Network Stake-Dependent Interest actually mitigates this potential attack. As the PoS difficulty drops, so too does the probability of receiving a stake and the corresponding probability of scoring excessive blocks without an insane amount of coin.
Unfortunately not, the attack involves an exploit of the difficulty
adjustment, similar to the multipool problem in PoW Dogecoin faced once. Note that the comparably old NovaCoin also has a dynamic interest rate, and many people consider it as a flaw.
Time will tell what happens - if it would be so obvious to fork a coin with 1% stake then I would assume some people would have done it before (and we would have noticed). Remember that full PoS still must be considered as an experiment, a successful experiment so far, but still an experiment.
I agree crypto and PoS is experimental, but if you analyze deeply the variance between average difficulty adjustment trends and our log stake adjustment, you find the probability of attack is extremely low. But definitely worth pursuing research further! thanks.