I'll present here an argument I've heard presented so far. It goes like this:
Powerful miners could raise the block size limit by mining larger than usual blocks, e.g. by adding some volume of self-transactions which wouldn't cost them all that much. Raising the limit would allow them to mine bigger blocks (no surprise), which would disadvantage less powerful miners who are unable to mine such large blocks, and would increasingly lose out on fee revenue. Rinse and repeat until smaller miners are successively eliminated, leading to increased centralization in mining.
It sounds like a valid concern, but real life is usually more complex than I imagine.
Would such an attack run into an orphan risk barrier before it becomes a real concern?
Would it only work if the attack managed to essentially starve the network of transactions available to the smaller miners, ie. they would have to endure substantially reduced block sizes than before the attack?
Curious to hear your views.
