I hear this criticism bandied about a lot. I still think it's nonsense though. If there is a cookie cutter similarity to point buy characters you see, the problem lies at the conceptual stage not at the distribution of stats.
If every paladin is a single classed front line bruiser with a bastard sword and large shield, they'll all look the same with point buy (and with rolling too). On the other hand, if a player has a nifty concept for a dextrous rogue/ranger/paladin/Consecrated Harrier in one game and wants to play a melee focussed paladin/sorceror in the next and goes for a warhammer and shield focussed fullplate bruiser paladin in the next and a greatsword wielding, mobility (and spring attack) focussed fighter/paladin in the fourth campaign, you'll have four different sets of "optimal" stats to account for the difference in planned abilities and tactics. (And if you start complaining that these are examples of cookie molds because they're stat based remember you're the one complaining about stats being the same--I'm just pointing out there's a reason for them to be different; of course all the characters will be monotonous despite their mechanical dissimilarity if they're all played the same. But what does that prove? Only that the monotony is not a function of mechanics but of concept which is what I said at the beginning. (I do think mechanics ought to mirror the concept though)).
And if players have cookie cutter character concepts (did you one better on the tongue twister

) they'll have cookie cutter characters whether the game is point buy or dice rolling. If players don't have cookie cutter concepts, they'll have different characters. Players are perfectly capable of looking at the die rolling output and deciding which of the optimal cookie molds the set of scores fits into ("Well, I've got one really good score and everything else is lousy--I'll have to be a cleric or a wizard (a gnome wizard with a toad familiar)"). In fact, point buy gives players who want it the freedom to explore character concepts at will rather than depending upon the luck of the draw. (It's hard to play a Forrest Gump cleric if you roll all 14's or to play an Errol Flynn swashbuckler if you roll a 9 for charisma, or to play a Maximus clone if your constitution is too low to take the frequent beatings--thus rolling for stats limits the number of concepts available to any given player at any given time).
In any case, why do cookie cutter stats necessarily equal boring characters? Every 20th level fighter in the world may have a +5 weapon. That doesn't mean they're all the same. I don't think anyone would say that all paladins are the same even if they all got +5 armor of heavy fortification and a Holy Avenger sword. And if having the same items doesn't make Glorfindel into Aragorn, Midnight into Elminster, Artemis Entreri into Drizz't, or Cattie Brie into Legolas, why would having a 16 dex and a 15 wisdom make two elf clerics into copies of each other? If one's a LG cleric of Torm raised by humans and the other is a CN cleric of Shevarash bent on vengeance I'd say there's still a world of difference, wouldn't you?
Halfabee said:
Point buying, on the other hand is akin to making cookie cutter characters (say that fast, three times
). It may work if the only goal is to create a character which does well in combat (it seems that a lot of folks would like to run an LP model to determine what the optimum stats would be for each class, and run with that . . . while following the optimal progression of skill and feats to arrive at the most powerful character in the end). While I admit that I like to have a character develop into a powerful hero-type, the approach I see a lot of today's gamers taking is purely mechanical.