In the third post in this series (see Part 1 and Part 2), I said that despite the risks presented by being blindsided with the Craig Con, it would be a pity to avoid face-to-face debates with religionists altogether because such exchanges are often sponsored by religious groups and the audience often consists of religious people. These debates can serve as a means of reaching audiences who may not otherwise hear the scientific and atheist point of view.
I always try to keep in mind that I am not trying to persuade my opponent. That is impossible. Although I am addressing him or her, the real target of my remarks are the members of the audience many of whom are there hoping to actually learn something. Even in the most partisan audiences there are a few open-minded people who can be influenced by you and can be swayed. It is that fact that prevents me from trying to adopt the same tactics as the Craig Con but in reverse, seeking to wow them with sophistry and debating tricks. As soon as one does that, as Marcus Ranum pointed out in a comment, one has lost the moral high ground by signaling that you too are not really interested in getting at the truth. Besides, to play that game successfully, one needs the quick wit and prodigious memory of a Christopher Hitchens, and not all of us do.
There is no alternative but careful thought and negotiations prior to the actual debate.
The first thing to do is to try and negotiate a better format. The traditional one consisting of a long opening speech by each person with one shorter follow up is not the best because these kinds of set pieces enable the debater to simply throw out highly rehearsed arguments in a scattershot fashion that can impress audiences. It is based on high school and college debate competitions and is not meant for deep exchanges but to enable fairly uniform scoring rubrics by judges to quickly determine the winner.
What one needs to do is to negotiate a more conversational format with frequent short exchanges, as was suggested in a comment by Mike D. Although on the surface this looks friendlier, it allows for the equivalent of a legal cross-examination, where you can pin down the opponent to the main point and not allow him or her to skip from one to another without substantiating any.
The next thing you need to do is not allow your opponents to frame the debate either by allowing them to determine the question and/or by speaking first. This is a favorite tactic of Craig who often initiates such debates and asserts the right to choose the topic in such a way that he has the affirmative position and gets to speak first. By doing so, he can not only set the agenda for the debate, he can try and put you in the difficult position of having to prove a negative. He then gives a laundry list of ‘arguments’ and insists that you rebut each and every one in addition to making your own case. Even if one can offer a sound point-by-point rebuttal, the person playing defense is usually seen as the loser. So the decision on who should speak first should not be treated casually, and neither should the wording of the topic. If you should get trapped in that situation, you should start by simply dismissing his long list of debating points as a childish form of debating and say that you want to have an adult discussion on an issue of substance.
The third thing is one that has already been emphasized thoroughly and that is the need for careful study of your opponents’ argument content and style. Nowadays with the internet this is not hard to do. As Mike D says, Craig’s style and content have been honed over the years in many public venues that can be found quite easily.
But despite all these efforts, one should still prepare how to respond if one is surprised by a novel argument supposedly from science that comes out of nowhere. One should not get stuck on the particulars of that issue but instead use that as an occasion to quickly pivot and teach the audience something about the nature of science, that its conclusions depend on the starting assumptions and the paradigmatic framework in which it is done.
As philosopher of science Imre Lakatos showed, results in in science often require using auxiliary hypotheses and accepting certain background knowledge as unproblematic. Any conclusion is conditional on those and anyone advancing a proposition has to be able to specify all those other things if they are to be taken seriously. You use this to turn the question around by saying that scientific conclusions can only be taken seriously if the theoretical framework in which it is being made, and the assumptions being used, are carefully laid out and it is made clear how the conclusions follow from them. Unlike in mathematics, in science the axioms are not only not that straightforward, they are prone to much greater challenge. One should immediately ask the religious person what the assumptions were that went into the theorem, how the theorem follows from those assumptions, and what is the evidence in favor of the assumptions and conclusions.
One can also make the point that if the conclusions quoted are as conclusive as your proponent claims they are, then the scientific community would have accepted them and there would be nobody working on alternative areas, just like no one now works on the phlogiston theory of combustion or the ether theory of electromagnetism. The fact that this is not true shows that the conclusions are far more tentative than what your opponent is implying.
If one is dogged and skillful enough one can force the religionist to concede ignorance on these issues because they will not have the necessary technical knowledge. Their excuse will usually be lack of time, at which point one can simply dismiss the argument with something along the lines of the famous quip “Anything that is asserted without proof can be dismissed without proof”.
While all these preparations may not fully protect you from the Craig Con, they will at least force the religious person to go beyond superficialities and at the same time can be used to educate the audience on the actual way that science is done.