1) It's not really clear to me how that would work, but assuming it does then it will go into a basket of attacks that are possible but unlikely due to the economic disincentives to do so.

2) That said, is the Achilles heal of this proposal the lack of a mechanism to lower the block size? 

3) Let me put it another way, I've read that both Gavin and yourself are favorable to a dynamic limit on the block size. In your view, what is missing from this proposal, or what variables should be adjusted, to get the rules to a place where you and other Core developers would seriously consider it?

On Wed, Sep 9, 2015 at 12:18 AM, Adam Back <adam@cypherspace.org> wrote:
> A selfish mining attack would have to be performed for at least 2000 blocks over a period of 4 weeks in order to achieve a meager 10% increase in the block size.

You seem to be analysing a different attack - I mean that if someone
has enough hashrate to do a selfish mining attack, then setting up a
system that has no means to reduce block-size risks that at a point
where there is excess block-size they can use that free transaction
space to amplify selfish mining instead of collecting transaction
fees.

Adam



--
-------------------------------------------
Co-founder, OB1
Core developer of OpenBazaar