Dear Commissioners:

Thank you very much for the time and attention that you 
gave to my previous letter.  Please allow me to expand
on it and try to answer the specific questions that
you posed.

1. The proposed fund, if approved, would be the first exchange-traded
product available on U.S. markets to hold a digital asset such as
bitcoins, which have neither a physical form (unlike commodities) nor
an issuer that is currently registered with any regulatory body
(unlike securities, futures, or derivatives), and whose fundamental
properties and ownership can, by coordination among a majority of its
network processing power, be changed (unlike any of the above). [ ... ]


1a. What are commenters’ views about the current stability,
resilience, fairness, and efficiency of the markets on which bitcoina
are traded?
  
  Between the sixth and seventh amended filings, the largest BTC-USD
  trading exchange -- Bitfinex, based in Hong Kong -- allegedly
  suffered a security breach and lost 72 million USD in bitcoins (3).
  
  To my knowledge, the Bitfinex management did not report the incident
  to law enforcement authorities, and there has been no audit or
  investigation by any independent entity; only a re-evaluation of
  their security practices by a private company contracted by Bitfinex
  itself (4). To handle the loss, the Bitfinex management unilaterally
  decided to apply a 36% "haircut" on all user accounts, and created
  an unsecured and unbacked internally traded token ("BFX") to
  nominally compensate them for the cut.
  
  Apart from ts implications for the security of bitcoin holdings
  (addressed elsewehere in this letter), the way the incident was
  handled by the management highlights the fact that bitcoin trading
  largely happens in exchanges that are not subject to any of the
  safeguards and regulations that investors expect from stock and
  commodity exchanges.
  
  Because of the lack of regulations and oversight, the largest
  bitcoin exchanges -- that determine the currency's price -- may be
  engaging in many practices that would be illegal in other financial
  markets, such as wash trades, insertion of fictitious entries in
  order books, front-running, and even trading with non-existing
  bitcoins.  There is no clear evidence of such practices, but the
  CEO of one of the largest Chinese exchanges accused his rivals 
  of engaging in them (7).  Anyway, it
  would be surprising if such practices did not occur, since they
  would be easy to implement, impossible to detect, perfectly legal --
  and extremely lucrative.
  
  While US-based exchanges, such as the sponsor's own Gemini, are
  subjected to stricter regulations and auditing for the holding of
  client accounts, the trading itself seems to occur in a regulatory
  vacuum, and seems impossible to audit effectively.
  
  (3) Reuters Technology News 2016-08-03: 
  "Bitcoin worth $72 million stolen from Bitfinex exchange in Hong Kong"
  http://www.reuters.com/article/us-bitfinex-hacked-hongkong-idUSKCN10E0KP
  
  (4) EconoTimes 2016-08-19: 
  "Bitfinex suspends use of BitGo segregated multi-signature wallet solution"
  http://www.econotimes.com/Bitfinex-suspends-use-of-BitGo-segregated-multi-signature-wallet-solution-264659
  
  (5) Reuters Technology News 2016-08-06: 
  "Bitfinex exchange customers to get 36 percent haircut, debt token"
  http://www.reuters.com/article/us-bitfinex-hacked-hongkong-idUSKCN10I06H
  
  (7) Coindesk 2014-01-28: 
  "The Reality of Chinese Bitcoin Trading Volumes"
  http://www.coindesk.com/reality-chinese-trading-volumes/
  
1b. What are commenters’ views on whether an asset with the novel and
unique properties of a bitcoin is an appropriate underlying asset for
a product that will be traded on a national securities exchange?
  
  I believe that the filing, in spite of its overal thoroughness and
  frankness, does not adequately describe the risks and negative
  expectations of the fund.
  
  Indeed the instrument would be unique, for being a derivative of an
  entity that itself has no material existence, no backing asset, no
  accrued revenue, and no responsible entity. But its most worrisome
  "innovation" is that it would be (as far as I know) the first
  investment instrument that is *guaranteed mathematically* to give a
  net total loss to its investors.

  Bitcoin as a negative-sum gambling game
  
  Any investment instrument can be analyzed as a game where investors
  put money in when they buy the asset, and take money out when they
  sell it or receive revenue from it (dividends, interest, rent,
  royalties, etc.).  The profit realized by an investor, up to a
  certain date, is the total money that he took out, minus the total
  money that he put in.  The total profit of the game is the same
  difference, summed over all investors.
  
  To my knowledge, every investment instrument that is publicly traded
  allows at least the possibility of yielding positive gains for all
  its investors.  In the case of a common stock, for example, the
  company is required to provide at IPO convincing arguments that
  its total profits, over (say) the coming 10-20 years, are likely to
  be greater than the value to be collected at the IPO.  In that case,
  the "investment game" above will eventually have a positive total
  profit; and the way profits are distributed will then ensure that
  every investor who held the stock over that entire period will have
  a positive profit.  A company that cannot show at least a
  substantial chance of positive total profit should not be formed in
  the first place, and should be avoided by rational investors.
  
  Some investment instruments are not expected to yield profits, but
  are marketed primarily as hedges against collapse or other
  instruments.  Such instruments must ensure, at the very least, that
  the loss of each investor will be limited to a fraction of the
  invested amount.  A gold fund, for example, provides such assurance,
  because there is a non-speculative final consumption demand for gold
  (for industrial and decorative uses) that is almost certain to
  persist for decades to come, and will ensure a positive sale price
  for the metal.  While it is possible, or even likely, that the total
  profit of gold investors will ultimately be negative, the loss will
  surely not be 100% of the investment.
  
  In the case of bitcoin, however, there is no input of money into the
  "investment game" other than what the investors put in.  There is
  not even a non-investment demand due to final consumption, as there
  is for ordinary commodities.  On the other hand, there is a steadly
  flow of money out of the game that goes to the bitcoin "miners", who
  only sell created bitcoins to investors but never buy them back.

  It has been claimed that the sale of bitcoins to "users" -- people
  who buy bitcoins for use as a currency of commerce, rather than for
  investment or speculation motives -- would be such "final
  consumption".  However, as discussed further on, that demand is
  unlikely to sustain a significant price for bitcoins.  
  
  Moreover, while those "users" derive new value from the use of
  bitcoin as currency, the value is provided by the bitcoin mining
  network, not by the bitcoins themselves; and those users pay for
  that value through transaction fees, that go to the miners and not
  to bitcoin holders.
  
  One should note, in fact, that the value generated by using bitcoins
  for commerce is received mostly by those who spend them quickly;
  whereas the "inflation" losses caused by the divestment of old
  hoards are borne mostly by those users who hold the bitcoins for
  longer periods (weeks or months) before selling or spending them.

  More importantly, the use of bitcoins as currency does not destroy
  them.  Every bitcoin that is bought by a "user" will be re-sold (for
  money, or for goods and services) to another "user", and will
  eventually return to the same market where investors trade.
  
  Thus, for the purpose of understanding the flow of value in bitcoin
  trading, it is not useful to separate the investors from the users.
  The users are just investors who hold the coins for shorter periods
  and trade them for different motivations.  It follows that the sales
  to "users" do not affect the negative-sum character of the "bitcoin
  investment game": the sum total of the losses of all those who buy,
  earn, sell, or spend bitcoins will be greater (by hundreds of
  millions of dollars) than the sum total of all the profits that they
  may achieve.
  
  Therefore, bitcoin is unique among investment instrumens in being a
  matematically guaranteed negative-sum game.  At any time in the
  future, the sum total of the money spent by investors in the
  purchase of bitcoins will be greater than the sum total of the money
  that the investors obtained by selling them.  The difference will be
  the money that miners collected by selling their bitcoins; which at
  present, grows by approximately 1.1 million dollars every day.
  
  Thus, it is mathematically impossible that all investors will obtain
  a positive profit, at any future time.  For every investor in
  bitcoins that will obtain a profit, there must be one or more
  investors who will lose money, whose losses will have provided that
  person's profit.  And there must be many more investors whose losses
  will have provided the miners' revenue.
  
  Indeed, the expected profit of an investor chosen at random, which
  is the same as the average investor profit, will always be negative
  -- by mathematical necessity.

  The bitcoin "investment deficit"
  
  Since there is no other input of money to the "bitcoin investment game"
  other than purchases by investors, and every sale occurs
  simultaneously with a purchase by another investor for the same
  value (plus trading fees and taxes, if any), the "investment
  deficit" -- the difference between total money provided and total
  money withdrawn, by all investors -- is not less than the price paid
  by the last purchaser of each coin, summed over all coins in
  existence (1).
  
  This parameter cannot be determined exacly because bitcoin trades
  are almost all anonymous.  We can however obtain lower and upper
  bounds by considering the minimum and maximum market price that
  could have been used to acquire each created coin (2).  That method
  puts the investment deficit of bitcoin between ?? and ?? billion
  USD.  That is, in order for the current and past bitcoin investors
  to recover their investment, as a whole, a minimum of ?? billion USD
  would have to be provided by new investors.  And that of course
  would not clear the investment deficit; it would only pass it on to
  those new investors.
  
  (1) Some coins, including the first million coins created by Satoshi
  himself, are still in possession of their miners, and thus were
  never purchased.  For those coins, one should consider instead the
  cost of mining, which the miner would like to recover.  However, the
  difficulty adjustment mechanism and the open competition among
  miners are such that the cost of mining is usually a large fraction
  of the market price.  Thus, for the purpose of investment deficit
  estimation, we can seeume that each miner sold his to a fictitious
  investor at some point fater the creation.

  Bitcoin's potential as currency of commerce

  The alleged value of bitcoin would come from its demand for use
  as a currency of internet commerce and other payments, such as 
  international remittances.  If that demand could one day be
  high enough, investing in the ETP could be justified by considering
  users of the currency a separate group from the ETP investors.
  Then, while negative-sum as a whole, the "game" could give
  a positive expected return to ETP investors by pushing all 
  losses to the users.

  However, for that to be possible, the demand for currency use 
  should be high enough to sustain the price above its present level.
  There is no rational argument to support that expectation.

  The use of bitcoin for illegal payments is indeed significant and 
  apparently growing.  Those uses include online gambling (in the US and
  other jurisdistions where it is prohibited), purchase of illegal
  drugs for consumption or distribution (11), purchase of 
  fake identity documents, weapons, and other illegal items,
  cashing gains from stolen credit cards (13), ransomware (14),
  child pornography (15), tax evasion, and more.  

  Bitcoin has also become a popular payment medium demanded by many
  classical frauds, such as prepaid sales of merchandise or services
  that are never delivered, investment in phony enterprises (through
  "Initial Crowfunding Offerings" or ICOs), and ponzi funds (16).

  On the other hand, the use of bitcoin for legal payments does
  not appear to be growing.  There is no reliable data on this
  (or any) sector of the bitcoin economy, since almost all bitcoin
  related companies are privately owned and do not publish financial
  statements.  

  A rare exception was a report by BitPay, one of 
  the largest bitcoin payment processors, that summarized their
  operations for 2014 (18).  They claimed to have processed 
  almost 160 million USD worth of payments in that year.  
  Of these, about 60 million are payments for general goods and
  services; the remainder are payments related to bitcoin mining
  and conversion of bitcoins into other payment media (precious metals
  and gift cards).  Thus the volume of e-commerce through BitPay 
  was only about 170,000 USD per day in that year.

  It must be noted that merchants who accept payment through BitPay 
  do not actually accept bitcoins.  Rather, the customer who chooses
  that payment option is directed to the BitPay server, that receives
  his bitcoins and sends the equivalent in dollars (or other national
  currency) to the merchant.  Still, that conversion can be considered
  a use of bitcoin as currency, for the purposes of estimating the
  "fundamental price".

  Other bitcoin payment services, like Coinbase and Circle, 
  keep custody of the client bitcoins, but not necessarily in the
  form of bitcoins.  When a customer of such a company
  needs to make a "bitcoin" payment to a merchant, the company
  simply sends the dollar equivalent to the merchant, and deducts
  the proper amount of bitcoins from the client's entry in the
  company's private ledger.  Thus, those "bitcoin payments" do not
  really entail use of bitcoin as a currency.  The same caveat can
  be made about various "bitcoin debit cards", that are charged
  with bitcoins but dispense national currencies to merchants.

  The price P of a unit of a currency (in USD, say) is related to the 
  volume V of payments done with it (in USD/day), the time T between
  reuses of the same currency unit (in days), and the number N of
  currency units in circulation, by the equation P = V x T / N.
  For bitcoin, in the distant future when all coins have been mined,
  N will be 21 million BTC; at present, it is about 17 million BTC.
  Using a generous estimate of V = 10 million USD/day for legal
  payments, and T = 17 days, we get P = 10 USD/BTC.  

  The illegal payment volume is probably many times the legal one, but
  even that is clearly not nearly enough to justify the current price
  of over 700 USD/BTC.  And I suppose that there is no need to discuss
  the desirability of an ETP whose success depends on considerable
  expansion of its illegal uses.

  The current bitcoin price is therefore largely speculative, based on 
  hopes of a substantial increase of its use as currency in
  some indeterminate future.  How realistic are those hopes?

  For starters, a price level of 1000 USD/BTC would require more than
  100 times that generous estimate of the payment volume, namely over
  1 billion USD/day.  But there is no reason to expect significant
  growth of adoption beyond present leves; and many reasons to expect
  stagnation, or even the demise of bitcoin.

  For legal payments, bitcoin's only alleged advantage over
  traditional payment systems, such as credit cards, would be its low
  transaction fees -- currently less than 0.20 USD in most cases.
  However, users often need to convert bitcoin from and to national
  currencies, and these conversions may easily add to more than the
  fees of other media.
  
  By and large, users do not seem to find the alleged fee savings
  sufficient compensation for the hassles of using bitcoin, such as
  the need to use special software and the limited acceptance of the
  currency.  BitPay claimed at one time to serve 100,000 merchants
  worldwide, but many of those apparenly have seen so little BTC sales
  volume that they have stopped accepting it (20).

  An analysis of the blockchain shows that there are less than 1.5
  million addresses ("accounts") that contain more than 0.1 BTC
  (presently worth about 70 USD) (22).  Admittedly, that is not the
  number of users.  On one hand, many bitcoin users let companies like
  Coinbase or exchanges keep custody of their coins, and therefore
  would not be counted in that statistic.  On the other hand, bitcoin
  users who handle the coins themselves typically keep them split into
  several separate addresses.  All things considered, however, that
  statistic is strong evidence that there are less than 1.5 million
  active bitcoin users in the world.
  
  Inherent limits to growth

  A more serious obstacle to increased adoption is that the bitcoin
  network is currently saturated and cannot handle more than the
  current traffic (about 230,000 transactions per day on average, or
  about 2.7 transactions per second).  This limit is imposed by a
  parameter -- the maximum size of a block in the chain -- that is
  hard-coded in the current reference implementation. Some
  improvements have been proposed that would increase the capacity by
  70--100% over the next year, but, due to dissentions among
  developers of the code and other players, it is not certain that
  they will be implemented (23).

  Beyond the next year, the possibility of capacity increases are
  uncertain.  The developers who are in control of the reference
  implementation (supported by Blockstream, a company with 70 million
  USD capital) are opposed to further increases in capacity, arguing
  that the network should not attempt to serve everyday e-commerce
  payments, but process only high-value transactions.  

  That camp claims that the bult of the bitcoin traffic should be
  carried out by a separate network, with radically different
  design. There is however no proposal for this "overlay network" that
  is technically and economically viable; and the obstacles are such
  that there is no reasn to believe that such thing is possible.

  While this view is not wholly shared by the community, it is likely
  to prevail in the coming years.  Unable to grow, it is quite likely
  that bitcoin will be superseded by other cryptocurrencies, for both
  legal and illegal payments.

  Inability to adapt

  The ongoing dispute over a relatively minor technical issue --
  the expansion of the network's transaction-processing capacity --
  shows that bitcoin is unlikely to incorporate improvements 
  that other new cryptocurrencies may introduce.  Therefore,
  it is almost certain that, if cryptocurrencies have a long
  enough future, bitcoin will be superseded by some better coin.
 
  Bitcoin was only the first prototype of this radically new form of
  payment system.  The fact that it worked as intended for two years,
  and is still running in some fashion, is proof of the competence of
  its inventor.  But his design did have some fatal flaws, such as the
  capped issuance (that made the currency an object of wild
  speculative trade, leading to its incurable volatility), and a
  mining reward mechanism that inevitably led to the concentration of
  70% of mining in 7 companies, all in China.

  Many other cryptocurrencies were created since 2013, and more than 200
  have been traded in the last month (31).  While most of
  them were simple copies of bitcoin, created to profit from its
  "private currency scam" aspect, some had interesting innovations,
  that could make them more attractive than bitcoin for both legal and
  illegal trade.

  Ethereum, for example, expanded bitcoin's blockchain to include
  executable programs, rather than just one-time coin transfer orders.
  The programs would be executed by the Ethereum miners, in stages,
  over an indeterminate period.  Such programs were intended to
  implement so-called "smart contracts", that would dispense payments
  automatically and irrevocably on certain computable conditions.
  Ethereum smart contracts were supposed to make lawyers and courts
  superfluous.

  Until a few months ago, Ethereum's popularity and price were
  growing, and it was on its way to displace bitcoin as the dominant
  cryptocurrency.  That did not happen only because a flaw in a smart
  contract caused many Ethereum investors to lose coins worth 70
  million USD to a hacker who noticed it.  In an attempt to recover
  those funds, the developers and miners agreed to "rewind" the
  blockchain to an early state, cancelling that smart contract.  That
  decision made the world realize the fragility of the smart contract
  concept, caused the coin to split in two (24) (26), and apparently
  destroyed its chances to take bitcoin's place.

  (11) Motherboard.com, 2016-10-14:
  "Cocaine Bust Shows How Close the Dark Web and Street Crime Really Are"
  http://motherboard.vice.com/read/cocaine-bust-shows-how-close-the-dark-web-and-street-crime-really-are

  (13) Tom's Guide, 2014-02-27:
  "How to Buy Stolen Credit Cards from the 'Amazon of Cybercrime'"
  http://www.tomsguide.com/us/how-to-buy-stolen-credit-cards,news-18387.html

  (14) The Atlantic magazine, 2016-06-07:
  "The New Economics of Cybercrime"
  http://www.theatlantic.com/business/archive/2016/06/ransomware-new-economics-cybercrime/485888/

  (15) International Business Times, 2016-06-06:
  "Britain's worst paedophile Richard Huckle: How monster preyed 
  on Malaysian children and wanted Bitcoin for child porn"
  http://www.ibtimes.co.uk/britains-worst-paedophile-richard-huckle-how-monster-preyed-malaysian-children-wanted-bitcoin-1563911

  (16) SEC Office of Investor Education and Advocacy, 2013-07:
  "Investor ALert: Ponzi schemes Using virtual Currencies"
  https://www.sec.gov/investor/alerts/ia_virtualcurrencies.pdf

  (17) George Markides on Medium.com, 2014-04-11:
  "Neo and Bee: Enthusiasm for new tech clouds investor judgement 
  and how Cypriot authorities failed to act YET AGAIN"
  https://medium.com/economic-thoughts/neo-and-bee-31667a1d1243#.4kw4tzxza

  (18) Tim Swanson, Great Wall of Numbers, 2015-04-17:
  "A gift card economy: breaking down BitPay’s numbers"
  http://www.ofnumbers.com/2015/04/17/a-gift-card-economy-breaking-down-bitpays-numbers/

  (20) Kevin Collier, The Daily Dot, 2016-01-02:
  "The great Bitcoin experiment that failed"
  http://www.dailydot.com/layer8/bitcoin-bowl-bitpay-one-year-later/

  (22) BitInfo Charts, 2016-10-29:
  Distribution of bitcoin addresses by value
  https://bitinfocharts.com/top-100-richest-bitcoin-addresses.html

  (23) John Hardy, SeeBitcoin, 2016-10-17:
  "The blocksize debate: is an end in sight for the civil war
  that has engulfed Bitcoin?"
  https://seebitcoin.com/2016/10/the-blocksize-debate-is-an-end-in-sight-for-the-civil-war-that-has-engulfed-bitcoin/

  (24) Paul Vigna, Yahoo Finance, 2016-08-01:
  "Ethereum: A Digital Currency Split in Two"
  http://finance.yahoo.com/news/ethereum-digital-currency-split-two-230200061.html

  (26) Frances Coppola, Forbes Investing blog, 2016-06-26:
  "Ethereum And The 'Old Catholics'"
  http://www.forbes.com/sites/francescoppola/2016/07/26/ethereum-and-the-old-catholics/#46db9fde544a

  (31) CoinMarketCap.com, 2016-10-29:
  "Crypto-Currency Market Capitalizations"
  https://coinmarketcap.com/all/views/all/
 
  Insufficient disclosure
  
  In conclusion, the negative-sum characteristic of bitcoin investment
  makes it qualitatively different from other common investments, but
  similar to ponzis and other pyramid investment schemes, to
  penny-stock pump-and-dump scams, and to lotteries and other gambling
  games.
  
  Because this guaranteed negative-sum character is so unique among
  ordinary investment instruments, it should be clearly spelled out in
  the COIN fund propspectus.
  
  Merely saying that the investor MAY lose some or all of his money is
  absolutely not enough.  When bitcoin holders or supporters try to
  convince others to buy bitcoins, they generally dismiss that risk by
  saying that it also exists in any other investment, including the
  stocks of solid and highly lucrative companies; thus inducing the
  prospective buyer (or even explicitly telling him) that investing in
  bitcoin is not fundamentally different than investing in stocks or
  gold.  The prospectus must dispell this common misconcemption by
  explaining the negative-sum character of bitcoin investment,
  observing that it implies a negative expected profit (independently
  of what happens to the bitcoin network), and warning that many of
  the investots will necessarily lose money.
  
  Indeed, the prospectus should also observe that the number of losers
  is likely to be much larger than the number of winners.  This is not
  a mathematical certainty, but an estimate supported by US government
  studies of other negative-sum games like ponzis, chain letters, and
  of course lotteries, in which 90% or more of the "investors" are
  seen to exit with loss. It is also supported by the observation that
  gamblers and speculative investors are more likely to exit the game
  when they are losing than while they are still winning.
  

1c. What are commenters’ views on the risk of loss via computer
hacking posed by such an asset?

  Please see the answer to question 4 below.
  
1d. What are commenters’ views on whether an ETP based on such an
asset would be susceptible to manipulation?

  Please see the answer to question 5a below.

2. According to the Exchange, the Gemini Exchange Spot Price is
representative of the accurate price of a bitcoin because of the
positive price-discovery attributes of the Gemini Exchange
marketplace. What are commenters’ views on the manner in which the
Trust proposes to value its holdings?

  Please see the answer to question 5a below.

3. According to the Exchange, the Gemini Exchange is a Digital Asset
exchange owned and operated by the Custodian and is an affiliate of
the Sponsor.  What are commenters’ views regarding whether any potential
conflict of interest or other issue might arise due to the
relationship between entities such as the Sponsor, the Custodian, and
the Gemini Exchange?

  I do not know what standards exist for other assets and
  exchange-traded instruments in this regard.  As a layman, however, I
  find it peculiar that the value of the backing commodity is defined
  by an entity under full control of the fund's operators, instead of
  an independent marketplace.

4. According to several commenters, there is a need for the Exchange
to provide additional information regarding “proof of control”
auditing, multisig protocols, and insurance with respect to the
bitcoins held in custody on behalf of the Trust, in the interest of
adequate security and investor confidence in bitcoin control.
What are commenters’ views on these recommendations regarding
additional security, control, and insurance measures?

  Some of those measures, such as proof of control and auditing, will
  only make the loss of assets (through accident, theft, or
  embezzlement) evident some time after the fact.  They will not
  reduce the likelyhood of such losses, and will not be of much help in
  discovering the culprits and recovering the assets.  One expects
  that losses by theft or accident will be promptly communicated by
  the fund operators and investgated by law enforcement.  Losses by
  embezzlement will either be falsely attributed to theft (38), or the
  responsible parties will flee after the incident (39).  In all these
  scenarios, the periodic auditing and proof of control exercises will
  not be of any help.

  As for the use of multi-sighature to protect the holdings, one thing
  that the recent Bitfinex invasion showed is that such security
  measures are much less robust in practice than predicted by theory.

  In an attempt to secure its bitcoin holdings against embezzlement or
  theft by hackers and insiders, Bitfinex maintained a separate
  bitcoin wallet for each client account.  The coins in the wallet
  were protected by 2-out-of-3 multisignature.  Specifically, in order
  to remove coins from the wallet, two of these three parties had to
  sign the transaction: the client, and/or Bitfinex operators, and/or
  the independent bitcoin security company BitGo.  Each party created
  the necessary private keys without knowledge of the other two.

  However, in order to keep those wallets up-to-date accounts with the
  trades executed by the clients in the exchange, Bitfinex had to move
  coins from and to thousands of such wallets every day.  Those
  transfers had to be countersigned by BitGo.  However, BitGo had no
  way to verify whether those moves were legitimate, so they set up
  their system to automatically countersign them.  Thus, when a hacker
  (allegedly) invaded Bitfinex's system and proceeded to transfer all
  coins from those wallets to his own, BitGo promptly countersigned
  all those moves.  Somehow Bitfinex operators noticed the attack and
  stopped it after the hacker has stolen "only" 70 million USD worth
  of coins.

  That incident should be a lesson for all parties who trust multsig
  for securing their bitcoins: when countersigning is a frequent
  operation, there is a definite risk that that the secondary
  signer(s) will treat it as a mere formality -- that is executed on
  trust of the primary signer, without an independent check of the
  legitimacy of the transfer.  Or that it will even be automated, as
  in Bitfinex's case.

  (38) James D. Sallah, Crptsy Receivership, 2016-08-02:
  "Second Report of Receiver", page 16 "The alleged hack"
  http://cryptsyreceivership.com/v1/wp-content/uploads/2016/08/Notice-of-Filing-Receivers-2nd-Report-8-2-16-full.pdf

  (39) Emma Lee, TechNode, 2014-05-20:
  "Hong Kong Crypto Currency Exchange HKCEx Collapses with 
  Founding Team Suspected Fled"
  http://technode.com/2014/05/20/hong-kong-crypto-currency-exchange-hkcex-collapses-founding-team-suspected-fled/

  Lack of true "bitcoin security" experts

  The Bitfinex case, and specifically the way "bitcoin security"
  company BitGo failed to perform, also highlights the fact that many
  "bitcoin security experts" are inexperienced amateurs, not even
  competent in ordinary computer security.

  There are no established practices in that profession, and no
  certification programs.  Indeed I would think that there are no real
  "bitcoin security experts" at all, because the only sensible advice
  that a competent security expert should give to its employer, in
  my opinion, is "stay away from bitcoin".

  Risk of loss from "weak" keys

  Another lesson about the (in)security of bitcoin holdings was
  involuntarily provided by Blockhain.Info (BCI) some years ago (8).
  BCI is one of the largest providers of bitcoin wallet software and
  supporting services.  Unlike bitcoin exchanges and certain "bitcoin
  banks" like Coinbase and Circle, BCI does not hold the bitcoins or
  private keys of their clients.  Instead, each client keeps that
  information in his own computer, and uses BCI-provided software
  (embedded in their webpages) to manage it.

  On 2014-12-08, BCI released a new version of their software for use
  by their clients.  That version included spurious changes to the
  random number generation routine, which caused it produce only 256
  possible values, instead of the astronomical variety required by the
  bitcoin protocol.  As a result, any new private keys generated with
  that software, while looking just as random as properly geberated
  keys, were in fact easily guessable: one only needed to generate the
  256 possible keys, and check whether they "unlocked" the
  corresponding address.  Moreover, if two transactions taking coins
  from the same address were signed with that software, an observer
  would have one chance in 256 of extracting from them the private key
  of that address.
  
  Fortunately for BCI, the problem was noticed by an independent
  bitcoin researcher who was monitoring the blockchain for the second
  kind of vulnerability above; and BCI released a fixed version of the
  software less than three hours later.  Nevertheles, in that short
  period thousands of client had their private keys exposed, and
  hackers were able to steal some of their bitcoins.

  That incident (and a few others like it) highlight an important risk
  of the bitcoin protocol: the signature mechanism is secure only if
  the keys are generated truly at random.  However, there is no test
  that can be applied to a private key to determine whether it is
  indeed random.  One must trust that the software that was used to
  generate it did not have accidental or intentional flaws, that would
  result in "weak" keys that are easy to guess by someone who is aware
  of the flaw.  Yet, it is practically impossible for the users of
  such software to verify that it does not have such flaws.  And it is
  impossible to rationally assign a probability value to the risk of
  the software being compromised.

  (8) Brave New Coin:
  "Blockchain.info Bug Exposes Users Private Keys"
  http://bravenewcoin.com/news/blockchain-info-bug-exposes-users-private-keys/

  Security is wholly dependent on secrecy of private keys

  It is easy to overestimate the security of the bitcoin protocol by
  comparing it to online banking and other financial services that are
  accessed via secret passwords, and are generally trusted by their
  users.  

  However, for these services the password is only the first
  of several layers of protection.  If a hacker steals one's bank
  balance after gaining access to one's password, the funds can often
  be recovered by blocking and reversing bank transfers and cash
  withdrawals, and ultimately by insurance.  

  None of these additional layers of protection are available to
  bitcoin holders: if the private key is obtained by a hacker, the
  coins are permanently lost.

5. A commenter notes that the Gemini Exchange has relatively low
liquidity and trading volume in bitcoins and that there is a
significant risk that the nominal ETP share price “will be
manipulated, by relatively small trades that manipulate the bitcoin
price at that exchange.” 

5a. What are commenters’ views on the concerns
expressed by this commenter? What are commenters’ views regarding the
susceptibility of the price of the Shares to manipulation, considering
that the NAV would be based on the spot price of a single bitcoin
exchange?

  In the seventh amended filing, the proponents replaced the
  4:00 pm spot price at the Gemini exchange by the price of an auction
  that is to be held at 4:00 pm every day, on that same exchange.

  The change does not seem to affect the concerns that I expressed in
  my previous letter.  The auction has been occurring for six weeks
  only, and it is not clear how it will evolve.  It is not obvious
  that the auction will be more attractive to traders than normal
  trading.

  The auction closing volume has shown a slight decreasing trend since
  its inception (9) and is now under 1 million USD during work days,
  and considerably less during weekends.  With such low volume,
  it seems possible to manipulate the NAV value by entering 
  suitable bids or asks in the auction.

  If the observed downward trend in the volume continues, it also
  seems quite possible that, on some days, the auction may not execute
  any trades, because the bids and asks fail to cross over.  In that
  case, the nominal asset value for the day would be undefined.

  (9) Bitballoon.com
  "Gemini Auction Price History"
  http://geminiauctionhistory.bitballoon.com/
  (Select "total $ amount" option to see the volume in USD)

5b. What are commenters’ views generally with respect to the
liquidity and transparency of the bitcoin market, and thus the
suitability of bitcoins as an underlying asset for an ETP?

  Since 2013, the price of bitcoin has been defined mostly by the
  major Chinese exchanges, whose volumes dwarf those of exchanges
  outside China.  As I pointed out in my response to question 1a
  above, those exchanges are not regulated or audited, and are
  suspected of engaging in unethical practices like front-running,
  wash trades, trading with insufficient funds, etc.

  As for liquidity, the charts of prices at those exchanges have a
  peculiar pattern (10).  Quite often there is a sudden increase or
  decrease of the price by several percentage points, which seems to
  be a large purchase or sale by a single trader.  The amounts do not
  seem large: while I am writing this letter, the sale of 1500 BTC
  (about 1.2 million USD) on the exchange BTCC (formerly BTC-China,
  one of the largest of the world by trade volume) would push the
  price down by more than 8%.  Thus, it would seem that the world
  bitcoin market has rather limited liquidity.

  Moreover, after such a "whale move", instead of returning to the
  approximate value that it had before the move, the price remains for
  hours hovering around the new level.  I interpret this behavior has
  evidence that the price is defined entirely by speculation, without
  any ties to fundamentals.  That is, the traders have no reason to
  think that the new price after the move is "too high" or "too low",
  and just continue trading at the new price, indifferently.

  (10) Bitcoinwisdom.com
  "OKCoin BTC/CNY" price chart
  https://bitcoinwisdom.com/markets/okcoin/btccny
  (Select 5 min intervals to see the abrupt changes)

6. The Exchange asserts that the widespread availability of
information regarding Bitcoin, the Trust, and the Shares, combined
with the ability of Authorized Participants to create and redeem
Baskets each Business Day, thereby utilizing the arbitrage mechanism,
will be sufficient for market participants to value and trade the
Shares in a manner that will not lead to significant deviations
between intraday Best Bid/Best Ask and the Intraday Indicative Value
or between the Best Bid/Best Ask and the NAV. In addition, the
Exchange asserts that the numerous options for buying and selling
bitcoins will both provide Authorized Participants with many options
for hedging their positions and provide market participants generally
with potential arbitrage opportunities, further strengthening the
arbitrage mechanism as it relates to the Shares. 

6a. What are commenters’ views regarding these statements?
Do commenters’ agree or disagree with the assertion that
Authorized Participants and other market makers will be able to make
efficient and liquid markets in the Shares at prices generally in line
with the NAV?

  This question brings up another major difference between bitcoin and
  almost any other tradeable asset: there is practically no reliable
  or meaningful information about the state of the bitcoin economy.

  Although the blockchain offers an open record of all bitcoin
  transactions, the anonymity of the addresses prevents useful
  analysis of that traffic.  It is known that large fractions of it
  are not payments, but transactions made with other purposes.  A
  major fraction generated by "mixers" or "tumblers", money laundering
  services that move client coins through thousands of addresses,
  combining and splitting them thousands of times.  Another large
  fraction is due to online gambling, where the bitcoin protocol is
  used only as a secure way to place bets and throw fair dice.  Other
  non-payment uses include moving coins between "cold storage" and
  "hot storage", depositing and withdrawing coins at exchanges, and
  wallet housekeeping.

  There may also be a significant amount of "spam" traffic:
  transactions from one owner to himself, that are intended to
  simulate adoption growth, or to harm the system by reducing its
  effective capacity.  

  There have been several obvious instances of the latter, in the form
  of anomalous surges of incoming transactions.  These "spam attacks"
  created backlogs of unprocessed transactions that sometimes took
  days to clear, delaying the confirmation of many legitimate
  transactions by many hours.  These unpredictable events are another
  reason why bitcoin is unlikely to ever gain wide use in commerce.
  
  As I mentioned before, there is no data on the volume of 
  legal payments executed with bitcoin.  The bitcoin payment
  processors are all privately owned, and generally do not release 
  any useful data on the volume of payments that they process,
  or its distribution by country, class of goods and services, etc.
  No one knows how many bitcoin payments occur without the
  intermediation of those companies.  

  Various lines of evidence indicate that legal commercial payments
  make up only a small fraction of the total blockchain traffic, which
  may be as low as 5% or less.  Therefore, it is not possible to use
  the total traffic as a proxy metric for the volume of legal payments
  and its growth trends.  (Illegal payments are probably many times
  the legal ones, but they still make up a minoritary fraction of the
  traffic.)

  Without any data on the volume of legal payments -- the only
  parameter that is alleged to provide value to the asset -- investors
  will have no way to estimate its fair price, not even within an
  order of magnitude.  Investing in the fund, like investing in
  bitcoin, would be gambling in a crazy lottery with unknown odds,
  unknown payouts, and unknown drawing date.

  An efficent market requires that sufficient information about the
  asset's future value be available to the investors.  Therefore, the
  answer to question 6a must be "no".

6b. What are commenters’ views on whether the relationship between the
Gemini Exchange and the Trust’s Sponsor and Custodian might affect the
arbitrage mechanism?

  Please see the answer to question 3 above.

  
  Concluding remarks

  Bitcoin was created as a computer science experiment, to validate a
  solution that "Satoshi Nakamoto" believed to have found for a
  decades-old problem.  It was not designed to be an investment
  instrument -- a role that it assumed due to exaggerated projections of
  its future usage.

  The developments of recent years have not improved its prospects; on
  the contrary, they have exposed its many flaws -- severely limited
  capacity, 10-minute minimum confirmation time, centralization and
  unsustainable cost of mining, inherent volatility, uncertain
  survival after the block reward disappears, inability to evolve, and
  more.  Its future is now more uncertain than ever.  The offering for
  public trade of such a questionable asset, packaged as an ETF, is,
  at the very least, highly premature.

  Strictly speaking, it is POSSIBLE that bitcoin will one day become
  used by hundreds of millions of people and millions of merchants.
  Just as it is POSSIBLE that a land plot in the middle of the Sahara
  will become as expensive as real estate in downtown Las Vegas,
  because someone MAY build a popular casino right next to it.  That
  mere possibility, however, should not be enough to make it a valid
  investment.

  The Commission may also want to consider that, if the bitcoin ETF is
  approved, there would be no reason to deny the same provilege for
  the other 600+ cryptocurrencies that have been created, or for
  other equally immaterial and unbacked assets.


?? The instability after reward is gone

?? Fees and hashrate as the reward dwindles

?? The "private currency scam"

?? Other cryptocurrencies will want ETPs too.

?? Bitcoins are not scarce: altcoins, increased issuance, fractional banking...

?? Like real estate in the middle of the Sahara: