JE does no better than a person flipping a coin?

[CFLarsen]

I thought Clancie made a very important admission in the thread "A count of JE's hits on Larry King Live" (http://www.randi.org/vbulletin/showthread.php?s=&threadid=24032):

Originally posted by Clancie
I find 50% a very probable hit rate for JE.

JE does no better than a person flipping a coin?

What Clancie is suggesting is even better odds than Andrew Ersby's "Psychic Dice" (http://www.skepticreport.com/funnies/psychicdice.htm), where we have 6 choices. With a coin, we have a 50% chance of success, since each question is decided 50/50, and in the end, we get a reading that is enough to persuade Clancie of paranormal powers.

A very important admission indeed. Amazing how little it takes.

[Martin]

That's not really fair, Claus. A coin will get the answer to 'is the person I'm thinking of male?' 50% of the time. I'd like to see you answer the question 'what's my dog's name?' with a coin. Whether or not a 50% hit rate is impressive depends upon the questions.

[CFLarsen]

Originally posted by Martinm
That's not really fair, Claus. A coin will get the answer to 'is the person I'm thinking of male?' 50% of the time. I'd like to see you answer the question 'what's my dog's name?' with a coin. Whether or not a 50% hit rate is impressive depends upon the questions.

Oh, it's very fair.

Cat or dog?

Dog.

Dog's name is Fluffy?

No.

Spot?

Yes.

Cold reading is working from what the sitter tells you. We are not talking about a string of unrelated questions here. Each question leads to the next, depending on the answer for the first.

On the flip side (sorry, I had to!), we don't get to edit the tapes like JE does. That just makes his performance worse provided Clancie's estimate is correct.

I know that she doesn't have any real data to base it on, only her memory, which is - to put it mildly - flawed in favor of JE. But it's her claim, and she claims a superior knowledge compared to anyone else here, neofight as a possible exception.

[Martin]

Sure, you can use binary questions to progressively narrow down any search space until you hit the right answer. But the probability of a hit at each stage is not 50%, it's the volume of the search space corresponding to the answer you're asking for divided by the volume of the entire search space. So, if your question cuts the search space exactly in half, we expect a hit-rate of 50%. But generally it won't. Take an old favourite 'is the sixth month, or the sixth of a month, significant?'. A binary question, which can be answered only 'yes' or 'no'. But the probability of a hit is manifestly not 50%.

[CFLarsen]

Originally posted by Martinm
Sure, you can use binary questions to progressively narrow down any search space until you hit the right answer. But the probability of a hit at each stage is not 50%, it's the volume of the search space corresponding to the answer you're asking for divided by the volume of the entire search space. So, if your question cuts the search space exactly in half, we expect a hit-rate of 50%. But generally it won't. Take an old favourite 'is the sixth month, or the sixth of a month, significant?'. A binary question, which can be answered only 'yes' or 'no'. But the probability of a hit is manifestly not 50%.

The further a reading goes, the more precise the hits seem to be. This is true both with JE and a cold reader, and this is because we narrow down the choices. We lead desperately common lives: "House, Dog and Volvo" is how we describe it in Danish. :) Once we have e.g. a house, we also have a spouse. Then, we have at least some marital problems, and certainly family quarrels.

Overall, the hit rate could easily be 50%, using a coin - and, of course, some knowledge about demographics.

Are you saying that JE has a hit rate of more than 50%?

[Lurker]

Well, I have no idea what JE's hit rate is.

It is not a binary situation but it swings both ways. For example, JE may say something like

"I see an "R" name connected to you." - In this case, the odds of an actual "R" name connected to you may be close to 100%.

Without mapping out the probability tree with each question having myriad answers there is no easy way to know how well JE is doing.

But if I had to guess, he is doing far worse than 50%.

Lurker

[Martin]

Originally posted by CFLarsen
The further a reading goes, the more precise the hits seem to be. This is true both with JE and a cold reader, and this is because we narrow down the choices'Both with JE and a cold reader'? There's a difference? :DOverall, the hit rate could easily be 50%, using a coin - and, of course, some knowledge about demographicsIf you used only binary questions with equal probabilities, yes. But that's hopelessly inefficient. Cold reading makes much better use of the available information. A coin tosser will hit the right conclusion eventually, but a cold reader, or even a real tosser like JE, will hit it in less time, using more specific questions and therefore seem much more impressive.Are you saying that JE has a hit rate of more than 50%?Hell, no. 50% is pretty generous, I'd say. And, of course, it's important to look at what he gets right and what he screws up, not just the averages.

[CFLarsen]

Martinm, Lurker,

So, using a coin, we'd actually get more hits than JE?

Thought-provoking, isn't it? :)

[Martin]

If we construct our questions correctly, we could certainly maintain a hit-rate of 50% (or thereabouts), and furthermore we would eventually be able to deduce any arbitrarily specific scenario from these questions. But it would be painfully obvious what we were doing, and how we were doing it. So, while we'd beat JE, it would only be by blatantly extracting information from the sitter in such an obvious manner that no one could possibly miss it.

Of course, what he does is painfully obvious too...to us, at least.

[Thanz]

Originally posted by CFLarsen
Martinm, Lurker,

So, using a coin, we'd actually get more hits than JE?

Thought-provoking, isn't it? :)
You don't actually think this do you?

As Martinm has explained, the guesses that JE makes are not binary 50/50 guesses. To take the simplest example, when he calls for a "J" connection to the "older male" energy, it is not a 50/50 proposition - we have seen that J makes up about 13% of the population.

[Lurker]

Originally posted by Thanz

You don't actually think this do you?

As Martinm has explained, the guesses that JE makes are not binary 50/50 guesses. To take the simplest example, when he calls for a "J" connection to the "older male" energy, it is not a 50/50 proposition - we have seen that J makes up about 13% of the population.

Yes, Thanz, but that percentage is only for 1 person. IF JE said "Your dead father starts with "J"" then it would apply.

But JE says, "older male with "J""

So if you know say 8 older males the odds of NONE of them starting with "J" are (1-0.1336)^8 = 32%. As you can see, JE has a 78% of getting a hit by not being specific on the relationship (or friendship etc).

Lurker

[Walter Wayne]

Better yet, he doesn't just ask one person, he often says from somewhere in this area. How many older males are we dealing with.

Walt