If I am not mistaken, you’re proposing that by locking up the transaction fees earned for the standard 600 blocks, same as is already the case for the coinbase minted rewards, then you posit no one would want to cheat because the system would not converge on consensus (or confirmations would become very slow) and everyone would lose their value.

But that is refuted in the research paper. Even Vitalik wrote about how about altruism-prime is an undersupplied public good. I refer you to Vitalik’s explanation. IOW, if you do not cheat, the others will, so you also have to cheat. There’s no Nash Equilibrium, which is precisely the point. I believe that is a Prisoner’s dilemma. Everyone must defect, which is also what has happened to Steem’s reward system.

I like this explanation of Nash Equilibrium.