Why would only -one- brown-eyed person remain, or many blue-eyed? And does this idea of yours, which I don't understand by the way, also apply to someone with red, orange, purple eyes? Who kills who and why is everyone aware of this being the most logical decision?
I fail to see any logic in this. By killing brown-eyed people, you would find out your own eye color if a blue-eyed person tried to kill you. This, to me, is a way of communicating between the islanders, which is forbidden. And murdering that many people for a chance no one leaves does not seem very logical to me. It is more logical to wait a 100 days, when all 100 people leave.
Look at it this way, which simplifies my solution to its most basic:
You are on the island. You know you will either escape, or not. There is nothing at all you can do to improve your chances of escape, but there is something you can do to improve your speed of escape. There is no logical reason to consider the welfare of the population as a whole as being important.
Slaughtering anyone you see with non-blue eyes reduces the set remaining to people who KNOW they must have blue eyes (as they're still alive, and no longer being attacked), and can therefore leave on the very first night, as opposed to on night 100. You have not affected your personal odds of leaving at all, merely increased the speed with which you do so. Therefore, it's worth doing it.
The only snag is if the people being attacked fight back. If that is the case, they would surely end up killing some blue-eyed people, reducing your personal odds of leaving, and raising the theoretical - albeit unlikely - possibility of everyone fighting to just one survivor, who might not have blue eys.
However, if the primary aim of the exercise is to leave the island (as opposed to survival on the island being an acceptable outcome), then actually the non-blue eyed people would have no reason to fight back as they would know that if they are being attacked, they do not have blue eyes, and therefore can never leave anyway so may as well let themselves be killed.
(Put another way, essentially my solution simplifies the set to only one type of content right from day one, therefore the problem can be solved on night one, in conceptually just the same way it would be if there was only one blue eyed person among any number of brown-eyed people).
Suppose you are one of the people on the island, and you can see N people with blue eyes.
Let N=10 for sake of this argument. Whatever colour someone's eyes are, you know that they will also be able to see someone with blue eyes. So everyone on the island can see someone with blue eyes, and everyone knows that everyone can.
So the guru isn't adding anything. Before the guru speaks, everyone on the island already knows that the guru can see someone with blue eyes. So what they say is already a fact known to all. Nobody has any new information to work with. Perfect logicians won't be able to deduce anything new, so nothing will change.
If my answer is wrong, then what's wrong with my logic?
To know that the other parties know other people have blue eyes requires a leap of faith (it's inductive, not deductive). Humans like you or I can do this, but as
Kommander jokingly - but correctly - pointed out, these people on the island are not human like you or I; they work solely on the basis of fact - perfect strong logic - and can't make that inductive leap of faith. Therefore it's only when actively
provided with the absolute factual knowledge that other people know that other people have blue eyes, that then can solve the problem using their perfect deductive logic.
THAT'S why they need the Guru under the original solution; to provide 100% truthful data which they cannot gather objectively via any other means. Without that fact, the problem ends up being insoluble as they won't induct it for themselves. You or I (being capable of induction) would not require the guru, you're correct... but we could be wrong in our inductive logic (whereas deductive logic cannot ever be internally wrong under any circumstances) and therefore we're not being perfect logicians in the way the islanders are being defined as. Incidentally, it points out how humans solve complex problems quicker and better, with less data, than an algorithmic computer program would need... at the admitted expense of being wrong sometimes.
(At least, that's the rationale you have to use to make the solution correct... I'm not necessarily saying it's a good problem... and it
certainly was a too loosely phrased one).