There's a lively discussion going on over at Monsters and Manuals, prompted by Noisms' question about why we fantasy gamers care so much for historical research, when such research has yielded very little evidence of dragons, orcs, or hippogriffs.
My take on it is that it's in the genetic code of our hobby.
Remember that before there was D&D there was Chainmail, and Chainmail was a wargame. The wargaming hobby is steeped in the tradition of historical research and analysis. Every game designer looks at what came before and tweaks it. What if we changed the focus from battalions to companies? What if we changed the combat dynamic? What if we consider the logistical challenges of this campaign? All of this requires research to fill in the blanks that the game designer wants to fill.
Good wargames have serious research behind them. Some are so good, they get used as reference materials. Some grognards even advocate the use of wargames to more thoroughly study and understand historical realities. The Holy Grail of wargame research is discovering that there was an apple orchard outside of Leningrad that forced a Panzer division to approach from a different direction.
Chainmail developed as a way to simulate medieval war, bringing combat stats to individual soldiers. Gygax then took the next step and said, what if these guys went and fought a dragon?
And Gary saw that he had made fantasy role playing, and it was good.
Even if we fantasy gamers look to different sources, this spirit of inquiry and invention runs strong. The same sorts of folks that like to ponder these questions are drawn to these different branches of gaming for this very reason - it's just that some of us like to fight Napoleon and some like to fight orcs. In the end, we're all branches of the geek tree.