Aparently, health dowsing is used to determine whether certain "spots" are good or bad for your health. It's usually done over a map to find out where a "healthy" spot to build a building is, but it can be done on the actual ground as well. And, as this video from James Randi's 1991 show Psychic Investigator evidences, it offers up a heaping cup of bad math along with its pseudoscience.
The basic test that Randi gives is fairly simple: if certain "spots" indeed have a "bad" or "good" impact on ones health, then each dowser should agree on the impact of a particular spot. Randi placed three blue circles on the floor of the studio, one of which has been determined to be a "bad spot" by a dowser beforehand. The test is for another dowser to determine which is the bad spot. If dowsing is real then the same spot should be picked. Note that picking the same spot does not prove that dowsing works - the test really ammounts to a 1 in 3 guess.
In the video, the dowser does manage to pick (guess) the correct spot, but he does something else. He "works in percentages", id est he tells you what percent of the spot is "good" or "bad". The math behind this should be fairly simple, shouldnt it? Well the guy gets it totally wrong. He finds that the first spot is bad; moreso, it is 30% bad. The second is a good spot which is 50% good. The final one is bad at about 70%.
Now, anyone (except maybe this guy) can tell you that a total always adds up to 100%. In the case of the "good" and "bad" spots, there's only two options: good or bad. Therefore the percent good and the percent bad should add up to 100% (if something is 10% good then what is the other 90%? It has to be 90% bad, of course). But this doesnt really match what the guy is saying. The first one he first determines to be definately bad. But then he says it is only 30% so. This would indicate that the spot is 70% good. In other words, it's almost 2.5 times as good as it is bad. Why, then, does it count as a "bad" spot?
His second attempt makes even less sense. This time he determines that the spot is 50% good, "quite good" he pronounces. Well, that would mean that the spot is also 50% bad. If 50% good is "quite good" then the spot is also "quite bad". If it's equal parts good and bad, then what does this mean? Does it do nothing?
The last one he finds 70% bad (30% good), which he thinks is very detremental to one's health. If the first one was only 30% bad and is unhealthy, then why is the one that's 30% GOOD not healthy? Then he claims that the second one was the best for your health. Yet, that spot was 50% good whereas the first one was 70% good. So shouldnt that be the best spot?
These numbers make absolutely no sense. How anyone could believe this kind of nonsense is beyond me. How people can actually hire and pay people like him to determine where their homes and offices should be constructed is even more astounding.