That behavior is already countered under this HF20 proposal :) Under HF20, any vote coming in early sends rewards to the pool (or more accurately, does not claim as much of the pool as the rshares provided by the vote would normally suggest). If a whale used a second voting account to vote at 0 minutes, the vote would give 75% to the author but would claim 0 curation rewards at all. It would be a loss for the whale.
I think there are two main questions - first, is HF20 an improvement on current situation? I think the answer to that is clearly yes.
Second, is HF20 the best way to solve current situation? Here we have some disagreement, and I didn't cover alternative proposals in my analysis. Competing proposals keep the chunk of curation that a vote misses for coming in before 15 minutes in the reward pool for the individual post rather than "returning" it to reward pool at large. Reasoning is that the largest accounts get a larger share of the reward pool, so returning rewards to the pool rewards them; I counter that the largest accounts are the ones giving up large rewards under HF20, and there is no way to game the reward pool to recover a greater % beyond whatever % of the reward pool the account already controls. I believe it would be trivially easy for a large account to recover all the "missing" reward through use of alt account voting strategies if the reward is kept in the pool for the post at hand.