[bobdsmith]

but can you remove intelligence from the analogy and make it work? You're equating random selection with intelligence. An intelligence is manipulating the decks to bring about a desired result. The desired result isn't known in real life and there's (supposedly) no intelligence guiding it.

Well there are a lot of differences because it is only an analogy. One key difference is that decks of cards don't reproduce like organisms do. So in order to make the analogy work a person must manually simulate the reproduction and mutation. Of course a better idea is to get a computer to do it.

Most uses of evolutionary algorithms on computers involve a mathematical based problem to be solved in which the desired result (the solution) isn't known beforehand. Running the same algorithm multiple time can yeild many differnent results. Sometimes the result can be suprising.

[DouglasKC]

Well there are a lot of differences because it is only an analogy. One key difference is that decks of cards don't reproduce like organisms do. So in order to make the analogy work a person must manually simulate the reproduction and mutation. Of course a better idea is to get a computer to do it.
Most uses of evolutionary algorithms on computers involve a mathematical based problem to be solved in which the desired result (the solution) isn't known beforehand. Running the same algorithm multiple time can yeild many differnent results. Sometimes the result can be suprising.

So is it fair to say that the theory is untestable without inserting some type of intelligence into the testing?