An psych experiment on the OOT crowd....

[pshreck]

[ QUOTE ]


Im honestly dumbfounded as to how many people actually flipped coins.



[/ QUOTE ]

I would estimate their is a 50/50 chance of people who now claimed to have flipped the coins are telling the truth. Surely your testing would account for this. If someone didn't actually flip but posted in the firstplace, we already know they were being dishonest. The fact that you are asking them now whether or not they actually did it, even after explaining thats the point of the test, doesnt mean most, or even a few, will now tell the truth.

[CallMeIshmael]

[ QUOTE ]
It would've taken me literally 5 seconds to do 40 samples on my calc, how does that dumfound you?

[/ QUOTE ]

Most people arent you? [img]/images/graemlins/tongue.gif[/img]

[tbach24]

[ QUOTE ]
[ QUOTE ]
It would've taken me literally 5 seconds to do 40 samples on my calc, how does that dumfound you?

[/ QUOTE ]

Most people arent you? [img]/images/graemlins/tongue.gif[/img]

[/ QUOTE ]

True, but I suspect that a large majority of OOTiots are smarter and have better logic skills than me. This should've been obvious to most (who have these calcs anyways).

[CallMeIshmael]

[ QUOTE ]
I would estimate their is a 50/50 chance of people who now claimed to have flipped the coins are telling the truth. Surely your testing would account for this. If someone didn't actually flip but posted in the firstplace, we already know they were being dishonest. The fact that you are asking them now whether or not they actually did it, even after explaining thats the point of the test, doesnt mean most, or even a few, will now tell the truth.

[/ QUOTE ]

I was thinking about this, but I dont know if its quite that strong. Though I do agree that I cannot assume that everyone who came back and said I flipped the coins, actually did.

Simply because I feel people would tend to lie quite a bit less in the second situation. Partly because the "trick" part of the experiment is over. Also, because I sort of stated that lying was OK in the first part, when I asked them to fool me.

(though someone may wish to fool me by saying "I flipped" when they didnt and scored well away from 19.5 (or vice versa), I didnt really care that much about the predictions part, I just wanted to see the data)

[CallMeIshmael]

[ QUOTE ]
True, but I suspect that a large majority of OOTiots are smarter and have better logic skills than me. This should've been obvious to most (who have these calcs anyways).

[/ QUOTE ]

I think you might over estimate the % of OOTers who have access to a RNG, be it through a Graphing Calc, or programming skills. (also, they have to know how to use the RNG on their calc... I have one, and dont know this off hand, and I sure as hell wouldnt look it up for some punk poster)

Though I could be WAY OFF on this one, because this is a pretty eclectic group.

[daryn]

please mention that you own a graphing calculator again. please.

[tbach24]

[ QUOTE ]
please mention that you own a graphing calculator again. please.

[/ QUOTE ]

I actually don't. I got lent one by the school cuz some faggot at my correctional school stole and broke mine and I couldn't prove it past that I saw him do it.

[NoTalent]

thhtthhthh
hhhhtttthh
thhttthhht
ttthhhhtth

[DougOzzzz]

[ QUOTE ]
[ QUOTE ]
By the way, given a sufficiently large sample size (maybe >1 million), I think that even the lists made by random number generators could be spotted from those made using a "fair" coin. I've done tests on this, and I get some pretty strange results. i.e. 80% of flips after 7 consecutive "heads" result in another heads (over 1000 flips) and things of that sort.

[/ QUOTE ]

What language?


I used C++ to test some things, and afer a million was basically dead on in flipness.

[/ QUOTE ]

I tested this in Excel using VBA - just tried it again, with 20,000,000 flips.

Heads came up 49.999 % of the time, there were 9,993,914 "runs", vs. an expected 10,000,001. This is 2.72 standard deviations away from the expected runs. It's not way off, certainly better than any human could do.

However, it's the distribution of the "runs" that makes it really odd. There were 0, absolutely NO streaks above 10 or greater for either heads or tails. There should have been nearly 10,000 of these streaks.

There were 225,000 streaks of EXACTLY 6 heads or EXACTLY 6 tails - vs. an expected 156,000.

After flipping 5 heads in a row, the chance of a 6th heads was 61.96%. After the 6th head, the chance was just 40.35%. After the 7th, 43.23%. After the 8th, 7.7% (not a typo, 2174/32321).

The results for heads and tails were almost identical, with the same odd patterns. So clearly something is wrong with VBA's RNG.

[CallMeIshmael]

Wow that is odd.


I'll run the same tests tomorrow, and see what I come up with.