Comments for Patricia Putt Million Dollar Challenge Test Results In! at http://www.randi.org/site , comment 1 to 42 out of 20 comments http://www.randi.org/site Sun, 22 Nov 2009 02:08:37 +0100 FeedCreator 1.7.3 http://www.randi.org/site/index.php/swift-blog/546-patricia-putt-million-dollar-challenge-test-results-in.html#comment-5951 [b] [s]Mark P, Said[/b] smilies/angry.gif Are you mental? How would lotteries work if that were true? [/s] [i]Ad hominum[/i] attack aside, it's been 30 years since I took a statistics course, and it wasn't my intent to get involved in a math argument. But this is not the same as a lottery, with 10! possible combinations. Each person has a 1 in 10 choice/guess. Each of the profiles has a 10% chance of being picked. Also, what if all 10 picked the same one as being uniquely theirs? Does that count as 10/10? I don't think the lady is any more psychic than my dog, but this protocol isn't designed to prove anything. I'm aware it was an agreed to protocol, but I'm saying it wasn't a good one. Most psychics and astrologers provide readings that are vague enough and have enough opposites ( "you are outgoing, yet also can be shy at times") so that most people can identify with something. Randi himself has demonstrated this by giving the same astrologers report to groups of people and them being convinced it was accurate. So, I'm a little suspicious that NOBODY picked one of those profiles. That, to me, is just as suspicious as 10 of them picking one. So, good for her for going through with it, and good for JREF for continuing to offer the $million challenge But, I still don't think this test proves anything, and the sCAM community will tear it apart, as we would if they presented a test like that. - NewCoaster Mon, 11 May 2009 20:29:08 +0100 http://www.randi.org/site/index.php/swift-blog/546-patricia-putt-million-dollar-challenge-test-results-in.html#comment-5887 The figures given above are a bit of a puzzle. There is an online probability calculator at http://stattrek.com/Tables/Binomial.aspx . There are ten trials, and p=0.1 for a successful guess. For five correct answers, the probability is p=0.0014880348, or 1 in 672. When Derek Ogilvie took a similar test for the full Million Dollar Prize, he was required to get six correct answers, with the probability at p=0.000137781 or 1 in 7258. - ianmacm Sun, 10 May 2009 09:28:43 +0100 http://www.randi.org/site/index.php/swift-blog/546-patricia-putt-million-dollar-challenge-test-results-in.html#comment-5850 I hope this clarifies the odds (which is to say I hope I'm right but it's been a long time since I last studied statistics - if I'm wrong, I'm sure someone will correct me.) Here are the [i]approximate[/i] odds for the ten possible outcomes: [code] odds of 0/10 right: 1 in 3 odds of 1/10 right: 1 in 26 odds of 2/10 right: 1 in 232 odds of 3/10 right: 1 in 2091 odds of 4/10 right: 1 in 18817 odds of 5/10 right: 1 in 169351 odds of 6/10 right: 1 in 1524158 odds of 7/10 right: 1 in 13717421 odds of 8/10 right: 1 in 123456790 odds of 9/10 right: 1 in 1111111111 odds of10/10 right: 1 in 10000000000 [/code] - kenhamer Sat, 09 May 2009 16:29:13 +0100 http://www.randi.org/site/index.php/swift-blog/546-patricia-putt-million-dollar-challenge-test-results-in.html#comment-5837 One in ten billion it is. The 10! only applies for lotteries etc, not this particular protocol. - Mark P Sat, 09 May 2009 14:01:33 +0100 http://www.randi.org/site/index.php/swift-blog/546-patricia-putt-million-dollar-challenge-test-results-in.html#comment-5742 The situation was not: Patricia Putt, write ten readings for ten people. Ten people, pick from one of the ten readings. Okay, bye. If this was the case, testers like Prof. Wiseman and Prof. French wouldn't be necessary, the JREF could have just put an ad out on craigslist for some testers. The random volunteers were a step in the testing protocol, but were not the be all/end all, the readings were most likely then compared to the volunteers to see how accurate they were by Wiseman and French. In any case, the simple fact that these two men were overseeing this testing process should lead one to imagine that the protocol was sufficient for the claim being tested, as part of their livelihood is, indeed, their scientific integrity. I understand being skeptical, but come on, at this point you're calling these men's integrity into question in an effort to 'get your point across', which is to assume you're smarter than they are. >:-P - UglyLikeMe Fri, 08 May 2009 18:05:00 +0100 http://www.randi.org/site/index.php/swift-blog/546-patricia-putt-million-dollar-challenge-test-results-in.html#comment-5739 kenhamer has it right. Each subject is picking one reading from ten, so has one chance of picking the right evaluation and nine chances of picking a wrong one. That's a one-in-ten (or 0.1) chance of being correct. Since there are ten subjects, you multiply 0.1 by itself ten times to get the probability of ten correct choices, which gives one in ten billion. - nervinjapan Fri, 08 May 2009 16:18:37 +0100 http://www.randi.org/site/index.php/swift-blog/546-patricia-putt-million-dollar-challenge-test-results-in.html#comment-5737 I think it's worse than 10! -- much worse, in fact. The 10! presumes that one person picks their reading, then the next only has to choose from the 9 remaining, then the 3rd person chooses from the remaining 8, and so on. But in this case, each volunteer was given all 10 readings, and asked to choosed their own. This means that 2 or more people could select the same reading ("two men say they're Jesus, one of them must be wrong"), but in each case random chance would give each and every volunteer a 1 in 10 chance of selecting the correct reading. Therefore, the true probability of all 10 volunteers selecting the correct reading, [u]by chance[/u] would be 0.1^10, or 0.0000000001 (perhaps more easily understood as 1 in 10 billion.) - kenhamer Fri, 08 May 2009 16:08:23 +0100 http://www.randi.org/site/index.php/swift-blog/546-patricia-putt-million-dollar-challenge-test-results-in.html#comment-5736 [quote]First of all, 0/10 is just as likely as 10/10[/quote] >:( Are you mental? How would lotteries work if that were true? The chance of 10 out of 10 is = 10! = 1 in 3,628,800. Yes that can happen by chance alone. One time in every three million or so. In which case the protocol specified the next step. [quote]This isn't the sort of test that's going to win support from the sCAM community who accuse JREF of being biased and making up their own rules[/quote] Since, like you, they haven't the foggiest idea of real testing of subjects. This was an [b]agreed [/b]protocol, so not only of JREF's making. - Mark P Fri, 08 May 2009 15:49:44 +0100 http://www.randi.org/site/index.php/swift-blog/546-patricia-putt-million-dollar-challenge-test-results-in.html#comment-5735 I also agree this was not a valid test, but am looking forward to the full report. First of all, 0/10 is just as likely as 10/10, and if she had scored that 10/10, the testers would certainly have mentioned that it could happen by random chance alone But you can't apply statistics for random events to something that is judged so subjectively. We don't know the pre test bias of the judges, but just as pro-sCAM are biased to see positive results, anti-sCAM may be biased to see negative results. Also as a very small study, it has very limited power./ A better test may have been for biographies and personality profiles done of the test subjects ahead of time, according to validated methodology. Then after the test have an independent panel of judges view both profiles and see if there were any valid hits. I haven't really thought this through for other sources of error or bias, but I think its a far better protocol than the one used. This isn't the sort of test that's going to win support from the sCAM community who accuse JREF of being biased and making up their own rules. - NewCoaster Fri, 08 May 2009 15:22:37 +0100 http://www.randi.org/site/index.php/swift-blog/546-patricia-putt-million-dollar-challenge-test-results-in.html#comment-5728 come see the MASK OF NOSTRADAMUS [url]http://www.thesupernaturalworld.com/forum/index.php?showtopic=24429[/url] randi..some little no-name...... - dennisjones Fri, 08 May 2009 10:24:36 +0100 http://www.randi.org/site/index.php/swift-blog/546-patricia-putt-million-dollar-challenge-test-results-in.html#comment-5722 ^ | Loon - Kuroyume Fri, 08 May 2009 08:51:07 +0100 http://www.randi.org/site/index.php/swift-blog/546-patricia-putt-million-dollar-challenge-test-results-in.html#comment-5719 [url]http://www.thesupernaturalworld.com/forum/index.php?showtopic=24429[/url] to see the REAL MASK OF NOSTRADAMUS - dennisjones Fri, 08 May 2009 07:58:32 +0100 http://www.randi.org/site/index.php/swift-blog/546-patricia-putt-million-dollar-challenge-test-results-in.html#comment-5716 visit [url]http://www.thesupernaturalworld.com/forum/index.php?showtopic=24429[/url] to see how we WON the Million Dollar Paranormal Challenge - dennisjones Fri, 08 May 2009 07:49:24 +0100 http://www.randi.org/site/index.php/swift-blog/546-patricia-putt-million-dollar-challenge-test-results-in.html#comment-5710 Instead of blindly accepting and cheering the results we expect (and maybe hope for) it's peer review extravaganza :D Look for that in the land of psychics, CAM etc, you'd have a hard time. - Blonk Fri, 08 May 2009 02:17:42 +0100 http://www.randi.org/site/index.php/swift-blog/546-patricia-putt-million-dollar-challenge-test-results-in.html#comment-5699 Since I cannot edit the post above, the zero out of ten result will occur approx 34.86% of the time, and the other results 65.1%. - Ian MacMillan Thu, 07 May 2009 21:11:46 +0100 http://www.randi.org/site/index.php/swift-blog/546-patricia-putt-million-dollar-challenge-test-results-in.html#comment-5698 Assuming that each of the ten test subjects can choose any of the readings, the probability of getting zero correct answers by guesswork is (approx) 1 in 2.87 or 65.1%. For five correct answers the probability is (approx) 1 in 672. See http://faculty.vassar.edu/lowry/binomialX.html for an online probability calculator. - Ian MacMillan Thu, 07 May 2009 20:52:24 +0100 http://www.randi.org/site/index.php/swift-blog/546-patricia-putt-million-dollar-challenge-test-results-in.html#comment-5686 If the psychic was merely claiming the ability to give accurate personality readings, I'd have called the test a farce... Consider a pool of 9 jerks and 1 good person. A 100% accurate personality-psychic would write a corresponding 9 negative readings and 1 positive. Human psychology predicts that the 1 positive reading would get picked almost every time (ironically, the 1 "good" person might, of humility, be the only one *not* to pick that). As such, we see a worst-case scenario where a 100% accurate psychic could reasonably be expected to scor 0 out of 10. In fact, this will be somewhat vindicated if, in the results, a particular description is picked 4 or more times (not expected by chance). NOTE: It's important to realize that she could be an accurate psychic and no very little of human psychology or statistics; and might therefore agree to a protocol without recognizing it's great potential for a false negative. So I do not consider her acceptance of the protocol to be a valid reaction to my objection. However, I retract that objection if the psychic felt she could predict concrete, objective facts about her subjects. If that were the case, she would have been best served to focus on simple, hard-and-fast calls like number of pets, number of syblings, names of deceased relatives, etc. Assuming she did so, and failed, I would agree that this was a solid demonstration of her lack of psychic ability. - baldape Thu, 07 May 2009 12:05:16 +0100 http://www.randi.org/site/index.php/swift-blog/546-patricia-putt-million-dollar-challenge-test-results-in.html#comment-5682 [i]If we judge these tests successful by results which exceed probably enough to indicate influence outside of random chance, should we also not say that a result significantly below probability also indicates outside influence of some sort?[/i] But, by pure chance there was a better than one out of three probability that zero matches would happen. Do you really consider that to be "significantly below probability?" - jadebox Thu, 07 May 2009 11:04:00 +0100 http://www.randi.org/site/index.php/swift-blog/546-patricia-putt-million-dollar-challenge-test-results-in.html#comment-5671 If we judge these tests successful by results which exceed probably enough to indicate influence outside of random chance, should we also not say that a result significantly below probability also indicates outside influence of some sort? - Griz Thu, 07 May 2009 08:53:04 +0100 http://www.randi.org/site/index.php/swift-blog/546-patricia-putt-million-dollar-challenge-test-results-in.html#comment-5667 Granted much of my probability theory is limited and please correct me if I'm wrong, but shouldn't the probability that at least one picked the correct reading be about 65%? Assuming probability that no one gets the right reading = .9^10 ~ 35%? - okmet Thu, 07 May 2009 07:02:22 +0100