I would like to respond to a few remarks in K.E.M.'s letter in In-Genius #40.
I don't agree that mechanical manipulation is meaningless with respect to measuring intelligence. The human brain is mechanical. The neurons fire automatically when they get a certain amount of input; the brain activity is entirely mechanical. In the gigantic complexity of this mechanism, something happens that we call "awareness" or "intelligence".
It is very hard to understand indeed how awareness can exist in mechanical complexity, but we have to face the fact that it does; man has too long been fantasizing about an independent soul with a free will. The brain activity, fed by sensorial input, mirrors the outside world, and mirrors itself while mirroring the outside world. In this embeddedness, something "extra" apparently occurs that is awareness. It is instructive to think about is as analogous to the "extra" aspect that occurs when two mirrors are opposed to each other, or when one films a television screen with a video camera connected to it. The infinite embeddedness causes new phenomena not present in the original picture. Analogous to that, one can imagine that the brain activity, on its deepest level nothing more than mechanical counting, contains similar embeddedness, resulting in phenomena not present in the original sensorial information.
Although it is hard to understand exactly how this works, it will some day become comprehensible that what you think is YOU is really a complex embedded structure of chemical/electrical events.
So it is not so illogical that IQ tests contain mechanical items; the brain is mechanical itself. Having said that, I point out that GOOD test items are mechanical only on their deeper level, but appear non-mechanical on the surface, just like awareness appears non-mechanical. But to solve them, you need to see right through the superficial level and recognize the mechanics behind it - and THAT takes a lot of awareness.
The ultimate test is to understand the mechanics behind awareness itself; how many are capable of that, even when it is explained to them? Only the super-geniuses.
About the program that scored IQ 84 to 92: that sounds like below-threshold scores that you get when you fill in the test at random. Have you tried that? If the test is valid up to IQ 174, it will probably not be able to measure below say IQ 130. So the program did not do well.
It has been done better. I am aware of a program called "HiQ-Solver" who scored IQ 160 on the Eysenck numerical test and solved half the problems in the Super Brain test it was matched against and which was widely published at the time (1984). One of his creators was ISPE founder CHEVALIER BARON Dr Christopher Harding. HiQ-Solver subsequently failed the ISPE test, but became the all time top dog on the Mega Test with 47 right, be it under the false name of Eric Hart (others have unrightfully claimed that name). Currently Mr HiQ-Solver is a member of TOPS and writes, pseudonymously, brilliant articles on topics like genius, IQ tests, running, etcetera etcetera.