The maximum block-size is one that can be filled at zero-cost by
miners, and so allows some kinds of amplification of selfish-mining
related attacks

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. 

If there goal is to simply drive up fees to gain acceptance into the block, we're in exactly the same position we are in today (as in nothing stops a miner from doing this). 
If the goal is to increase the block size to push out smaller miners, they'll have to perform this attack over the course of years and destroy any economic incentives they have for mining in the first place.

 why give this power up to a subsection of the ecosystem in order to make it easier to change or game

Well this same could be said for developers trying to predict what the appropriate block size should be over the next 20 years... it's a hallmark to a group of bankers trying to predict the appropriate interest rate for the entire economy. Just as it is impossible to predict the appropriate hash rate to secure the network, so it goes for the block size. Both need to adjust dynamically to the scale/adoption of the network.

On Tue, Sep 8, 2015 at 11:13 PM, Adam Back <adam@cypherspace.org> wrote:
The maximum block-size is one that can be filled at zero-cost by
miners, and so allows some kinds of amplification of selfish-mining
related attacks.

Adam


On 8 September 2015 at 13:28, Ivan Brightly via bitcoin-dev
<bitcoin-dev@lists.linuxfoundation.org> wrote:
> This is true, but miners already control block size through soft caps.
> Miners are fully capable of producing smaller blocks regardless of the max
> block limit, with or without collusion. Arguably, there is no need to ever
> reduce the max block size unless technology advances for some reason reverse
> course - aka, WW3 takes a toll on the internet and the average bandwidth
> available halves. The likelihood of significant technology contraction in
> the near future seems rather unlikely and is more broadly problematic for
> society than bitcoin specifically.
>
> The only reason for reducing the max block limit other than technology
> availability is if you think that this is what will produce the fee market,
> which is back to an economic discussion - not a technology scaling
> discussion.
>
> On Tue, Sep 8, 2015 at 4:49 AM, Btc Drak via bitcoin-dev
> <bitcoin-dev@lists.linuxfoundation.org> wrote:
>>
>> > but allow meaningful relief to transaction volume pressure in response
>> > to true market demand
>>
>> If blocksize can only increase then it's like a market that only goes
>> up which is unrealistic. Transaction will volume ebb and flow
>> significantly. Some people have been looking at transaction volume
>> charts over time and all they can see is an exponential curve which
>> they think will go on forever, yet nothing goes up forever and it will
>> go through significant trend cycles (like everything does). If you
>> dont want to hurt the fee market, the blocksize has to be elastic and
>> allow contraction as well as expansion.
>
>
> _______________________________________________
> bitcoin-dev mailing list
> bitcoin-dev@lists.linuxfoundation.org
> https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev
>



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