SkepticblogSkepticblog logo banner

top navigation:

Randi and the race-car driver

by Donald Prothero, Dec 05 2012

Most of us have read about homeopathy, and its claims to be an effective form of medicine. It was first developed by the German physician Samuel Hahnemann in 1796, and based on the old Greek and Medieval “Principle of Similars,” or the idea that “like cures like.” For example, Hahnemann observed that cinchona bark helped treat malaria in sick people, and caused similar symptoms in healthy people, so he reversed the logic and reasoned  that whatever causes similar symptoms can be used to treat it. (Cinchona bark does contain natural quinine, a cure for malaria). He argued that if poison ivy causes skin rash, then diluted poison ivy is a cure for skin diseases. But his theory of “likes cure likes” was pure Medieval alchemy and mumbo-jumbo, completely invalidated when modern chemistry developed in the nineteenth and twentieth centuries. This archaic version of pre-modern medicine has persisted virtually unchanged until today. The ingredients used in homeopathy sound like those of a witch doctor or medieval apothecary  (or something out of Harry Potter’s potions class): snake venom, ground honeybee, crushed bedbugs, live eels, wolf milk, arsenic, poison ivy, diseased tissues (including pus, tumors, feces, plus urinary discharges, blood and tissues from sick individuals), quartz, gold, oyster shell, and common salt. Their other ingredients, known as “imponderables”, supposedly “capture” electromagnetic energy by exposing alcohol or lactose to sunlight, X-rays, or lightning.

What makes homeopathy different from other alternative medicines that use bizarre ingredients is their key method: dilution. All these substances are then ground into fine powder to make a “tincture”, diluted in water, and then diluted over and over again. The “hitting” or “shaking” technique by the homeopath is very important because of “kinetic energy input” and its effect on their medicine. (This is just a fancy way of saying that more shaking probably mixes a solution better, but it is given a mystical mumbo-jumbo meaning in homeopathy). A typical homeopathic remedy has a “strength” of “30C” in their terminology. This means that the original agent has been diluted 30 times by a factor of 100 each time. A simple calculation means that this material has been diluted by a factor of 1 x 1060, or 1 followed by 60 zeroes (in common terms, one part “tincture” in one million trillion trillion trillion trillion trillion trillion parts of water). There are only 3 x 1025 molecules of water in a liter, so for all intents and purposes, there are no molecules of the “active ingredients” left and the homeopathic remedy is just water and nothing more. At this concentration, a patient would need to swallow 1041 pills (a billion times the mass of the earth) or 1034 gallons of elixir (a billion times the volume of the Earth) to consume even a single molecule of the substance. Even a stronger homeopathic solution of “12C” has only a 60% chance that one molecule of the tincture is present. Yet according to homeopathic theory, the medicine is stronger the more it is diluted. Most homeopathic remedies have been diluted so much that they are just small bottles of water. In effect, homeopathy has become the modern “water cure.”

Of course, drinking lots of clean water is good for you, but this is something different. Homeopaths are claiming that a “medicine” made of nothing but water somehow has acquired “magic properties” or a “memory” of the substance that was once diluted in it, and somehow this water “memory” transfers the therapeutic value of the original material. Anyone with common sense, and especially anyone who has taken even an introductory high school or college chemistry class, can immediately tell that something is wrong with this theory. First of all, in chemisty lab you learn by repeated experiments that dilution only makes a solution weaker and less reactive. There is absolutely no known instance where a less concentrated substance has a stronger effect. Secondly, you learn early in chemistry class that water is a simple molecule—just two atoms of hydrogen and one of oxygen. It does have some remarkable properties, like its high heat capacity and its ability to become less dense after it freezes (which is why ice floats). But water (and even the crystalline forms of ice) is still an extremely simple molecular structure, with no potential for “memory”. There are complex organic chemicals like DNA, or the complex sheet structure of a clay mineral, which do have the complexity to record changes in their “code” and thus have a memory. But water (even in the case of the most complex ice crystals or clathrate structures) has no such ability. The Australian Council Against Health Fraud wrote, “Strangely, the water offered as treatment does not remember the bladders it has been stored in, or the other contents of the sewers it may have been in, or the cosmic radiation which has been blasted through it.”

But couldn’t homeopathic remedies work like a vaccine, where we introduce tiny amounts of some germ to a patient to stimulate immunity? Superficially, they sound like the same thing, but in reality there is no comparison. A typical vaccination shot has billions of killed viruses or virus fragments in a few milliliters of solution, and these work by immediately stimulating the immune system to kick into high gear and recognize their chemical signal, so that the white blood cells quickly attack when a full-strength virus invades at some later time. By contrast, a homeopathic remedy is pure water, or at best one or two molecules of an active ingredient, and there is no reason to believe these things stimulate the immune system or anything else. The analogy is so ridiculous that the famous nineteenth century doctor Oliver Wendell Holmes wrote that this comparison was like “arguing that a pebble may produce a mountain, because an acorn can become a forest”.

Reading the basic theory of homeopathy (“likes cure likes”, “tinctures”, dilution increases potency, water has memory, and the long list of peculiar substances used in cures), it is clear that it was based on medieval notions of alchemy (as homeopathy was when it was founded in 1796). It bears no resemblance to what chemists have discovered in the past 200 years about the nature of matter and chemical compounds and their behavior. In fact, homeopathy would have probably vanished like many other forms of quack medicine in the early nineteenth century, except for the fact that conventional medicine at that time (with leeches, many types of toxic “medications”, and no sanitation, antiseptics, antibiotics, or anesthesia) was often deadlier to the patient than leaving them alone. Thus, early homeopathic hospitals in England had greater success than conventional hospitals of the time because all they did was give their patients a water placebo and a lot of attention, and did not make the patients sicker. Those who had stronger immune systems survived, while similar individuals died after being treated in conventional hospitals and being bled and infected by doctors who didn’t wash their hands.

But what about rigorous clinical studies where many patients are involved, which all real medicines are required to pass in order to meet FDA approval? For years, nearly all the evidence favoring the efficacy of homeopathy were individual anecdotal cases, or examples like the fact that early nineteenth century homeopathic hospitals were less deadly than conventional hospitals (as we pointed out already, only because conventional medicine of that time was worse than no treatment at all). Every other study that had favored homeopathy suffered from one deficiency or another: insufficient sample size, bias by the investigator, improper statistical techniques, inability to replicate the results in any other lab, and many other flaws.

Then in 1988, Jacques Benveniste, a former French racing driver who had taken up medicine after suffering a back injury, submitted a paper to Nature (Davenas et al., 1988) that appeared to be a rigorous study supporting homeopathy. The Nature editor, John Maddox, was skeptical, but after it passed peer review because there were no obvious flaws in the reported methods, he consented to publish it—but with a disclaimer that Nature would send its own panel to investigate Benveniste’s lab and see if the results were credible. The team consisted of Maddox himself (a trained chemist and physicist), the chemist Walter Stewart, and the famous magician, skeptic, and mythbuster James “The Amazing” Randi. They watched the laboratory closely as they repeated the experiment, and at first the lab kept getting results consistent with their original report. But then the panel realized that the investigators might be biased in their sampling, so they set up a procedure where the investigators could not know which test tube was which (a “double blind” test). As the results were being analyzed, Randi performed some sleight-of-hand magic and card tricks to amuse the people in the lab while they waited. Sure enough, when the new results were reported and the panel revealed the actual identities of each test tube, there was no positive effect of the homeopathic medicine (Maddox et al., 1988). All the results were no different from random chance reactions.

After Nature reported the debunking by the panel, and the probable reasons for the experimenter bias, they discovered that Benveniste had been partially funded by a French homeopathic company, so there was a clear conflict of interest. Benveniste was not an impartial scientist, but stoood to gain if he found results that improved the profits of his sponsor. Since Benveniste’s study, three different independent labs failed to replicate his results, and Benveniste lost all credibility in the scientific community. Subsequently, he left the conventional medical community and went private. With big funding from homeopathic medicine producers, set up his own private lab without scientific oversight where he continued to make amazing pronouncements about his “miraculous” discoveries. Fancying himself a scientific martyr like Galileo, he felt he deserved a Nobel Prize, but instead has twice (1991, 1998) won the satirical “Ig Nobel Prize” for the worst pseudoscientist of the year.

The claims of homeopathy have been tested over and over again, and none has passed muster (Ernst E, 2002). In 2005, the premier British medical journal The Lancet published an analysis of 220 studies about homeopathy, half of them conventional studies of medicines, and half of them of controlled homeopathic experiments. They found no evidence that homeopathy had any real value except as a placebo. In 2006, the European Journal of Cancer surveyed 6 studies, and found homeopathy had no effect (Milazzo et al., 2006). Even studies by homeopaths themselves (Jacobs et al., 2006) often show that their products have no significant effect, despite their biases to prove otherwise. As Singh and Ernst (2008, p. 136) document, the trials continue, but not once has any study shown a statistically meaningful evidence that homeopathy works.


  1. Davenas E, Beauvais F, Amara J, et al. (June 1988). “Human basophil degranulation triggered by very dilute antiserum against IgE”. Nature 333 (6176): 816–8.
  2. Maddox, John, James Randi and Walter W. Stewart (28 July 1988). “‘High-dilution’ experiments a delusion.” Nature 334: 287–290
  3. Ernst E (2002), “A systematic review of systematic reviews of homeopathy”, Br J Clin Pharmacol 54 (6): 577–582
  4. Milazzo S, Russell N, Ernst E. “Efficacy of homeopathic therapy in cancer treatment.” Eur J Cancer. 2006 Feb; 42(3):282–9.
  5. Jacobs J, Guthrie, B, Montes G, Jacobs L, Colman N, Wilson A, DiGiacomo R. “Homeopathic Combination Remedy in the Treatment of Acute Childhood Diarrhea in Honduras.” The Journal of Alternative and Complementary Medicine. 2006 Oct.; 12(8):723–732

14 Responses to “Randi and the race-car driver”

  1. Max says:

    “not once has any study shown a statistically meaningful evidence that homeopathy works”

    Out of a hundred studies, you’d expect 5 studies to get p under 0.05 by chance alone.
    Brian Dunning mentioned this study, and said it’s “one of a minority of a few scattered studies that did find a small statistical improvement in symptoms among homeopathy users compared to a control group who took an identical placebo.”

    Seems that studying any implausible phenomenon carries the risk that any positive results raise suspicion of scientific misconduct or fraud, because the prior probability of fraud, as low as it might be, is still higher than the prior probability of the implausible phenomenon, and a positive outcome is more likely if there’s fraud than if there’s no fraud and the phenomenon exists.
    What’s more, the better controls and more subjects are used, the more likely the positive results are due to fraud rather than innocent bias or chance.

  2. MadScientist says:

    Since there is absolutely no evidence that homeopathy works, by homeopathic principles homeopathy is a huge success! They just can’t lose …

    @Max: “Out of a hundred studies, you’d expect 5 studies to get p under 0.05 by chance alone.” No, that’s not the case at all. Even with a Gaussian distribution and so on, you will get those results only once every so many 100 studies. Statistics is simple, but not quite that simple. It is even possible to have experiments where no matter how large the population studied is, you never see a case of >0% being ‘positive’. A good example of that is alchemy: in how many cases have the magical incantations turned lead into gold?

    • tmac57 says:

      in how many cases have the magical incantations turned lead into gold?

      All of the Dirty Harry movies perhaps?

    • Max says:

      You give me 19 vials of water and 1 vial of a homeopathic preparation, and I’ll identify the homeopathic preparation 5 times out of 100 with p=0.05.

      • MadScientist says:

        Well, on average and for very large sets of trials. If you were to do sets with 100 trials each, you will still see a variation around the 0.05. It is even possible that in one trial you get p=0 and in another you get p=1. You can try it out by writing a program which generates a random number from 1 to 20 and you guess at what number the computer has. It requires a hell of a lot of patience though.

  3. Brian says:

    What is the source for the image of the newspaper article?

  4. Ed Seedhouse says:

    Just to be clear I support the right of homeopathuals to equal treatment in a free and just society.

  5. Adrian Morgan says:

    I colour-corrected the homeopathic cartoon. The result is a bunch of image artifacts subject to pareidolic interpretation. Personally I see a bunch of people playing some violent, chaotic and implausible sport with lots of body protection.

  6. Max says:

    In homeopathy, is there some dilution of, say, arsenic right between toxic and therapeutic, where it has no effect?

  7. Zach says:

    “There is absolutely no known instance where a less concentrated substance has a stronger effect.”

    I don’t believe this is actually correct, unless you are using a very conservative definition of “effect”. Just off the top of my head, endocrine disrupters are now routinely argued to have atypical dose-response curves where lower doses have apparently greater effects on endocrine function. In fact, if we define “effect strength” in terms of specific beneficial outcomes all pharmaceuticals operate within an optimal range. Obviously for the vast majority of drugs increasing the dose will simply overstimulate the molecular target of interest and generalize to offtarget sites, but I think you need to be more specific in your definition of “strong effect” if you want to convincingly argue that increasing dose always increases effects. Regardless, any time we have to consider complex feedback and feedforward processes it is worthwhile considering that response curves may not be perfectly monotonic (particularly over long timecourses) with respect to a given effect of interest.

    Nonetheless, I think we can safely assume that in the area of homeopathy the absence of any dose whatsoever makes the all the caveats pretty much moot.

  8. Gary says:

    Funniest cartoon ever, still can’t stop laughing.

  9. scars says:

    Someone necessarily lend a hand to make seriously posts I
    would state. That is the first time I frequented your web page and
    up to now? I amazed with the analysis you made to make this particular post
    extraordinary. Magnificent process!

  10. Clement Doyer says:

    I don’t know why I bother to post on this blog since it seems that opinions are already formed and no amount of facts will modify them. To look at homeopathy with a high school chemistry mindset and to deduct all the arguments from this is simplistic and does not describe reality. For 23 years now, I’ve been curing animals, mostly dairy cows, some 100 000 of them with homeopatic formulas. Don’t bother to come again with the placebo effects on this… And don’t insult my intelligence by telling me that more than 300 farmers are stupid and keep buying products that do not work…
    I think Hanneman had an intuition when he developped this therapeutic method, he found a way to use quantum physics. By diluting, he reduced the particle parts and he multiplied the “message”, the wavelength of the base of the remedy. This radio-like emission in contact throught a mucus membranewith

  11. Clement Doyer says:

    with the water that composes 92% of the organism, by resonnance, affects the cell that was sick, with a false note, re-establishes the correct wavelength and the health of the cell. This is my view of the process. I will not spend hours to try to make people understand E=mc2, it’s been more than 100 years, and it’s still not integrated.
    See if you publish this!