Mr.Duck on 9/1/2007 at 01:16
Not at all, sailor!
Come and join us, we're sipping margaritas butt-nekkid.
:cool:
mopgoblin on 9/1/2007 at 01:28
Quote Posted by TheGreatGodPan
If the population of ttlg was a random sample, you'd be right about statistics saying that, but we aren't. We're highly self-selected based on certain factors (liking games by Looking Glass). The male/female ratio, for example is far from random. Note: I'm not saying there aren't any gay people here as some people have stated their preferences in the forum, but just that we cannot necessarily know it from statistics.
Not from probability theory alone, no. But you're not supposed to do the maths without considering other aspects of the question. I'm going to assume that 4% of the general population are homosexual, and that TTLG has 270 active members (there are 270 people with 1000+ posts, it feels like a fair estimate, and I'm not going to spend hours looking through all the forums figuring out a precise number).
Now, it seems reasonable to assume that people generally discover the forums independently of each other, and in particular that given any two forum members, the probability that one is homosexual is independent of that of the other. Consider the subset of the general population that are potential TTLG people - if the proportion p are homosexual, then the probability q that none of the 270 active members are homosexual is:
q = (1-p)<sup>270</sup>
Let's try setting q to 0.05 (95% sure there are homosexuals - generally good enough for "statistics tells us..."), 0.5, and 0.95:
<table><tr><td>q</td><td> </td><td>p</td></tr><tr><td>0.05</td><td> </td><td>0.01943211205402936</td></tr><tr><td>0.50</td><td> </td><td>0.002563919309782315</td></tr><tr><td>0.95</td><td> </td><td>0.0001899571202594554</td></tr></table>
So statistics does tell us that we can expect at least one homosexual amongst the 270 people considered, unless there's some bias related to TTLG that specifically repels homosexuals with 50% effectiveness, <em>after</em> negating any effects in the other direction. That's a pretty strong bias - where would it come from? Consider another influence that might oppose this bias: the internet has significant attractions to smaller or less well-liked groups, provided that their members can afford access.
If we're happy with a 50% chance that at least one person is homosexual, the bias has to change things by over 93% to prevent this. Or if we wanted "statistics tells us there aren't be any homosexuals" - a 95% chance that there are no homosexuals - the bias has to repel 99.5% of homosexuals. Incidentally, if 270 people are a true random subset of a population that's 4% homosexual, there's a 99.998% chance that at least one of them is homosexual.
Martin Karne on 9/1/2007 at 01:51
A family branch of mine starts with this name, Nones, is that a Greek name because of the ending S?
Crouches and hides from attacking barbarians.
Scots Taffer on 9/1/2007 at 01:53
Quote Posted by mopgoblin
stats
you sunk my battleship :(
Jackablade on 9/1/2007 at 02:38
yeah I lost my erection too
fett on 9/1/2007 at 04:28
Quote Posted by mopgoblin
Math Shit
Does this mean we should diversify by actively recruiting homosexual forum members? I thought that's what the new Morrowind forum was for. :confused:
demagogue on 9/1/2007 at 05:15
Can't we just use the same method to figure out how many Greeks are active around here?
I mean, math can't be wrong can it?
Martin Karne on 9/1/2007 at 07:20
LMAO at the new forum.
:laff:
N'Al on 9/1/2007 at 11:53
Quote Posted by mopgoblin
...given any two forum members, the probability that one is homosexual is independent of that of the other.
Not with fett and Duckeh struttin' their stuff around here, no.
fett on 9/1/2007 at 13:26
Come 'ere you! :sly:
