September 17, 2000

CANCER CURE, BUMBLING BROWNE, AND DIPPY DOWSING.

When my friend, magician Doug Henning died earlier this year, the Transcendental Meditation movement tried to explain away his passing, since he'd been convinced that TM had cured his liver cancer. Now I've heard from an angry TM follower who rejects the mystical aspects of the movement, and considers the relaxation techniques to be pretty well the whole value of it. He and his wife attended Doug's funeral.

. . . I was completely outraged when at his memorial a TM movement bigwig named "Dr." Nader had the temerity to get up and say something to the effect of, "At the time of Doug's death there was no trace of cancer in his body but it was his time to go so he died anyway."

As with most of the other movements of this sort, the TM folks will resort to any means necessary to shore up their shaky position. They will simply never admit that they are living in a house of straw. You'll see shades of that sort of "logic" and "spin-doctoring" in the DKL discussion, up ahead.

...............................................................

 

Amber Wilde, a 4-months-pregnant University of Wisconsin-Green Bay student, 19 years of age, disappeared September 23, 1998. Her car was found a week later parked near a Green Bay sports bar. Last July, in response to impressions offered the family by a "nationally known psychic," the police took trained dogs and digging equipment to a site and spent a day digging up the area. Nothing was found. Police referred to the tip as "plausible," despite the fact that they had previously used a psychic in another case, with exactly the same negative result. But you won't see this item featured in the resumé of Sylvia Browne, about whom we receive many inquiries every week.

This Browne bumble reminds me that eleven years ago, here in Florida, a group of six high school students headed away from a school dance in their van late one night, and were never heard of again. The police were contacted by psychics, who said the kids had run away, and located a few of them -- or so they claimed -- working as mechanics and waitresses in small Western towns. Then, last year, a rusted wreck was dredged up from a canal a few miles away from the school. Inside were a number of opened beer bottles. And six skeletons.

...............................................................

 

Last week's puzzle was rather different. You were asked to go to a web page report made by some folks who are selling a dowsing rod, the "DKL Lifeguard," to federal and municipal agencies, and who claim that it can detect living persons at great distances and through any sort of shielding. Your task was to figure out how this report had been "enhanced" to look better. We got some 20 responses that were pretty close, but no one actually penetrated the full depth of the scam. Here are the facts.

First of all, the "Government Laboratory Testing" referred to by DKL, was done by the renowned Sandia Labs in New Mexico, who DKL refers to only as "a government facility." Perhaps the reason that the DKL description you worked on did not mention the Sandia name is because the lab has a much different version of the March, 1998, tests they did on the dowsing rod. Look in at www.sandia.gov/search.html and enter "DKL" in the Search box. There you'll see just how poorly represented those tests are by DKL.

Note that the DKL folks write: "The data, when evaluated by two different independent scientific agencies clearly showed that LifeGuard . . . was able to detect the presence of a living person in a general direction 72% of the time." The "two different independent scientific agencies" are not named, either, nor are any details of their evaluation given. The bottom line seems to be that their dowsing rod was correct 72% of the time. Nothing is said to define their term "in a general direction," but in view of what you're about to be told, that's a minor detail.

The only way to make sense of their statement, "A statistical analysis showed that the operator's chance of making a correct detection was over 90%," is to assume it refers to the fact that the probability of getting six out of twenty five by chance is 90%. I'll skip that. Read on.

Their first claim about the tests, in which a human subject was concealed in one of five large plastic crates, referring to what Sandia calls Test A, is that "The operator detected the human target 25 out of 25 times for 100%." Remarkably, most of those of you who offered an analysis of this report rather ignored that startling statement. Only two things are wrong about the statement. One, it was ten trials, not twenty-five. Either DKL simply quoted the wrong number in ignorance, or they hyperbolized it for effect. Two, yes, the operator -- a fully-trained, fully experienced, top executive of DKL, chosen by them to demonstrate their dowsing rod -- did obtain 100% results. But it was impossible not to do so, since that was the "baseline" series I had suggested to Sandia, and that was "open," which is to say that everyone present -- including the operator with the rod -- knew where the target was in each of the 10 trials!

Let me explain why I always include a baseline series in any set of tests. We know from past experience that those being tested will offer any number of excuses for failure, and applying the baseline provision closes that escape hatch. They will want to claim that the target was not "strong" enough, the weather was wrong, there was too much moisture, the operator was inexperienced or untrained, or for some other imaginative reason, the test failed. But if the operator first goes through a baseline series, and gets 100% as they always do, because they know the answer -- those excuses cannot be used.

DKL then describes on their web page the actual double-blind tests, which are outlined on the Sandia pages as "Test B." Their first whopper here says, "The operator was able to detect the target 18 out of 25 times, or 72% of the time." Not true. Just not true! Look at the table published by DKL, and you'll see that they count as a "hit" any guess that is either right -- or next to the right one, on either side! DKL was right 6 times out of 25, or 24% of the time! That's what one would expect by tossing dice, or reading Chinese fortune cookies! Also, the dashes in columns 1, 3, 4, and 5 indicate misses!

Read that again. It's incredible, but true. To correct the DKL report:

"When human target was in Position 2, operator detected 100%." Fact: no, 50%
"When human target was in Position 3, operator detected 71%." Fact: no, 40%
"When human target was in Position 4, operator detected 75%." Fact: no, zero %
"When human target was in Position 5, operator detected 75%." Fact: no, 25%

And please note that there's no listing by DKL of results for Position #1. The target was in there twice. But DKL puts that under the category of "out of detection pattern" (indicated by a dash) as were another five of their guesses that you and I would call "total misses," but which they choose to merely omit. Inventive, but not honest.

In summary, the DKL dowsing rod, handled by a very competent, experienced, trained, trusted, executive of the DKL corporation, scored 100% in "open" tests where the target was known to everyone, but when the data was blinded he got exactly in the range of what chance -- simple guessing -- would call for: 6 in 25! And the DKL people published a spurious account and analysis of the data developed by this very responsible, careful, experienced, federal agency. Why? I leave that for you to decide.

What does Sandia Labs have to say about all this? Their 14-page report ends with a definitive statement: "This device cannot perform any better than random chance." The lab came to that conclusion after a very carefully-designed series of tests.

During the Sandia tests, every possible advantage was given to the DKL operator. For example, the Operator's Manual for the device said that "Most operators can get maximum range in open air past 20 meters (66 feet) using a walk-by . . ." In fact, DKL claims that it works at distances of up to 500 meters (1640 feet) through any known material, locating a human target "in three to five seconds"! Sandia placed the target only 15 meters (50 feet) away, well within the lesser operating range claimed by DKL. Plenty of separation between the 5 positions was arranged, 30 degrees apart, so they were separated by more than six times the minimum angular distance required by DKL. Under these circumstances, the device certainly should have been able to perform accurately and dependably.

The actual formal test, what the Sandia report designates as Test B, was a simple matter of having the target -- a person -- inside one of the randomly chosen five crates, but without anyone but the target person himself knowing where he was. The DKL operator was required to simply designate, using the device, which was the correct crate. He was allowed to use as much time as he wanted. He obtained six correct results in 25 guesses. Ideally, five out of 25 would be expected, and six or seven are well within expectation.

But how could DKL claim that they obtained 18 correct out of 25? Well, after the test results were revealed at the end of the day, the representatives from DKL declared that a guess on either side of the correct crate should also count as a hit -- that's what they mean in their report by "in the general direction" -- and that's how they claim the 18 out of 25 (72%) result that they reported. However, as Sandia correctly pointed out, since DKL stated in writing that their device had an accuracy of ±5 degrees, and the distance between adjacent crates was more than 30 degrees, this "out" could not be allowed. But DKL used it, anyway. They were working with a self-applied leeway of twelve times their own claim!

In any case, given completely random chance and considering the layout pattern, the probability of being correct even with the "improved" DKL way of looking at things, is still 92% of getting what they claimed! No matter how you look at it, DKL failed Test B spectacularly.

(There was also a third set of tests, Test C, done by Sandia on that occasion, but DKL chose not to mention that one at all. Perhaps we can guess why. This test, too, was a total dud. Where chance expectation was 7.2, the operator obtained just 6 hits.)

You should also know that the DKL operator took an average of 3 minutes for each of the Test A guesses, when he knew the answers, but in Test B, he took 11 minutes for each one. What happened to the "three to five seconds" they claimed in their advertising?

When the final results of this test were announced by Sandia, as we might expect, DKL was well prepared with reasons for the failure. During Test A, they had given no indication that conditions were anything but ideal, but during B, they said, "static charge," "residual signals," and the "sharp edges of the crates" were "distorting the field." However, as Sandia reminded them, their published literature clearly states that there are "no known countermeasures" that can defeat their device, and that it "penetrates all forms of camouflaging." The target subject could have been surrounded by steel girders, and still been located, according to the DKL claims. Then, in a parting shot, the DKL people declared that during the test, the "field had spread," whatever that might mean, though this problem had not been evident during Test A.

Even more astonishing is what DKL claims about the principle on which this thing is supposed to work. They say that it picks up the electromagnetic field of a beating human heart. No, I'm not kidding. Basic facts: the heart operates on a built-in timing system that triggers it independently of the rest of the body, and that consists of a very minute pulse that is so overwhelmed by the other EM fields in which we are immersed, that the claimed detection is just ludicrous. I need hardly add that any sort of shielding would defeat a detection system, even if it did work. Apparently best-selling author Tom Clancy didn't see the absurdity of the notion, since he featured it in one of his novels, "Rainbow Six," as if it worked.

As often happens with these quack devices, there is something useless but attractive in there for "flash" effect to catch the eye of the potential customer. DKL achieved this by building in a simple laser pointer for a "line of bearing," and they have it "GPS (Global Positioning Satellite) compatible" as well. That model is listed for $13,995.

The DKL people grossly misrepresented the results of the Sandia tests. No amount of waffling and shuffling will get around that. Casting down the JREF gauntlet, this week I sent the following letter to Howard Sidwell, Chief Executive Manager of DKL, this being a hard copy of the e-mail message I'd sent to the DKL web site earlier in the day:

Mr. Sidman:

I am James Randi, president of the James Randi Educational Foundation, Florida, USA. We are prepared to pay DKL the million-dollar prize offered by this Foundation, following a simple demonstration that your device actually works. TIME Magazine has agreed to cover such a test, and a major TV network will feature the test on a forthcoming TV special, if and when it takes place, or as a definitive report following the test. Should the test be successful, you will immediately receive the prize money in the form of negotiable bonds. The terms of our offer can be found at: www.Randi.org.

Please let us know of your interest in proving your claim, winning the million dollars, and obtaining international validation and recognition for your company and your product. The challenge is simple: if the DKL device works you will receive a million dollars. It would take less than a day of your time, and we are prepared to travel to your location or any location named by you, to witness any test -- which would be supervised by verified, qualified, independent, authorities approved by you, and would use an operator selected and trained by you, and a target of your choice, at any time, day, and location of your choice.

Should there be other requirements for a definitive test, please notify us and we will of course accede to whatever is required for a proper test. Should you require further information, you may call, fax, or e-mail us at the addresses shown on this letterhead.

This is a genuine, serious, offer to pay DKL one million dollars and fully validate the claims made by your company, following a simple test of your product. We await your response with interest.

Signed, James Randi

This was not the first time I sent such a letter to Mr. Sidwell. The first went out to him, with essentially the same offer, in March of 1998, and was followed periodically by others repeating the challenge. No response has ever been received, nor do I expect that this letter will evoke any reaction. The DKL people simply don't want to be tested, yet they are selling a device that does not work, and in many cases we are paying the bill. DKL sells to fire departments, drug & law-enforcement agencies, border patrol units, search-and-rescue teams, as well as security and military personnel, according to their literature. They appear at major conferences and conventions world-wide, peddling their toy to anyone who will buy it -- often with your tax money.

Yet, strangely enough, they will not accept the JREF offer to give them a million dollars. Why? At the top of every new web page, to keep you informed each week, I will insert the latest response from DKL.

...............................................................

 

You can anticipate a full report here soon of a TV program that was recently aired in Korea. Featured prominently on the show was Uri Geller (remember him?) and some quite delicious facts developed during the program that you will find most interesting. But that's for later . . .

...............................................................

 

The "hit" counter just went in on the home page last week, and we're already at over 4,000 hits! That's almost 600 a day, which is very encouraging.

But don't expect to get through to the "StarLight" page in reference to the Polk County fiasco outlined last week. As we "go to press," Sue Miller Hurst has closed access to her page. But strangely enough, we have the codes to get through . . .

...............................................................

 

 

Martin Gardner suggested this minor but interesting puzzle to me:

Put simply, what is the area of the smaller square? Do it without drawing, without calculating. Just look at it. A delightful solution. It should take you about 30 seconds. Start ---- NOW! (Responses directly to 76702.3507@CompuServe.com).