2. Explain what is wrong with
the example (the one with real numbers) that shows the reward per block is smaller for big miners.
Right here;
Reward per block for 1% miner: 5943 * 0.0006885 + 3.2223 - 0.4589 = 6.8552 BTC (more than in scenario 1)
Reward per block for 90% miner: 7251 * 0.0006885 + 3.2223 - 3.5294 = 4.68521 BTC (less than 1% miners in this scenario; more than the miners in scenario 1).
No miner is going to mine a block that costs him more than 2 BTC to mine. Since it's not economically viable to mine larger blocks, you're right that there isn't an economic advantage given to the larger mining entity, and what I wrote earlier wouldn't apply. What I wrote would only apply to penalties that don't reduce the reward below the target block size reward.
Why would a miner mine a block at a loss just to accept more transactions? Regardless, any market participant that engaged in this behavior would just get out-competed by another that didn't.
I don't understand what you wrote here. These are the results for the optimal strategy for each miner, which maximizes his profit. Should I show the derivation? It's just a bunch of unwieldy equations that
my silicon overlord took care of for me...
Each miner wants to maximize tx fees + collection - penalty. The small miner can't hope to affect collection because it depends on what the other miners do. But the big miner does meaningfully affect collection as he makes bigger blocks. As he adds transactions to his blocks, all of tx fees, collection and penalty increase. At some point, the increase in penalty outweighs the increase in tx fees + collection, so he stops there. Whether one or more of these summands is more than 2 BTC is irrelevant, it's their sum that counts.
In the part you quoted, no miner mines at a loss. The small miner has a profit of 6.8552 BTC per block, and the big miner has a profit of 4.68521 BTC per block. If the big miner built smaller blocks, his rewards per block would be smaller. If, for example, he included only 6000 txs/block, like the small miners, he would get only 4.50 BTC per block (instead of 4.68), just like the miners in scenario 1.
If the big miner takes a longer-term view, he may notice that creating supersized blocks will give him a profit at first, but then attract more miners, who will increase the difficulty, which will reduce his profits, to the point where he is not competitive with the small miners - so he can give up the whole thing and just build normal blocks (ignoring reclaiming his own penalty in his optimality calculations). And then we go back to the situation where big and small miners are completely equivalent, and get the same reward per block.