CMB homogeneity: Inflation vs. common initial conditions? As I understand it, inflation was proposed to provide a mechanism by which now-causally disconnected regions of the universe could have once been in causal contact, and thus explain correlations in CMB temperatures between regions that shouldn’t have had the ability to exchange information.  
My question is simple:  Why not instead propose that the initial conditions were sufficiently the same in distant regions to result in similar temperatures in areas not causally connected?
Forgive my ugly analogy:
If I find 2 coffee mugs in a house, one on the first floor and one on the second floor, and they are both exactly 200 degrees F, I can propose either that:
A) These mugs have once in the past been in causal contact, and thus had time to equalize their temperatures
Or
B) Both mugs were exposed to sufficiently similar conditions to result in similar temperatures.  
In this case, my microwave. 
The latter seems by far the better choice; why have we settled on Choice A to explain CMB temperature correlations instead?
 A: In your option B, your suggestion of a common microwave oven in the past of both coffee mugs is an example of causal connection. That is, each mug is in the forward light cone of the microwave oven. The situation with the universe is that on the huge surface from where CMB was emitted (I mean the part of the CMB now arriving at Earth) there are many different parts, all in a thermal state at the same temperature (to a precision around $10^{-5}$), but which are not in the forward light cone of a single past event, unless you introduce into the model of the Big Bang something like an inflationary scenario. This is not to say the inflationary scenario is the only way to do it; but it does indicate that the problem is not easy to solve.
There is nothing wrong, from a logical point of view, with arguing that similar conditions held in different parts of the universe in the early stages of the Big Bang without the need for those parts to have a common past or to have been influenced by events in a common past. This has not been ruled out as a possibility, but people are interested in exploring things like inflation because it offers a model which can tackle several puzzles at once. However it would be incorrect to say that inflation is an established fact; it is more a research program that merits further study.
A: Well, either you think that homogeneity is a fact in need of explanation, or you think it doesn't need to be explained at all. If you don't think it needs to be explained at all, and it just is that way, then that's a fair position to take, but you can see the other answers for why most people disagree.
But if you do think that homogeneity has to be explained, your solution (B) is not a solution at all. It is true that if the situation is homogeneous at time $t$, it could be explained by having homogeneous initial conditions at time $t-1$. But that's not a real explanation, it's just kicking the can down the road, because then you have to explain why the situation was homogeneous at time $t - 1$. (It's turtles all the way down.)
Or you might say homogeneity is because the visible universe came into thermal equilibrium with something else which was homogeneous. Again, that just moves the problem one step back because you have to explain why that other material was homogeneous. Once you start trying to propose any explanation at all, nothing is going to work unless you change cosmological history, e.g. by adding inflation.
A: I just realized your (A) and (B) are the same :)
In your analogy coffee mugs were in the microwave at initial time. Then inflation kicked them apart. Non-inflationary expansion is not enough to make as big a universe as ours from one causal patch.
If you open the wiki page on inflation, you can find three main points that motivate cosmological inflation: (1) horizon problem (the one that you are talking about), (2) flatness problem, (3) monopole problem. It should be clear enough there, but let me know if you have further questions.
