Eclipses can be predicted with sharp accuracy, and that applies backwards as well as forwards. There are confounding factors, but they are likely to be tiny. The earth and moon can wobble a bit and neither is a perfect sphere, and we can't know exactly which sides of the two were facing each other at the time of an eclipse. The higher your altitude the longer the eclipse, by a tiny amount (to do with optics, all those ray diagrams you used to draw at school). Topography on the earth and moon can cause tiny diversions from the expected. The most likely bigger error in looking at past eclipses is getting the track wrong by just a few miles. On the whole, you can probably trust the date, time and location given for any past eclipse to a high degree of precision. Can't speak for any before 4004 BC, of course. ;-)