Here it doesn't matter who is greedy or not.
If a situation/proposition was the subject of a vote, and one side won, the devs/team/system should implement the decision, no matter if you/they/anyone like it or not. I don't know the story but if what LaaMos Seeth said is true, that the winning side proposition wasn't implemented, then why they started a vote session in the first place?
Not talking about governance power and stuff. Simply don't do it, because if you do this, is like you said: "Let's vote! But if you win you'll loose anyway" - in other words - let's vote if you wish, as we will piss on the result anyway, because we can.
54% ... I would say "no consensus" rather than win because in statistics, we have something like a small sample error or statistical error, and this 4% is nothing more than that. I'm not surprised that it was not implemented. There are more important things now than satisfying 54% of network at the expense of going against 46%. There are things that will satisfy 100% of network (like sharding, mobile wallet, ledger support, option to encrypt wallet with password etc.)
Maybe it was 54:46 because both options are equally shitty and it is worth to repeat vote with additional option?
I personally prefer option A, because otherwise the network will be full of bad flips.
Let's say you are a human which has about 700 DNA as reward for the last validation. In the first scenario if you make 1 bad flip you will lose all your reward and next time will create better flips. In the second one you'll just lose 15 DNA, which is tiny 2% of your overall reward. You may even don't notice such a penalty and will continue to make bad flips.
If you want some trade off I would suggest the third option C:
C. 1 bad flip = 20% of overall reward unpaid instead of 1 flip reward unpaid like in option B. 20% means 1 flips out of 5 is bad. 40% means 2 out of 5 are bad, etc.