I understand, I read that too. That’s not what I’m talking about. The fitness effects of the mutations make a neglible difference here. That’s not what’s wrong with Sanford’s arguments.
I repeat, it’s that he is focusing on a particular set of mutations, for example imagine you have this sequence:
AAAAGCCCCTTTT
And it randomly suffers a single substitution once every 100 times it is copied. So you wait 500 copy events and find 5 mutations have occured. Then you say, how long would it take for those 5 particular mutations to occur, on average?
Say these are the mutations(in bold) that changed from the above sequence:
TATAGTCCCCTGT
How long would you have to wait, on average, before those specific 5 mutations occur? If we started all over again, an exceptionally long time. Because there’s no guarantee that at any point the mutation that happens to occur is the one you’re waiting for. So you could wait for the first 100 copies(lets just call them generations) to see whether a mutation has occurred, and then check if it’s one of the mutations you wanted.
Since there are many more mutations possible than those 5 I have highlighted, chances are much greater that it is not the one you’re waiting for that occurs, so now you have to wait again for another 100 generations. And then the problem repeats, there’s still many more possible than the one you’re waiting for, so it’s likely another one than the one you want. And so on.
This would be very different from asking how long you would have to wait, on average, for just any 5 mutations to occur. We already know that, we waited 500 generations to get 5 mutations. If you just need to wait for any 5 mutations, you just need to wait, on average, for 500 generations to have occurred. Because on average there’s a new mutation every 100 generations.
But since we already know the first 5 mutations evolved in 500 generations, what use is it then to turn around and calculate how long we would have to wait on average for those 5 specific ones to evolve again if we started over? It doesn’t matter which 5 mutations that evolve in those 500 generations, you would always be able to calculate that it should take an extraordinary amount of time for those 5 specific mutations. And yet, we just need to wait 500 generations for some such set of 5 mutations.
Sanford and colleagues look at the differences between humans and chimps, count how many there are but then treat the waiting time problem as if these specific mutations had to occur, instead of just calculating the waiting time for the total number of them.
I hope this made the problem clear. There’s a name for what is wrong with this kind of argument Sanford is making. It commits the Texas Sharpshooter fallacy.
It’s not. What I am explaining above really is what this is all about. I agree that even if the mutations are all unrealistically beneficial it has a neglible effect on the waiting time, since we still have to wait for specific ones to occur. Making the mutations beneficial just reduces the probability of loss due to drift, and reduces the time to fixation. While those matter too on evolutionary timescales, they’re not the real problem here. It’s the specifics vs general thing.
Whether you think the mutations are “devolutionary”(or whatever you’d call that) doesn’t change the problem at all. You still have to wait for them. Sanford’s argument applies to every species that “devolves” from the pairs that leave the Ark. No creationist shows any sign of having realized this.