You've just described a clear majority of pre-modern battles, so we're walking a bit of a tight rope as far as the rule against trivia-seeking questions goes. If you don't mind, though, I'll talk a bit about the kind of sources pre-modern historians use.

So, historians use primary sources, which are documents, artifacts, monuments, art, and such, from a given time period to produce scholarship about that time period. How much evidence is available, and what kind it is, varies with time and place. You'd probably be surprised how limited this evidence can be. As a general rule, the farther back you push it, the sparser the evidence becomes. Go back as far as the second millennium BC and you're likely dealing with increasingly limited evidence: relatively brief engravings on steles and tablets, not lengthy narrative accounts. There are any number of ancient Mesopotamian battles that we know almost nothing about, other than that they supposedly occurred.

However, that's not always the case. Some classical battles are supported by either multiple sources or archaeological findings or both. Relatively recently, archaeologists working in western Germany turned up the probable location of the battle of the Teutoburg Forest, where Varus' legions were famously wiped out more than 2,000 years ago. Likewise, there are more recent battles that we know much less about. I do 10th-12th century Normandy and England, and the early part of that period just plain sucks for a military historian. Chroniclers love to list battles, but other than specifying who won them, they almost never go into any real detail unless it was a VERY important fight. Hastings (and the Norman Conquest generally) is the exception, rather than the rule, in that it spawned a lot of detailed accounts which can be used to mostly reconstruct the campaign. We know more about Alexander the Great's army than we do about most medieval armies prior to the 15th century.