James Randi Comments on StereoPhile (1 Viewer)

[Brian L]

Thats because you are too damn sensible, John!

Your reviews in WSR are so free of voo-doo and other non-sense that it amazes me that they still see fit to publish your stuff

BGL

 

[Lee Distad]

This is of course absolutely fascinating.

The end result will, you understand, be lots of fulminating invective in the pages of these magazines, but exactly zero takers on Randi's offer.

 

[RobertR]

I've just had a well known person bet me that he or a person of his choosing can tell the difference between cables 75% of the time.

 

[Seth_L]

Why only 75% of the time?

 

[RobertR]

I think he's leaving himself some wiggle room.

 

[Yogi]

quote:Why only 75% of the time?

 

[Jack Briggs]

Correct -- no one is going to take Randi up on his offer. They never do, no matter how many ghosts, UFOs, alien abductions, Big Foots (Big Feet?), or vacuum-tube amps they've experienced.

 

[Yogi]

BTW James'challenge doesn't cover tube amps I asked him specifically if he would cover tube/ss amps and he said he only covers silly/absurd claims. I wanted him to elaborate if amps would come under his definition of silly claims but haven't heard back from him so far.

OK I'll let you guys get back to your hillbilly bashing ritual.

 

[Chu Gai]

But there's a 10,000 challenge on that. Just think Yogi, you can get a month of fill-ups!

 

[Yogi]

And we all agree (even some of the die hard objectivists like RobertR and RichardHOS) that the sole purpose of that challenge was for Mr. Clark to not lose his 10K. Period. So that challenge is pretty much another one of the cons.

P.S. If you like I could dig up the impressions of your fellow objectivists on RC's 10K challenge from a thread few months ago.

 

[Yogi]

BTW, I did receive James' reply on my amp question a few minutes ago:


It feels good that JR and I agree that amps are out of the paranormal claims regime.

OK guys sorry again for the interruption. Please continue your ritual.

Let the beer flow.

 

[Chu Gai]

Regarding RobertR's scenario of 75% right answers...

Let's examine the claim of a 75% in terms of probability and with the contraint that the number of attempts must be a multiple of 4. This makes it easier for the claimant to exactly hit the magic number of 75% correct guesses...errr...identifications

If we assume that the true probability of correctly identifying a cable is 50% (in essence a coin flip), then the probability is 0.5. Therefore, what is the probability of correctly identfying it 3 out of 4, 6 out of 8, ...12 out of 20? Looking at it as a coin toss, one can look at this as saying what are the odds of you tossing 3 Heads out of 4 tosses?

Using probability this can be computed relatively easily as follows:

A. Determine the # of ways that one can get 3 Heads.
B. Determine the # of total possibilities.
C. Determine A/B.

Let us assume that n is the number of tosses and that r is the outcome. These are the equations.

A1. n! / [(n-r)! r!]

B1. 2^n

For 4 tosses, the odds of getting 3 Heads are: {4!/[(4-3)! 3!]}/2^4 = 0.25 or 25%

For 8 tosses, the odds of getting 6 Heads are: {8!/[(8-6)! 6!]}/2^8 = 0.109375 or 11%

For 12 tosses, the odds of getting 9 Heads are: {12!/[(12-9)! 9!]}/2^12 = 0.0537109375 or 5.4%

For 16 tosses, the odds of getting 12 Heads are: {16!/[(16-12)! 12!]}/2^16 = 0.02777099609375 or 2.8%

For 20 tosses, the odds of getting 15 Heads are: {20!/[(20-15)! 15!]}/2^20 = 0.0147857666015625 or 1.5%

 

[Yogi]

The number of attempts aren't specified by either Robert or the claimant. So we dont exactly know if he would guessing or identifying

So in RCs test (24 attempts) if someone was able to correctly identify amp A from B 75% of the time what would your analysis say? Are the amps different or was he just guessing?

 

[Chu Gai]

Well we need a sufficiently large # of samples. After all Yogi, you could claim to be able to toss heads 75% of the time and do 4 tosses. I tossed that out for Robert's consideration. 12 tests seems about right to balance time and results.
The rest of the question involves too much effort right now for a thoughtful response.

 

[Yogi]

Good, so if you are able to pick out amp A from amp B 75% of the time out of 12 attempts then that is statistically significant to say that those two amps sounded different. That's all I wanted to hear. Thank you my friend.

 

[Chu Gai]

Depends on the level of signficance you want to achieve. Bias control is very important and when one gets an unexpected result, one examines the results and methodology more carefully. Otherwise you wind up with Cold Fusion, or that funny stuff that Tiffenbaum from Linn came up with, eh? Science after all should be rigorous.

 

[Yogi]

quote:...when one gets an unexpected result...



But expecting a certain result beforehand could lead to a bias in itself.




quote:Science after all should be rigorous



I never doubted that.

 

[BruceD]

For statistical significance to really mean anything, not only does the statistical result of the experiment have to reach a minimum level of correct responses, but the sample size (number of participants for experiment) is generally considered to be inadequate if it's below 100 in size.

For me this means one person getting 75% right answers is just as likely to be lucky guessing.

 

[RobertR]

Just for everyone's info, I suggested 16 trials to the guy, and briefly outlined the double blind methodology. I said that the comparison would be with ordinary 12 gauge wire, and that a simple test would be conducted to make sure that the "high end" wire isn't significantly varying the level. It's been two days and I haven't heard back from him. I'm starting to think he really isn't interested in the challenge, which wouldn't surprise me. We'll see.


I'd say a 97.2% chance that he won't get the 75% from sheer luck is good enough odds.

 

[Chu Gai]

That's why I'm sure Robert will call Jesse Jackson to monitor everything