First, Oblivion. This RPG scales monster difficulty roughly in line with player level, to keep the game challenging. Player level rises to follow usage of the player's major skills, and is independent of usage of minor skills. For many players, this works ok, but for many others, rises in major non-combat skills result in the monsters growing stronger more quickly than the player does, making the game harder, even when travelling in areas that were previously safe. Ultimately, the player is forced to emphasise use of their minor skills, or to stop electing to level up, in order to stay competitive - surely not what the designers had in mind. Scaling monster difficulty according to player success as opposed to player level is one way they could have improved this. Or they could have stuck to the geographically-bounded difficulty levels that work well in most other RPGs.
Secondly, Empire Earth. This RTS attempted to bring more tactics to the genre by adding explicity Rock-Paper-Scissors mechanics to the game. Spearmen beat shock troops (range advantage), shock troops beat archers (shock troops carry shields and can chase down archers), and archers beat spearmen (spearmen cannot block missiles). All well and good in theory. However in practise - at least in the single player scenarios - the opponents tend to come at you as one heaving mass of all three unit types, meaning you may as well do so as well, or just pick one type arbitrarily, and hope for the best in each case. This mechanic might well work a lot better in games like the Total War series, which give you more time to manoeuvre individual troop types. Maybe it works better in Empire Earth 2, as well?
It's a shame when actual gameplay doesn't match the implied 'correct' gameplay as directed by the overall system. It implies a design element that wasn't fully thought out to the degree where it would be guaranteed to work.