I just wanted to point out that Schroeder was not speaking to probability. He was speaking to the number of combinations. And also he was remarking how strange that nature would use that same combination for all visual systems, that it must have been pre-programmed in the lower life forms which had no use for eyes.
If he were talking probability, a lot of statistical issues would have come up including Bayes theorum no doubt.
After reading it closely, that is what Schroeder is saying, though he does it in a circuitous fashion. In an evolutionary system, one would expect there to be a conservation of basic bootstrap systems (its practically required) at some fundamental level. Even though the particular protein system found in animals is generally the same, the number of independent photo-reactive expressions (e.g. "eyes") is actually quite large (40+ known independent systems IIRC), and the number of photo-active expressions is too large to count. Photo-activity is a very active part of biological systems and evolves rapidly, mostly because photo-activity in chemistry is so common. There are also a number of other independent unrelated systems that have the same function in non-animals, but it seems like those were ignored for the sake of painting a pretty picture. It is roughly analogous to saying that because all cars are made of steel, all cars are the same. The raw material says almost nothing about the nature of the useful form, and it ignores the fact that while most cars are made of steel, there are also some that are not.
Since photo-receptivity is a common property in chemistry, there is nothing intrinsically special about the system protein system in question. If anything, conservation of the protein framework that animal photoreceptive systems are bootstrapped off of (in a pretty indirect fashion) suggests an evolutionary history, since there is no good design reason to have that particular framework in all animals unless it is a trait that has been conserved through evolution. The protein framework was not particularly optimal, but it was highly adaptable which is a trait only of value in an evolutionary schema.
From where I'm sitting, it looks like a perfect argument for evolution, where adaptability of a useful protein system was selected for rather than optimality of function because it made the system survivable. If I'm building a race car, I don't borrow parts off a pickup truck, but I probably could if I really had to. That's kind of where the eye biochemistry thing is at. There is evidence of many other eye type systems in biology, but most of the ones that we find today are those best suited for conservation in an evolutionary system rather than ones that would work best for a particular species.
Schroeder almost boils down to an argument from incredulity. I'm mostly extremely dubious of his argument because it carefully omits certain facts and avoids properly accurate representations that would damage the construction of his position. That is the behavior of a snake oil salesman, not a true scientist.