Also, can you quantify the words "small" and "high" as they apply to probability?
ID people typically use an arbitrarily chosen "
Universal Probability Bound", such as 1 in 10
150. (Note that the 1 in 10
50 number attributed to Borel is a "rule of thumb", not a mathematically established bound, as discussed at
Borel's Law and the Origin of Many Creationist Probability Assertions). Since it's easy to demonstrate that events which are more improbable than that do happen, or that events that are more improbable than any finite number you care to pick do happen, the probability
should depend on the situation ...
Note also that steps 1 and 2 cannot be accurately carried out unless we know
all possible ways in which the item under investigation could have arisen. Since we certainly don't know that, or even an approximation to that, for any item of practical interest, the EF is at best an academic toy (and there's good reason to believe it's not even that).
The only attempt (of which I'm aware) to apply the EF to a biological system was by Dembski in "No Free Lunch", in which he calculated the probability of the bacterial flagellum arising from random assembling of proteins, and concluded that it couldn't have happened that way. Well, duh! He then, of course, concluded that it arose by intelligent design, illustrating a major problem with the EF; if we don't have enough knowledge to make an accurate calculation but calculate anyway, the "conclusion" of ID may be in error (a false positive) and may be overturned in the future when more knowledge is available. This is all discussed in much more detail at
Not a Free Lunch But a Box of Chocolates. There's also some recent criticism of the EF at
Wrongly Inferred Design. Other interesting articles are
The advantages of theft over toil: the design inference and arguing from ignorance and
Information Theory, Evolutionary Computation, and Dembski's "Complex Specified Information".