[DrPraetorious]

If something is drastically broken, it is better to nerf it than to upgrade absolutely everything else.

But the example I keep coming back to is Shadowfist (which is an excellet game, by the by.) They hardly ever nerfed anything - any underpowered card would get minor flavor advantages until it was competitive.

In dominions, there are a huge number of relatively minor advantages you can give that would actually enhance the flavor of units, and you can balance them to the most robust, useful members of their type, rather than taking that unit and nerfing it. I think this is preferable even if it introduces more changes, especially if it only introduces changes in stuff people seldom use.

For example - pretenders. The Prince of Death, the Wyrm, the Dragons, the Cyclops and a a bunch of bless chassis (esp the Oracle), there is a general consensus, are better than the other base pretenders.

It is easier on the learning curve to boost all the other pretenders than it is to nerf those dozen which people with carefully refined strategies actually use, even though this is more changes. That way, I don't even have to look at the mod, I can show up with my unmodified prince of death strategy and keep using it without even learning the new and improved abilities of the Titan (who can hurl lightning at 5% of enemy units at the start of combat, or something.)

You can do the same thing with nations - give them national spells (or, if absolutely needed, boost their troops) until they are competitive.

Personally, I don't think that units within a nation need to be particularly balanced against eachother. If a unit is actually useless, this is a problem, but I think it's actually *important* that some units (likewise spells, but obviously not whole nations) be niche, useful only in specific circumstances, while other nations and spells are more robust and broadly useful.

All of this is also a lot more work, of course.