Well, we are talking about TWO distinct processes here, Yaddy, both pyrolysis and the Maillard effect. The one you mention involves the reactions of amino acids with sugars to form complex flavor compounds. When one effects a charring light, you are caramelizing some of the tobacco right there which is probably why it too affects flavor.
Since it has been mentioned that baking a tin on a dashboard effects the process, and much bakery occurs in well under an hour, I am not sure I'd want to bake the stuff for hours on end. I might try baking some at a higher heat for a shorter period of time, in a sealed container able to withstand the pressure (gas is given off) and where as much air as possible has been removed first. Maybe 325° for 20-30 minutes?
But I would do it with controlled samples that I could smoke in the near future. Whatever the positive effects, let's not kid ourselves that we are altering the tobacco in a way that very likely will make the tobacco more changeable. Just as uncooked food can remain for a long time, once you cook something, it must be consumed soon after, likewise, once these flavor compounds are forced out of the tobacco, the process goes on, and they change into others and yet others over time, even if you do not do anything else.
I would not do this to any large quantity of tobacco until I knew for sure the long term outcome. If anyone has done this to any great length, they should list the brands, the conditions, the heat and duration, and the results! I'd be interested to study that list. And finally, I wonder why no blenders have offered any tobaccos that are "pre-cooked?"
Finally, I would think Virginia blends might do better at this while perhaps most aromatics and Latakia / Balkan blends perhaps not as well?
