That sounds like a nice idea, actually. I hate those rounds where 11 or 2 occur more often than 6 or 8. I know it can happen, but it can make the game rather annoying at times.

I actually think that makes the games more fun. Especially so when you have a larger game so that there are people on those tiles.

For me, it makes the game feel less "solvable" by basic analysis. Sure, statistically certain observations will hold. For the game you are in, though, you have to be ready to adjust your strategy based on what has happened. (Clearly not in anticipation of future rolls, but more based on the resources you have managed to get, not the ones you wanted.)

Admittedly, we already tend to take the number planning a bit out of the game by initially turning them on the other side and you can only reveal the numbers to hexes you have built at (for the inital round of placing settlements/streets they are revealed after everyone has placed both). Makes for a nice variant where exploration holds surprises, or you can expand to hexes that are already known.

I've played once with this variation. I liked it, not sure why we didn't let it stick.

Of course, I haven't played at all in a long while.

One of the nice features of the Settler of Catan assistant for the iPhone is that it will roll the dice for you, and it keeps a chart of the results. Thus one has data to back up the complaint that "Your 3 is out producing my 6!"

FWIW Catan gets even worse if you play with 1d12. It was worth a shot - there's got to be _some_ easy way to make that game better.