A Test for Sandsted

[Jean310]

Here are the results of one analysis I did, taking into consideration that each individual date is really a range of 7 dates, with the exception of the dates at (or near) the extreme ends of the set S{40}. In this case, rather than using 1/40, I used 7/40. Pbsuccess = 0.175 And, I progressed the Pb as each trial was completed.

0 Correct, Pb .14606, 1 in 6.85
1 Correct, Pb .30983, 1 in 3.23
2 Correct, Pb .29783, 1 in 3.36
3 Correct, Pb .18037, 1 in 5.54
.
.
6 Correct, Pb .00551, 1 in 181
7 Correct, Pb .00094, 1 in 1061

(Note, these Odds are what could be expected from a Chance result, no dowsing involved)

There are other ways to analyze this problem, this is just one, but one interesting thing that can readily be seen, is that the Odds for getting 1, 2 or 3 correct (by pure guessing) are better than getting Zero correct.

I'm still looking at this..............

Also, I might be mistaken here, but doesn't TBS usually attempt to Pre-test individuals such that they must perform successfully at a significance level where the Odds of a Chance Result would be 1 in 1000 or better?

 

Amazon Forum Fav 👍

Diving for Pleasure and Treasure - Grab it through Amazon!

[Jean310]

LMAO!!! NOW we are finally attempting to establish the sample group and odds of guessing.....based on the guessing skills of "Jean".
Sandsted.....you are getting this...right? LOLJudy

LYAO? I don't think so. Looks to me like it's still there.

​

.....single mom, huh? What a surprise! (I didn't need no stinkn' dowsing rod to figure that out......)

 

[Jean310]

Come on, Jean....you can do better than that.
Answer the question......leave your bitter catty remarks for someone who might actually be offended by them.

Answer the question...please. What protocol did you use?

Judy

Meow!​

Doncha jus luv it honey?

​

I'll come back and play with you later...... right now I have other stuff to do. Gotta admit, this is a lot more fun though.....

 

[ClonedSIM]

Btw Jean, how was your "test" for guessing set up? Did your husband hide the coins under an egg or a bowl? Did you wrap the coins in card stock and mail them to yourself? Just exactly what protocol was used in YOUR test?


Judy

I'm curious, Judy. Why is Jean's testing protocol so vital? It would be impossible for her to reproduce exactly the test between Carl and Sandy. I don't know what her protocol was, but it would seem that she only self-tested herself to get some sort of baseline, but not one that would apply with Sandy's test, since the procedure and variances have already been established. It's just more for curiosity and informational sake, I would believe.

Unless Carl wanted to send a similar batch of coins to Jean for a true baseline, that is.

 

[Carl-NC]

I'm not sure why the exact protocol is so important, as there are a number ways to reasonably replicate the statistics of this test. You can even do all by yer lonesome. Take 40 pennies, dates don't matter... number them 1-40 with a marker. Put them in a bag, mix 'em up, draw 10. Write down the numbers IN ORDER. Those 10 become the target numbers. Put the coins back in the bag, mix 'em up, draw 10 more. Write the numbers IN ORDER next to the first 10 numbers. See how many correlate within the parameters of the test (+/-3 in this case). Rinse and repeat until you get 6 or more "right".

By far, the easiest way of trying out the statistics of this test, is to use a software-based RNG. There is a web-based one somewhere on the Internet that is perfectly suited, but I'll have to find it again.

- Carl

 

[Jean310]

:::chortles:::
A "non-answer" is what you get when you try to goad someone into giving an "opinion".
"Avoiding an answer".. is when someone asks you a specific question...such as "what protocol did you use"....and instead of answering you make nasty personal remarks intended to offend and then run away.
That clear it up for you?

Judy

I told you I'd come back and play with you.... and I did.

Actually, Judy, I’d be very happy to give you the details of how I ran my five test runs. The thing is, when you ask folks for answers and info and do it with your “crusty attitude” showing, I would guess that more often than not the responses you get would also come with some crustiness and really not be what you were looking to receive. I’m guessing this isn’t exactly news to you, since I can’t be the first one to mention it to you.

Now, are you looking for a reasonable answer, or are you just interested in obtaining some cannon fodder to kick around on the forum between you and Real Deal, no matter what I answer?

 

[Jean310]

I'm not sure why the exact protocol is so important, as there are a number ways to reasonably replicate the statistics of this test. You can even do all by yer lonesome. Take 40 pennies, dates don't matter... number them 1-40 with a marker. Put them in a bag, mix 'em up, draw 10. Write down the numbers IN ORDER. Those 10 become the target numbers. Put the coins back in the bag, mix 'em up, draw 10 more. Write the numbers IN ORDER next to the first 10 numbers. See how many correlate within the parameters of the test (+/-3 in this case). Rinse and repeat until you get 6 or more "right".

By far, the easiest way of trying out the statistics of this test, is to use a software-based RNG. There is a web-based one somewhere on the Internet that is perfectly suited, but I'll have to find it again.

- Carl

Judy - Carl is correct here. Naturally, there is an intent to remove the human element from "trying out the statistics of this test". We are not dowsing here, so there is no human element that has to play into the equation.

Why do you think there should be?

What am I missing here?

Can you reword or rephrase your concern that "trying out the statistics of this test" without a human element is wrong?

_________________________________________________

The whole purpose of me performing a few test runs, with conditions and constraints similar to the Sandsted Date Dowsing exercise, was for me to observe how my test run data would compare to the calculations I made for the Probabilities associated with this problem. The conclusion is, my test run results match almost exactly with what the calculated Probability of this problem would yield. That said, had I of run the test 1061 times (instead of just 5), I could very likely have seen an instance when I would have gotten 6 correct, with no dowsing involved whatsoever.

 

[ClonedSIM]

I'm curious, Judy. Why is Jean's testing protocol so vital? It would be impossible for her to reproduce exactly the test between Carl and Sandy. I don't know what her protocol was, but it would seem that she only self-tested herself to get some sort of baseline, but not one that would apply with Sandy's test, since the procedure and variances have already been established. It's just more for curiosity and informational sake, I would believe.

Unless Carl wanted to send a similar batch of coins to Jean for a true baseline, that is.

Jean, is a her?? Is that a fact?? Dell

Fact enought for my response, Dell.
If I go through and substitute the word her with his would it change the jist of my post?

And since when have you concerned yourself with facts, Dell?!?

 

[ClonedSIM]

The whole purpose of me performing a few test runs, with conditions and constraints similar to the Sandsted Date Dowsing exercise, was for me to observe how my test run data would compare to the calculations I made for the Probabilities associated with this problem. The conclusion is, my test run results match almost exactly with what the calculated Probability of this problem would yield. That said, had I have run the test 1061 times (instead of just 5), I could very likely have seen an instance when I would have gotten 6 correct, with no dowsing involved whatsoever.

KA-POW!

 

[Jean310]

It doesn't matter what attitude is showing...crusty...or catty....

That's where you are wrong...... you get back what you sow, and you always will.

there was a question...and you ran off without answering it.

Sorry honey, but some of us have other things to do in life, and living on this forum is not one of them.

Now that you have had the time to come up with one...and so graciously offer it....sure, lay it on us.
I can only use cannon fodder when cannon fodder is given....

Judy

No........ I really haven't written anything up that wasn't already a fact. I completed the test runs several days ago, and as Carl has already pointed out to you, there are probably thousands of different ways to "try the statistics out". My way would not be unique or better than any others someone could think of doing.

 

[Carl-NC]

By far, the easiest way of trying out the statistics of this test, is to use a software-based RNG. There is a web-based one somewhere on the Internet that is perfectly suited, but I'll have to find it again.

There you go again...taking the "human" element out of the equation.

I would like to see your computer attempt to dowse.

The last few posts have been about statistics, not dowsing. You can quite easily demonstrate statistics without involving dowsing, so I've intentionally excluded dowsing to show that the statistics are quite easily demonstrated without involving dowsing. These results1 can then be used to compare against someone who claims they can dowse.

You could do it, too, were you interested in objective testing. No dowsing required.

- Carl

1 Or, if you believe the field of statistics is on pretty sound footing, you can forgo the experiment, and compare the dowsing results to calculated odds.

 

[ClonedSIM]

Dowsing is a human ability (or non ability).....comparing it to statistics gathered with a machine, without a human sample study for establishing the guessing part of the equation.....would be flawed.

Jean already did this, but you didn't like it.....

 

[ClonedSIM]

As for your "test run", Jean.....since you are so fond of using your superior intelligence as a comparison to the poor "high school dropouts"....then by your own words....how could your personal guessing ability be compared to their guessing ability??

Why would a dowser have to guess in the first place? I thought they dowsed?

 

[Jean310]

Dowsing is a human ability (or non ability).....comparing it to statistics gathered with a machine, without a human sample study for establishing the guessing part of the equation.....would be flawed.
Judy

Well, I don't agree with you.

Yes! Definitely dowsing is something that can only be done when a human is holding a dowsing gadget. Lots of dowsers have proved that many times over.

But, Judy, Statistics and Probablities are nothing more than measuring tools. It matters not if we run the statistics by hand or use a computer, the answers will be the same.

However, if it makes you feel better, I did not run my 5 tests with a machine of any sort. They were done with 40 different physical objects, hand-picked random selections (much like Carl's coins in a bag), and pencil and paper.

 

[ClonedSIM]

Dowsing is a human ability (or non ability).....comparing it to statistics gathered with a machine, without a human sample study for establishing the guessing part of the equation.....would be flawed.

Jean already did this, but you didn't like it.....

Hmmm....you didn't read the entire post...did you, AF?

I'm wondering now if I should group you in with Art, Judy.

Let's see. You don't like the statistics about guessing that machines spit out, and you told us this when Carl suggested it.

But you don't like the statistics about guessing that Jean gave us, for one reason or another, and you told us this when Jean gave her report.

And then you suggest that a skeptics guesses shouldn't be included since.....well.....why was that again? You just don't like skeptics, or did you have some rational explanation?

Carl gave the stats concerning guessing, and Jean ran an independent study that confirmed his results. Feel free to run your own battery of guessing tests to see if you reach a different conclusion. For that matter, Sandy could keep a different sheet of guessing responses right along with his dowsed responses.

 

[Jean310]

I am not requiring that you agree with me, Jean.

That's good, at least you understand I have that prerogative.

And, yes, Statistics and Probabilities are useful tools for measuring.....you just need to be sure you are measuring all the components that affect the test....such as the "guessing" data.

Judy

Okay. I think you've seen the various characteristics of the statistical test enough to probably know them by heart. Given the statistical test does not involve the practice of dowsing, but only is done to show the researcher a comparison of actual data vs calculated expected results; please explain what you mean by your statement, "the guessing data affecting the test"?

What did we miss here that could skew the results in such a way as to render them worthless or meaningless?

 

[Carl-NC]

Dowsing is a human ability (or non ability).....comparing it to statistics gathered with a machine, without a human sample study for establishing the guessing part of the equation.....would be flawed.

That might be true, if (a) we were not comparing dowsing to statistical guessing, or (b) statistics just didn't work. But we are comparing dowsing to statistical guessing, and statistics does work.

So, yes, we could perform a large human sample study to establish what the odds of guessing would be. But it's no more necessary than having the referees in yesterday's playoff games perform large-scale coin flip studies in order to ensure the actual coin toss event has 50-50 odds.

- Carl

 

[Jean310]

Judy, here is how I did my 5 tests -

The Sandsted Date Dowsing Exercise has the following characteristics and parameters, as I understand them. Since the act of dowsing is not a concern of mine, I am only interested in duplicating the important characteristics of the exercise. My whole point was to see if the results of pure random picks will compare to my calculated Probabilities.

The US Mail system involvement has no bearing on either Sandsted’s exercise or mine. It is required in his, to get the coins back and forth, but it is not a dependent variable.

The fact that Sandsted is dowsing coin dates IS important to him, so he needs actual coins, and they need to be sealed in such a way that he cannot visually see the date. Since dowsing is not involved in my test, I don’t have to use actual coins, but can simply utilize a like group (a Set) of 40 unique items. It could be 40 different buttons, or 40 different numbers written down on separate pieces of paper, or it could be 40 different playing cards from a standard deck of 52.

I chose 40 playing cards and assigned each of them a number 1 to 40. This way I could keep track of the agreed upon range of plus or minus three numbers. The pack of 40 cards was shuffled extensively, face down. Then, I randomly selected a subset of 10 cards from the pack of 40, obviously still face down. This subset of 10 cards remained face down, and I arranged them into two rows of 5 on the table in front of me. These 10 are now my target cards. Then, on a piece of paper, I drew two rows of squares, 5 each. In each square I randomly picked a number (from the range 1 through 40) and wrote it in the square. [Here is where I could’ve or should’ve used a different method for getting these numbers. Ideally, I should have used a 40-sided die, or a PRNG that emulates a 40-sided die, or just put 40 numbers in a hat and select 10 of them*.] Next, I would turn over the first card in the first row, and compare it to the number I had written down for that position. If it was plus or minus 3 of the number, I scored it as a Hit, otherwise it was a Miss. I did this for the remainder of the 10 cards, and recorded the results for those 10 trials, constituting 1 test. I did the same test 4 more times, each time thoroughly shuffling the pack of 40 before selecting the subset of 10. The results of these 5 tests have already been posted, further down in the thread.

BTW, Carl has already given a much simpler rendition of the same kind of test, I just didn’t bother to find 40 coins and a bag to put them in, and the playing cards were handy.

(*) with replacement

 

[Jean310]

I have no problem with your statistics, in and of themselves.
(Or the constant repetition of them... )

Constant repetition? Sheesh. Did you or did you not ask me how I performed the test?

The problem is that you are testing a human being using "guessing" as a measurable quantity....yet you do not have a "measure" of "human guessing" (a human sample group) on which to base your statistics. Therefore, your statistics are skewed and unreliable for determining a proper level of "confidence" on any results obtained from this "test".

There is no problem and you are totally in error with that statement.... you keep repeating. I did not, nor do I intend to test a human being using "guessing" as a measuring device.

A human being (Sandsted) made a "claim" concerning what he believes to be some sort of talent related to dowsing. SANDSTED IS NOT GUESSING! He is going to dowse for the dates; there's a difference. Dowsing is NOT guessing, according to those making the claims. IF there is a way (and I believe there is) to evaluate his dowsing claim, then it amounts to comparing his ACTUAL results against those that we already know to be statistical random results. It's really very simple. If his results are repeatable and significantly better than the statistical random results, then that would constitute very compelling evidence to support his claim.

I have no idea why you're trying to make this so difficult, but I suppose you have your reasons.

Finally, if you insist the measuring tool we are suggesting is greatly flawed; ---BY ALL MEANS, explain to us your method of measurement that would not be flawed. I'm waiting.............

If you can't, or won't ---but just want to cry about the statistical measuring tool, then I suggest you've been hanging around some of these dowsers too long and their traits have rubbed off on you. (i.e. "I have located a Viking ship in my back yard, but I'm not going to dig it up.")

What exactly do you want fixed? Do you even know? Can you verbalize it?

 

[Jean310]

Btw.....I am not saying you should not continue just as you are....with the test.
In fact.....I am really looking forward to seeing the whole exercise published in a reputable scientific journal.
Please....carry on......

Judy

Did you ever consider maybe Carl is just placating you, to avoid your wrath?

If you can prove there is this grandiose problem in the world of statistics, please save the world from making any additional mistakes and publish your paper on how to fix the problem. Crying and complaining without a viable alternative is futile.

If you insist the measuring tool we are suggesting is greatly flawed; ---BY ALL MEANS, explain to us your method of measurement that would not be flawed. I'm waiting............. zzzzzzzzzzzzzzzzzzzzzzzzzz yawn.....

​

..............in fact the Statistical World is Waiting for your revelations. Put down your club and tell us all about your fantastic measuring tool. Can you verbalize it?