I've recently read his old but still entertaining book, Flim Flam (tip - the Kindle version is a lot cheaper than paper equivalents). It is an effective dismemberment of various claims to supernatural and paranormal abilities and events. It might seem at first sight he over-analyzes some cases. For example he spends page after page on the detail of the Conan Doyle fairy photos. This might appear unnecessary, because those pictures are so obviously fakes - the images of the fairies are quite clearly paper cutouts. Yet Randi usefully examines the different ways those who were apparently taken in by these crude fakes supported their beliefs and ignored the obvious.
However, I do notice one thing. Randi is understandably heavy on scientists who have a tendency to be naive when presented with fakery. It isn't their field. But because of this, he ought to be careful himself when writing about science, because his opinions in the scientific arena can be more than a little naive too. I found one example rather entertaining. Randi writes:
Jack van Impe, a TV evangelist who perspires and preaches his version of science regularly to millions of believers, recently gave us an Easter message that reflected his ignorance of science. He referred to the preposterous "Jupiter Effect" so beloved of some nuts, which is supposed to cause wonderful catastrophes in 1982. The Earth should be a mess at the end of this claimed alignment of the planets, and I can hardly wait to see the show. Said Jack, "The Earth will be seven times hotter." Codswallop. The term has no meaning. "Seven" is a number, Jack. If you take the normal temperature to be 70 degrees Fahrenheit, that makes the new reading 490 degrees Fahrenheit. If you're in Europe or Canada, that same temperature is 21 degrees Celsius, giving the other folks a break with 148 degrees C, which is equal to only 298 degrees F.Yes, of course seven is a number - but that doesn't mean that something can't be seven times hotter than something else. Temperature is a measure of the mean kinetic energy of the atoms or molecules in a substance, and one thing can have seven times as much energy as another. So something can be seven times hotter than something else. (Or the world can be seven times hotter at one point in time than it is at another.) What Randi is really identifying is the arbitrary nature of the temperature scales he uses in his illustration. (And the stupidity of the prediction.) But the way he phrased his attack shows that he too can get it wrong when working outside his field of expertise.
Image from Wikipedia