I suppose combat XP rewards are also intended to provide a less 'bursty' form of levelup or ability gain in games with not so many quests. Bloodlines had lots and lots of dots to place, but it had so many xp giving quests and so many possible new dots at a time, that it effectively served the same purpose, even allowing you to bumrush a ability to get maximum ability-raising book benefits, or controversially to save xp to gain juicy disciplines early with no major problems; so there was variability still (some quests 'alternate solutions' put out more XP - thou mostly 1 pt - for 'harder' solutions with often more tedious - routes though, which annoys metas (things like the gang deal infiltration). But it's mostly well behaved).
I guess it can be a problem with simpler rulesets with less knobs to turn and things to fill used on a game with less quest density, like a early BG1 (though that generally had two quests per area and many more in the city).
I don't see it as a problem in a game like fallout, much less fallout 2 though.
There is always the 'simulationist' argument too, that combat is the main parts of the ruleset, so combat competence progression should be modelled as a result of doing it.
This argument sucks though. 'I have this model, that i modelled from the real world, but not like the real world works of course, since this is a game and we can't do that; but i think it's better because the real world sometimes works like this, whenever one of the many exceptions, provisos, simplifications we apply, complexities which we do not model, absurdities which we do model; do not interfere, so this model is totally self-consistent and what comes before precedes what comes after'