Ben Jeffares wrote a useful article “Testing times: regularities in the historical sciences” (http://www.ege.fcen.uba.ar/ecodes/Integrantes/Javier/cursos/PDFs/(02)%20Jeffares%202008.pdf):
The historical sciences, such as geology, evolutionary biology, and archaeology, appear to have no means to test hypotheses. However, on closer examination, reasoning in the historical sciences relies upon regularities, regularities that can be tested. I outline the role of regularities in the historical sciences, and in the process, blur the distinction between the historical sciences and the experimental sciences: all sciences deploy theories about the world in their investigations.
This is based on his dissertation “Testing Times: Confirmation in the Historical Sciences” (http://philpapers.org/archive/JEFTTC.pdf) which concludes:
To conclude, I will outline what I take to be the confirmation strategies of a good historical science. Firstly, a good historical science will utilise the understandings of regularities that the sciences in general use. These regularities will be well tested, using all the apparatus of experimentation, repeated observations, and intervention in processes that allow us to understand the relevant variables. They will be regularities that are well confirmed. Secondly, historical scientists will engage in research to determine how these regularities leave traces that can act as evidence for their occurrence. This dispersal of consequences also utilises regularities that can be tested, observed, and understood.
Utilising these regularities in dispersal allows researchers to choose between alternative hypotheses. Hypotheses about the past should have distinct signatures of downstream consequences. They should also make predictions about additional lines of evidence.
From the side of history Robert A. Hurley wrote “The Science of Stories: Human History and the Narrative Philosophy of Science” (http://researcharchive.vuw.ac.nz/xmlui/bitstream/handle/10063/2227/thesis.pdf?sequence=2) which argues that history is an epistemic narrative, different from literary narrative but essentially the same as the historical sciences.