The number of one in 1020 is not a probability calculation. Rather, it is statistical data.
But if Behe had read White’s 2003 paper (table 1) “The de novo selection of drug-resistant malaria parasites.” N J White and W Pongtavornpinyo Proc Biol Sci. 2003 March 7; 270(1514): 545–554. he would have read that
The estimates for chloroquine and artemisinin are speculative. In the former case, this assumes two events in 10 years of use with exposure of 10% of the world’s falciparum malaria (Burgess &Young 1959; Martin&Arnold1968; Looareesuwan et al. 1996; Su et al. 1997
Luskin is correct, the number is not a “mere guess”, it’s a speculative estimate. Glad we got that right. Why Luskin failed to mention this is beyond me since he does seem to quote the paper in question. Perhaps if Luskin had spent more time on reading the papers and less on emphasizing the academic achievements of White, he might have found the error in Behe’s claim himself.
What Behe meant when he said that
the 1020 statistic is an empirically derived fact
is less clear. Surely since ID proponents are so critical of evolutionary scientists when it comes to confusing fact and fiction, Behe may wants to revise his statement.
Now it has already been well document that the paper from which Behe lifted this single data point was careful to state that this was merely a guestimate by the author.
As I have discussed, Behe asserts that the probability associated with a “CCC” is 1 in 10^20. Where does this number come from? From footnote 16 in the first excerpt given above - White, N. J. 2004. Antimalarial drug resistance. J. Clin. Invest. 113:1084-92. Here is the actual passage from the review by White that mentions the number 10^20:
“Chloroquine resistance in P. falciparum may be multigenic and is initially conferred by mutations in a gene encoding a transporter (PfCRT) (13). In the presence of PfCRT mutations, mutations in a second transporter (PfMDR1) modulate the level of resistance in vitro, but the role of PfMDR1 mutations in determining the therapeutic response following chloroquine treatment remains unclear (13). At least one other as-yet unidentified gene is thought to be involved. Resistance to chloroquine in P. falciparum has arisen spontaneously less than ten times in the past fifty years (14). This suggests that the per-parasite probability of developing resistance de novo is on the order of 1 in 10^20 parasite multiplications. “
In a review paper published in TREE, Nick Matzke further demolishes Behe’s guestimate
The argument collapses at every step. Behe obtains the crucial 1020 number from an offhand estimate in the literature that considered only the few CQR alleles that have been detected because they have taken over regional populations. What is needed, however, is an estimate of how often any weak-but-selectable CQR originates. A study conducted in an area where CQR is actively evolving  showed that high-level CQR is more complex than just two substitutions but that it is preceded by CQR alleles having fewer substitutions; moreover, Behe’s two mutations do not always co-occur. As a result, CQR is both more complex and vastly more probable than Behe thinks. This sinks his one in 1020 estimate for CQR, in addition to his notion that protein-protein binding sites are more complex and, therefore, less probable than CQR.