airwalkrr
Adventurer
The next poll may very well exclude the sorcerer (and only the sorcerer) if the current trend continues. I have always found the XP cost of the wish spell extremely limiting though (and most DMs in my experience apply the XP cost liberally). I think sorcerers can do more consistently with shadow evocation and shadow conjuration than they can with wish.
However, I personally voted for bard. A half-elf bard optimized for Diplomacy at 1st-level can achieve a Diplomacy score of +13. At 2nd-level it blossoms into +18. At 3rd-level, it becomes +22 (or +24 if scrolls/potions of eagle's splendor are readily available). You can essentially stop most battles before they begin at that point. And let's not forget fascinate, perhaps the most powerful ability in the entire game. So many players and DMs overlook it, but in the hands of a skilled player, it can be used to single-handedly alter the very nature of the campaign from a very early level unless the campaign's adversaries consist exclusively of mindless enemies. A bard min/maxed for Perform and given a reasonable budget can achieve a Perform skill of +13 at 1st-level, which gives an average save DC of 23.5 and a MINIMUM save DC of 14, which is still reasonably effective. While the sorcerer might accomplish similar things via charm person, a sorcerer min/maxed for charm person will only have a save DC of 16 (effectively 11 when used in combat or trying to convince a person to do something against his nature), which is hardly as effective; this is not to mention the limitation of charm person to humanoids while fascinate works on just about anything with a mind. Once the bard gets suggestion at 6th-level, there is practically no contest, as suggestion can be used without limit according to the RAW. Even if the DM requires the rewording of a failed suggestion, a crafty player can still manage to accomplish what he wants eventually. I have seen bards dominate a campaign using the fascinate/suggestion combo alone. If you allow certain sources outside the core rules, bards can use their music on constructs, undead and pretty much anything else so any practical advantage a sorcerer might gain from non-core rules could easily be negated.
Those are just my thoughts and observations.
However, I personally voted for bard. A half-elf bard optimized for Diplomacy at 1st-level can achieve a Diplomacy score of +13. At 2nd-level it blossoms into +18. At 3rd-level, it becomes +22 (or +24 if scrolls/potions of eagle's splendor are readily available). You can essentially stop most battles before they begin at that point. And let's not forget fascinate, perhaps the most powerful ability in the entire game. So many players and DMs overlook it, but in the hands of a skilled player, it can be used to single-handedly alter the very nature of the campaign from a very early level unless the campaign's adversaries consist exclusively of mindless enemies. A bard min/maxed for Perform and given a reasonable budget can achieve a Perform skill of +13 at 1st-level, which gives an average save DC of 23.5 and a MINIMUM save DC of 14, which is still reasonably effective. While the sorcerer might accomplish similar things via charm person, a sorcerer min/maxed for charm person will only have a save DC of 16 (effectively 11 when used in combat or trying to convince a person to do something against his nature), which is hardly as effective; this is not to mention the limitation of charm person to humanoids while fascinate works on just about anything with a mind. Once the bard gets suggestion at 6th-level, there is practically no contest, as suggestion can be used without limit according to the RAW. Even if the DM requires the rewording of a failed suggestion, a crafty player can still manage to accomplish what he wants eventually. I have seen bards dominate a campaign using the fascinate/suggestion combo alone. If you allow certain sources outside the core rules, bards can use their music on constructs, undead and pretty much anything else so any practical advantage a sorcerer might gain from non-core rules could easily be negated.
Those are just my thoughts and observations.