Thanks for posting to clarify the change.
In other words, the private details of each transaction will be recorded on a proprietary ledger so that those details can remain private. But the fact of each transaction's occurrence will be recorded publicly on Bitcoin's blockchain, and recorded in a form that allows parties privy to the transaction details to verify cryptographically that the entries in the proprietary ledger correspond to the sequence of transactions published to Bitcoin's blockchain. Basically, this scheme uses Bitcoin's blockchain to provide security (immutability) and public accountability while keeping private information private.
(This insert-a-hash scheme is typical of "colored coin" implementations even when privacy is not desired, since each Bitcoin transaction can insert only a small amount of data into the blockchain.)
The amended S3 filing [0] spells all of this out in detail. The most relevant section starts on page 35, and the first paragraph of page 36 describes the core of the scheme:
In connection with a digital securities transaction, the
tØ software will publish the transaction to the proprietary
ledger ... . Concurrently, the tØ software will electronically
publish the proprietary ledger and commence the process of
embedding in the Bitcoin blockchain information necessary
to mathematically prove the validity of available copies
of the proprietary ledger. Specifically, after a set of
transactions in our digital securities have been executed
and recorded to the proprietary ledger, the Pro
Securities ATS will send a de minimis amount of Bitcoin
from an ATS-controlled Bitcoin wallet to another ATS-
controlled Bitcoin wallet using the blockchain protocol.
This blockchain protocol provides for an editable field
that can be used to implant code or other data within the
Bitcoin transaction that will be embedded into the
blockchain, and the tØ software will use this field to
implant anonymized cryptographic hash functions for the
digital securities transactions reflected on the
proprietary ledger into the Bitcoin transfer made by the
ATS. The blockchain will validate this de minimis Bitcoin
transaction and embed it, together with the implanted
anonymized cryptographic hash function, into the Bitcoin
blockchain. As a result, once the Bitcoin transaction is
immutably embedded into the Bitcoin blockchain, an
immutable record of the digital securities transactions
reflected on the proprietary ledger is also recorded
within the Bitcoin blockchain. ...
You get all the the security of a 6.6 billion network without the costs.
Private blockchains don't need to be as inefficient as cryptocurrencies either because they can have a very small, specific use case and feature set.
Even if we disagree on the merits of HFT, we can surely agree that settlement is an unnecessarily slow and unreliable process. In September, 2011 (the most recent date for which I could easily find data), there was a failure to deliver $200M worth of stock per day! [1] In the Treasury market, the daily average is currently $50B! [2] Blockchain technology has the potential to solve this problem.
[1] https://en.wikipedia.org/wiki/Failure_to_deliver
[2] http://www.dtcc.com/charts/daily-total-us-treasury-trade-fai...
Patrick Byrne from t0 will agree with you, but that's b/c he's against naked shorts.
[1] https://t0.com/
I read his speech from a charity event back several years back, periodically, to keep perspective.
http://www.pmc.org/blog/2015/9/17/pmc-threshold-moment-billy... (scroll down)
a) The blockchain protocol is the final arbiter of ownership, and those who purchased the stock in the other branch of the chain no longer own the stock. Caveat emptor.
b) Courts rule in favor of the people who actually bought the stocks and marks the attack branch as invalid. This is possible because the attack was witnessed by so many people.
It's unlikely that a fork would be resolved with (a), but if it does, there would be a lot of very pissed of investors and lawsuits. All it takes for such a fork to happen is someone with the motive to do it and the means to lean on a few mining pool operators.
It's more likely that we would see outcome (b), in which case, the courts are the final authority and then what the heck is the point of using an expensive, slow, poorly scalable, decentralized consensus system?
I totally see the value of a decentralized system despite the shortcomings, but if you're coming to do away with that aspect, what on earth is the point?
Of course stealing the private key of the hot wallet representing 10% of Overstock.com, well that would be a very juicy target indeed.