...
For me, optimizing, past casual 'pick useful options' stuff(which is still upper middle opt), optimization is a lot of work on every side of the table. It doesn't often expand capabilities to enjoy new types of adventures, the most common optimizations are directly related to combat, what with the numerous number of ways to dedicate to enhancing damage, and accuracy, while capability expanding options tend to actually close off types of gaming even as they open up new ones.
All that sums up is greatly increasing the amount of work to make enemies provide a challenge without being capable of finishing the party off in a round, providing challenges they cannot trivially bypass, providing events for when they use special abilities...in short, work.
Those are all the right reasons. But, it does nothing to address the character concept elimination problem.
Permit me to be more precise. I doubt many people would consider the following build low OP: Druid X with Natural Spell and a reasonably intelligent spell selection. I don't know if people would consider that high OP, or really high OP, or just baseline OP. Personally, it's kind of the baseline I operate against, but I make no judgments on it.
Now, if that's not low OP, then saying "I'm running a low OP game" cuts off that build. The relevant question is does it cut off that concept. Is there a credible way of building a nature-themed spellcaster that would be low OP? If so, then there's no problem. None readily occurs to me, though. And, if it takes massive levels of system mastery to do it, then it sort of defeats the low work, simplicity point.
That would be the advantage of my personal default, which would be probably "practical OP." Not the most powerful or TO characters ever (obviously broken), and not even the most cracked out version of God Wizards or whatever. But, that Druid character is fine, as would be something even slightly more OP than it. The reason being that while Ranger 20 is (typically) a crap build, one can build something that credibly fits the ranger concept -- and numerous variations on it -- with not too much work. Hell, a Tiger Claw-oriented Warblade gives you one cut on it out of box.
Two caveats. As I mentioned earlier in this thread, this perspective may be guilty of myopia. I have a value judgment -- playing the widest array of genre-appropriate and mechanically-interesting concepts is a good -- which I think is uncontentious. And, I have spent most of my time figuring out ways to make badass versions of the ranger concept (and monk, knight, etc.) within the admittedly flaw framework of 3E D&D's ruleset. And, I freely admit that using optimization to paper over the game's flaws in this regard is dangerously close to an Oberoni stance.
Second, in practice, none of the above really matters. Despite what I've just written, I'd happily play or run a low OP game. Wouldn't bother me in the slightest. I'd want to spend a few minutes to get a rough handle on what we meant by low OP, but after, game on. And, the reason is b/c I have had and continue to have plenty of opportunity to play Aberration Wildshape Druids or Supermount equivalents where the dragon mount is the character. So, it's no big loss for me.
Indeed, as I was just discussing with my girlfriend, participating in a low OP game, at least an intelligently designed one, doesn't strike me as much different than participating in any other game. I come up with a concept and look for options within the game to realize it mechanically. And, invariably, I choose not to use some options b/c I deem that they would be too powerful in my judgment. My suspicion would just be that a low OP game would just push that bar down. That has been my experience in the few I've participated in.
As a side note, I haven't found the optimization level I'm thinking of (see above) to be a whole lot of work. We tend to almost always use stock monsters with very slight modifications. I tend to up their hit points (I usually do double, but the DMs I know vary on this and it depends on the firepower of the party) and give them a few feats that I think are interesting or spell-likes or magic items. Nothing major. It's possible that these experiences are driven by gaming groups that are relatively well-behaved, and so while there are what would typically be high OP choices (druids, conjurers, supermounts, etc.) there's a lesser level of OP in practice.