Alright folks.
This is how "MasterMind" kinda works. MasterMind has different colored pins (6 colors), and there is no lying possible. This is the same, only with 5 "colors" and some people may lie.
Now, a little green bird tells me that there is no way that ShadowKellyGarnetLurker is going to lie, and that's a huge plus with this scheme. With this, I blindly trust the numbers he has chosen as unbiased.
We can get a "VERY GOOD" guaranteed, and if the scum tries to lie, we will catch them and string them up.
So we can do it this way:
SL asks: 74,14,39,23,51
Whether if he gets a "NO" or a "YES," the next player asks for OVERLAPPING NUMBERS:
Mighty Orbots asks: 39, 23, 51, 18, 52
So if SL gets YES, and MightyOrbots gets NO, that means that the "VERY GOOD" is within the first two numbers. See? We'd just need to narrow it down to two.
Also, we should be able to double-check whether someone is lying pretty quickly. No one can lie unchecked with this system. Overlapping with 3 numbers means that the scum is painted in a very tight corner and will be caught if lying.
The chances of there being ONE "VERY GOOD" in those first 15 numbers is 99%.
To show my calculations (actually, those are my
husband
's gay lover calculations, the probability of "VERY GOOD" is easily computed as the 1 - P{everything else}.
So,
1- (60/75*59/74*58/73* [....] *44/59) = .99
Instead of choosing from a set that contains one or more "VERY GOOD" we can actually narrow it down to an actual "VERY GOOD."
So here are the overlapping sets:
(1)
74, 14, 39, 23, 51
(2)
39, 23, 51, 18, 52
(3)
51, 18, 52, 37, 69
(4)
52, 37, 69, 63, 47
(5)
69, 63, 47, 66, 20
(6)
47, 66, 20, 34, 5
(7)
20, 34, 5, 56, 41
Now, if we get results that deviate from the expected number of "VERY GOOD" which is, I repeat, 99% chance of hitting one, this means we may have TWO LIARS IN A ROW. Should that happen, anyone is free to unexpectedly double-check a set of numbers that someone else already asked. Of course, if someone does that it reduces the span of numbers, so let's suppose that we ask for 12 numbers instead of 17. Investigating 12 numbers, the chances of a VERY GOOD is still 95%.
And to put it another way, if we FAIL to get a VERY GOOD, we will catch scum, for sure. Either way we win something today. A VERY GOOD, or a scumbag.
My plan to go last was about making a surprise spot check on whoever is more likely to be lying. But if hewitt can agree to do a spot check on the other players, either of his own choosing, or by concensus. I happen to trust myself more than anyone else, and I'm not sure letting scum have any control on how gets to spot check is a good idea.
This being said, we're probably better putting a very townie player in the last spot, rather than a less trusted player. Start with a presumed townie, and end with a presumed townie. That's the best combination to make sure we have true results.
[size=75]This is a block of madness that can be added to insanity you post. There is a 255 psychiatric limit.[/size]