplification and there are many other factors at play.  The 17.2% figure does not imply this value of coins has actually been sold, many investors may have either spent their coins in preparation to sell at a later point or even be splitting coins before doing the opposite trade (Selling Bitcoin and buying Bitcoin Cash).  In addition to this, the analysis says little about the demand for Bitcoin Cash and focuses primarily on supply.  In order for the price of Bitcoin Cash to increase, strong demand is also required.  

There are also many other challenges facing Bitcoin Cash, such as fixing the volatile and unusual difficulty adjustment system and improving the P2P network.  Therefore we are not recommending readers invest in Bitcoin Cash, however if you are considering it, we believe this type of analysis may be useful.