[Bitcoin-ml] Time to get agreement on new DAA for Bitcoin Cash (Eli)

[Scott Roberts]

There are 2 objectives to the difficulty algorithm:
1) consistent solvetime
2) no excess benefit to large miners which would cause profit in
collusion.  The coin reward is specifically meant to motivate
selfishness to avoid miner collusion.

Getting 1) on average is easy, even for small coins.  Satoshi's
equation does not work for small coins because he did not plan on
having a small coin.  Bitcoin core will have to fork to fix this
error.  It is well known Satoshi made errors.  The change to satoshi's
equation is simple, despite people of the past making it hard and open
to debate.  Bitcoin cash has somehow stumbled onto the realization
that simple is better.  What they're doing is just making Satoshi's
equation a rolling average with a smaller N. I called this Zawy v1 and
karbowanek coin was the first coin I know of that used it, with N=17.
Bitcoin cash has added two unnecessary mild complications that will
not have any noticeable effect as I explained. Bitcoin core will have
to fork to copy bitcoin cash's code.  I'm about to email their list to
recommend it, and mention Degenr8's as something slightly better.

Item 2) is similar to but different from miners reaching an agreement
that allows them to dictate code in the same way big banks dictate
monetary law, resulting in higher fees for users and bailouts when the
security fails from taking higher risks in pursuit of higher total
fees. Lower interest rates (fees per tx) seemed good for home owners
and also allowed more total interest (total fees per block) to be
given to banks. It seemed like a win-win until the security of the
system fell apart as a result of trying to achieve both these goals.
The current blockchain technology does not allow both.  The users had
to pay a $20 trillion bailout to the banks (there are papers that
explain this number) to keep from losing all their coin that the banks
were ostensibly protecting in their databases ("blockchains").
Bitcoin SegWit2x might be a good idea for now, but it sets a
frightening precedence that exactly parallels banks dictating law in
pursuit of more fees.  Until I saw the email to this list 4 hours ago,
I had no idea bitcoin cash is also deferring to miner wishes. I
disagree because they are employees, not employers.

The N=2016 window only works for the largest coin. Otherwise, it's not
a problem. That's why satoshi did not bother to fix it.  Similarly he
was unnecessarily blunt in causing coin reward to halve every 4 years.
Both choices are strange, almost as if he did not want to spend an
hour to figure out the continuous version of the math because he could
not see that being precise would help.  To be more gentle, the coin
reward should have been 69.3*0.5^(N/210,000) where N=block number to
get the same result as halving every 4 years with a total of 21M
produced. Some alts do something like this.

There is one thing incredibly clear about both of these "staccato"
math decisions by Satoshi. One suddenly changes the amount of work to
be done. The other brutally changes the amount to be paid for the
work. Both show concern for miners was the last thing on his list.
They are being paid to do a job, not to collude as monopolies do to
raise prices or dictate law. BTW I'm for as many bitcoin forks as
possible, and that they all have a good difficulty algo so that they
can compete efficiently and fairly. I was banned from SegWit2x mail
list for quoting Satoshi on this issue. But you asked why Satoshi did
things like he did. That was the right question.

On Thu, Nov 2, 2017 at 8:21 AM, Tom Zander via bitcoin-ml
<bitcoin-ml at lists.linuxfoundation.org> wrote:
> Something I’ve been missing is people actualy researching first what the
> wanted output of an algo is supposed to be.
>
> The majority of the people seem intent on focusing solely on getting a 10
> minute block, which is rather impossible due to the nature of the beast.
> But much more importantly, there are a lot of economic reasons why this is a
> bad idea.
> Nakamoto could have chosen a 144 block algo too, maybe we should realize why
> he didn’t.
>
> If you want to improve upon the system, I’d suggest the first thing to do is
> determine what the outcome is supposed to be of this control system.
>
> --
> Tom Zander
> Blog: https://zander.github.io
> Vlog: https://vimeo.com/channels/tomscryptochannel
> _______________________________________________
> bitcoin-ml mailing list
> bitcoin-ml at lists.linuxfoundation.org
> https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-ml