Recently someone on one of the casting groups commented that there was no such thing as a cement that can have all of the water baked out of it; I believe this was done in defense of the use of Portland cement in a homemade castable refractory. The official answer would be that the chemically locked portion of water in lime based cement cannot be baked out, but the chemically locked portion of water in refractory cements can be. I have always accepted this official version as the only reality...in the past. The limiting factors on "baking out all the water" are that the refractory must be taken to yellow heat in the first place, and that over time, water content can recollect in refractory if you're not careful to seal the refractory surfaces against water vapor in ambient air.
However, truth of any kind is seldom found effortlessly; including technical "truths". I suspect that a healthy debate, with both sides airing their views, might adjust what the majority of us accept as practical reality--to our mutual profit; this could be important for people wanting to make insulating refractories as secondary layers. Homemade refractory as an insulating secondary layer might be quite forgiving of official standards; standards useful for hot-face layers may constitute a waste of money in secondary refractory layers.
While casually dismissing the Portland cement idea, I have noted both resentment, and a strong hint of "I'll match your official facts with personal experience" in passing (heated) comments from the other side of this issue. Isn't it time they had a FAIR hearing; something open minded, maybe?