What do any resident /boo/ lads make of this study?
http://phys.org/news/2016-01-equation-large-scale-conspiracies-quickly-reveal.html >Equation shows that large-scale conspiracies would quickly reveal themselves
>If you're thinking of creating a massive conspiracy, you may be better scaling back your plans, according to an Oxford University researcher.
>While we can all keep a secret, a study by Dr David Robert Grimes suggests that large groups of people sharing in a conspiracy will very quickly give themselves away.
You mean 'things you already believe'. I'm not saying your belief is wrong, but to say you already know something without a decent base of justification, is the height of arrogance.
Surely that is what they would want us to think. Dr Cancer-Research is clearly just covering his tracks.
More realistically though, his methodology appears to use conspiracies that were exposed which will obviously lead to a result where any prediction on present conspiracy theories will have a 100% failure rate. We know of course that states can keep secrets for long periods of time from those that are revealed decades later.
It's a nice idea but complete bollocks in reality.
This equation has no more factual basis than the equation made up to "prove" that the third Monday in January is the most depressing day of the year. It's barely any better than the equations which pick random numbers from the bible to predict the end of the world.
Things like this are just cherry-picking numbers to arrive at the answer you want.
I was thinking that. How many thousand people must have known what Snowden knew and for how long and never said a word? What if Snowden never got that job, would the public ever know? How many morally questionable things have GCHQ and the NSA been up to for decades that the public have never heard of?
>>4301 >How many thousand people must have known what Snowden knew and for how long and never said a word?
Not many. Organisations like GCHQ and NSA are very good at compartmentalisation. You might not know what's going on in the next office unless you need to. Snowden was able to know what he knew because he was in a privileged position as a sysadmin.
They're also good at preventing cell formation, so as to ensure that a small group can't go rogue. This carries the rather dark implication that everything within the organisation happens with the proper authority and knowledge of the higher-ups. When the NSA's telephone programme was found to be partly illegal, you can be sure that whoever signed off on it knew what was going to happen and knew that it was partly illegal.
These organisations get away with this because they are in a unique legal position. They can rely on their secrecy to prevent disclosure and effectively lie in open court.
Most of what Snowden 'revealed' was already public knowledge, in large part due to the work of Duncan Campbell. He's largely forgotten today, but he was tried for espionage in 1978 for revealing the mere existence of GCHQ. He went on to reveal the existence of Echelon and numerous other bulk interception programmes.
The Snowden files revealed a lot of technical details and brought attention to the issue, but they didn't fundamentally change our understanding of the scope of mass surveillance.
>>4303 It merely brought it to light and reminded everybody that it's still a thing. Mostly that between the 5-eyes almost all unencrypted web communication is being tapped.
>>4298 >Things like this are just cherry-picking numbers to arrive at the answer you want.
The specific numbers aren't important, it's the orders of magnitude. You only need a back-of-the-envelope calculation to show that secrets held by hundreds of thousands of people aren't likely to last long.
This is something I keep coming back to when I think about stuff like this. Especially in the post-9/11 security hysteria, there was a lot of back and forth about the idea of privacy and whatnot. Those tedious posts on the primordial message boards and early social media saying 1984 WAS A WARNING NOT AN INSTRUCTION MANUAL and the equally tedious arguments in return about being paranoid and that if anything the government is surely too inept to construct such a vast surveillance state. Which side turned out to be right?
Anyone who still has an essential faith in the government and it's intentions, knowing what we know to be fact today, has got to be a bloody moron.
The worrying thing is that everything gets recorded forever, so that if a halfway competent totalitarian state gets into power in the future, they'll know who to stick in the gulag.
>>4310 What is interesting about this though is how in the past few years the surveillance has gone from a debate over personal privacy to a debate over surveillance upon the state apparatus itself. The debate has moved on from 'should we have CCTV?' to 'how can we make sure the surveillance is done fair and accountably and how can we employ it to keep an eye on the police?'.
Yeah I guess even just having lurked here will mean getting fucked over on the basis of some teenlad rant about killing posh is and bumming the queen five years ago.
>>4316 The real question now is whether the paper was part of the grand conspiracy, or whether the revelation that it's flawed is a deliberate limited hangout to put people off digging deeper?
>>4316 I have to be honest that I didn't really trust the writings of a Vice columnist over an Oxford mathematician, but on reading the paper the error is plain to see in the first two equations of the paper:
L = 1 - exp(-t \phi) (1)
\phi = 1 - (1 - p)^N(t) (2)
Having a time parameter t in (1) is completely wrong on dimensional grounds, since \phi is a dimensionless quantity and exponentiating any quantity with dimension is unphysical. Secondly, in (2) he is giving time-dependence to the rate of events (in this case, the rate of whistle-blowing) which violates the key assumption of the Poisson distribution (i.e. that this quanitity is constant). Both these errors together lead his paper to have a nonsensical plot of cumulative probabilities.
As far as I can tell, the best you can do is assume the number of conspirators to be constant over some finite period and apply Poisson statistics to each period seperately. A quick calculation shows that for constant conspirators the probability of failure tends to 1, whilst for exponentially decaying conspirators the same probability tends to some finite value < 1.
Also, while you have a decreasing component representing what I presume to be people dying off, have you accounted for an increasing component representing new people being added (new hires, gossip, etc.)?
>>4320 I had a bit of a failure of English there apparently.
No, I've not attempted to do anything significantly more complicated than was in the original paper. Though the decay doesn't necessarily need to be conspirators being killed off, just of them being removed from positions where they have access to the level of information needed for a credible whistle-blowing.
This has nothing to do with maths though. Being "exposed" has nothing to do with actually uncovering a conspiracy There are literally hundreds of people who claim to have been involved in conspiracies of all stripes. Depending on how much of a nutter they are or how believable the story, it's incredibly easy to discredit these would be whistle-blowers, and then that's the end of it.
Someone comes out and tells us everything about how the government assassinated so and so and with a little encouragement everyone rolls their eyes in synchronisation, "What a nutter, just trying to sell a few books isn't he..." meanwhile they actually did do it but nobody believes it regardless and this little piece of maths doesn't make a difference.
Why are you pigeonholing 'maths' as a single thing? Everything has everything to do with mathematics, it's just a matter of having enough data and looking hard enough for a pattern.
To your second point, I think this is why we should differentiate between credible whistle-blowers such as Snowden who publish direct evidence, and any random person who can and will confess to anything for any number of reasons.