What Diamond is saying is that he thinks that people really should not think that history is so much different from the sciences. He is saying that the study of history can be pretty "scientific" if it is done correctly (and that some sciences are not as precise as things like chemistry).
Diamond is trying to argue that historians should use what he calls "natural experiments." He thinks that there are many "experiments" where there are two or more places that are pretty similar in many ways but which then turned out differently (like his experiment in Chapter 2). He argues that historians can use these -- they can look at what was different and what was similar and how the differences led to the differences in the outcomes.
So he is saying that historians should do more stuff like what he did in Chapter 2. If they do that, he says, history would be more precise and it would perhaps be more useful (look at the last sentence of the book for a nice summary of what he means).