Discuss 73. Skeptoid’s Brian Dunning Finds Global Consciousness Project Lacking

[Sandy]

Originally Posted by hoggworks

I'm still unclear what they think a "statistically significant" variance from random numbers would even mean. Why would our attention have that affect?

The speculate that it's akin to observations of photons which cause the waveform to collapse.

It's hard dealing with the GCP because almost nobody has the expertise in statistics to address it. I keep wanting to say something and reminding myself of that fact.

[FastEddieB]

Alright. First some essential background...

Miguel,

Thanks you so much for taking the time to summarize some of the basic math involved.

It reads easily and clearly to a non-mathematician like myself, and summarizes many of the thoughts I've had while following this debate.

The paradox is that if you flip a coin 1,000 times, you're most likely to get 500 heads and 500 tails, when compared to any other result. However, the chance of getting exactly 500 heads and 500 tails is very small. This may seem counter-intuitive, but is very easily demonstrated. The "bell curve" of possible results clearly shows the majority of possible outcomes is NOT 50/50.

Alex consistently seems to misstate this by implying that any deviation from 50/50 is significant. Its NOT significant - its to be expected in one direction or the other most of the time.

[alextsakiris]

Originally Posted by Miguel

Alright. First some essential background:
1: You cannot measure randomness.
2: Random does not mean uniform

By tossing a coin and converting the heads/tails to 1s/0s you might get these sequences:
1010011010
1111100000
Which one is more likely? The answer is, of course, that they are exactly equally likely.
Now someone might point to the apparent pattern in the 2nd sequence. Don't be fooled by that. If we discarded those sequences with "patterns" then the data would no longer be random because we made a conscious selection.
Depending on how hard we look and how the data is presented a pattern may be more likely than a non-pattern. (See: Clustering illusion)

One last thing to bear in mind is the importance of how the data is interpreted. Let's say you treat those binary sequences as numbers and convert them to decimal:
666
992
Suddenly the innocent first sequence turns into the ominous number of the beast while the 2nd sequence becomes inconspicuous.
Obviously this would change yet again if we were to reverse the conversion of head/tails into 0s/1s and so on...

Let's turn to the GCP now...

Please run this past anyone familiar with how data like this is normally analyzed. I think you'll find that the GCP is on very solid footing when it comes to statistics.

[DoctorAtlantis]

Originally Posted by alextsakiris

Please run this past anyone familiar with how data like this is normally analyzed. I think you'll find that the GCP is on very solid footing when it comes to statistics.

Alex:

What if we ran it by five different statisticians holding professorships at American universities. What if all five said that the GCP was not on solid footing - what would that mean? Would it mean that you'd switch to Brian Dunning's point of view? Or would you find ways to disagree with them? Or what if we asked ten statisticians and seven of ten agreed with Brian - would the other three be better statisticians? Or would the seven, as a majority, be right?

Before anyone runs anything by anybody I'm curious as to what it would take to dissuade you from agreeing with the GCP?

[Miguel]

Originally Posted by hoggworks

I'm still unclear what they think a "statistically significant" variance from random numbers would even mean. Why would our attention have that affect?

They have a lot of implicit assumptions in their project. I hope I managed to highlight how arbitrary it gets when one fishes in (random) data without a clear idea of what one is looking for.

I'm very unhappy about how I explained the GCP work. I tried to keep it simple and in truth it doesn't matter how exactly they do their analyzing. So I pretty much left it out which is far from ideal (and possibly misleading wrt how they actually do it).

I'll make amends sometime later this week.

[Sandy]

Didn't Nelson mention "auto-correlations" between the RNGs? I take that to mean that all or most of the RNGs in the network were trending in the same direction during certain periods. I thought that was interesting, but I don't know enough about it to comment. Anyone else want to?

Also, Nelson stated that they did comparisons of 9/11 to all of the surrounding days and found that none of them deviated from the expected value to close to that degree. Is there some reason that's not interesting?

Finally, I get that 100 1's in the previous example is not special or even likely, but it is my understanding that when you have observations of 1 billion coin flips, the actual mean value ought to be very close to the expected value of 0.5, or something is wrong. That something wrong could be the RNG, the analysis, or the understanding of how chance works.

If one were to observe the mean value of the network for several days and come up with a series like:

0.501 0.499 0.498 0.511 0.502 0.500 0.499

then isn't the fourth/middle one interesting in some way? Is the argument against GCP that the middle observation doesn't exist, or that it's not interesting?

[hoggworks]

Originally Posted by Sandy

If one were to observe the mean value of the network for several days and come up with a series like:

0.501 0.499 0.498 0.511 0.502 0.500 0.499

then isn't the fourth/middle one interesting in some way? Is the argument against GCP that the middle observation doesn't exist, or that it's not interesting?

Why would 0.511 be significant in this example?

If you were to toss a coin a billion times, you could easily get 1,000,000 flips in a row of heads; that wouldn't be indicative of anything remarkable, though, just of random chance.


I'm still unsure why they think that statistical deviations by an RNG is indicative of psychic abilities, though.

[hoggworks]

Originally Posted by Sandy

The speculate that it's akin to observations of photons which cause the waveform to collapse.

Who's observing the RNGs? Wouldn't, if this is the case, the more likely explanation be that the person watching the RNG results is causing the deviation, not the rest of the world?

[Sandy]

Originally Posted by hoggworks

Why would 0.511 be significant in this example?

If you were to toss a coin a billion times, you could easily get 1,000,000 flips in a row of heads; that wouldn't be indicative of anything remarkable, though, just of random chance.


I'm still unsure why they think that statistical deviations by an RNG is indicative of psychic abilities, though.

Huh. It surprises me that you think that.

Since p(1) for a coin toss is 0.5, wouldn't the probability of 1 million heads in a row be 0.5 ^ 1,000,000? That's a vanishly small number. I don't think you're going to see that happen just by tossing a coin 1 billion times. I think you're off in that relationship by (very) many orders of magnitude.

Edit: This part below relates to your question about why the 0.511 sample would be interesting.

This would be an application of the central limit theorem:

Flipping a large number of coins will result in a normal distribution for the total number of heads (or equivalently total number of tails).

Central limit theorem - Wikipedia, the free encyclopedia

If you toss one million coins an arbitrarily large number of times, and take the ratio of heads to tails, that ratio should converge to 0.5 as the number of tosses approaches infinity. That's essentially what the GCP does, and that's why the value above would be significant.

It is a property of normal distributions that the most likely value is at the center, and moving out from the center values are less and less likely. For example, for the million coin toss above, the center should be at 0.5 (500,000 heads), and both 0.0 (0 heads) and 1.0 (1,000,000 heads) will both be vanishly rare. This distance from the center is quantifiable as a Z-score (http://en.wikipedia.org/wiki/Standard_score) and probabilities can be assigned to different Z-scores.

In out million-flips experiment above, 0.511 probably corresponds with a large Z-score and thus small probability.

I think I'm rambling here, but this goes to the basic idea of the GCP.

[Sandy]

Originally Posted by hoggworks

Who's observing the RNGs? Wouldn't, if this is the case, the more likely explanation be that the person watching the RNG results is causing the deviation, not the rest of the world?

The speculation is that there is a correlation between conscious attention and chance at fundamental/quantum level. The GCP assumes that certain events are correlated with "more" conscious attention than non-event days, and therefore that chance is less chancy on those days.

Edit: The who is your question is the human race, who is watching TV, reading the news, talking to neighbors about the given event (or not)