All right.
So I was in this 3.5 game where the DM allowed me to play an alchemist (a class from pathfinder). It mostly worked, but there were a few "PF fixes" to the 3.5 systems that aren't in 3.5 and it was at time a bit OP, in sometimes unexpected way. That way was alter self.
See, in that edition alter self gave you the natural AC of the creature. We eventually realized that the best choice was troglodyte: +6 natural AC, while still able to use weapons etc. On top of that, because I was an alchemist, I could share my infusions with someone else, so I gave it to my barbarian cousin (we were 2 dwarves, said barbarian was played by my spouse). Also, the AC bonus stacked with the AC bonus the alchemist's mutagen gave you. My dwarf could also use a shield and just had massive AC (about 35 around level 7?)
We did it so often that the GM just gave the barbarian a "troglodyte shape shifter" prestige class to the barbarian he made up.
As an aside, this is the game where another player played a bard with middling dex, light armor, took no defensive magical items, cast no defensive spell and had no con bonus. And back then bards had 1d6 hit dice... she had awesome party boosting powers, but one stray arrow and she was in mortal danger! It really opened my eyes to how intra-party balance was important, and how the importance of system mastery in 3.X was leading to some rather serious problems...
However, I should mention that I told this story to a fellow GM recently and he stated that this bard player was good for the game because she created opportunities for dramatic combat. I guess it's a different way of looking at it, and I'm still thinking about this different perspective...
