While I think optimizability (I don't think that's even a real word; but oh well) is outside the scope of this particular concept, I think you may be on to something -- maybe something that actually looks at the optimizability (look at that -- I did it again ... I gotta find a new word) of each class?
just a thought.
While there is merit in that idea, IF you start to differentiate the problem, length of progression must be a factor, too. Taking two levels of Fighter is Tier 3. Taking the third level drops him down back to 5, unless you know where you're going. Taking two levels of unconnected wizard is Tier... whatever.
However, I think there is a ceiling. NO mundane melee character will ever advance above Tier 3, no matter how much you cherry-pick abilities.
Second however: I don't think this was the aim of the Tier system. The key is finding the sweet spot in the system.
Most classes tend to optimize to relatively similar levels, so it didn't seem necessary.
I don't generally agree. Perhaps one should factor in how hard it gets to optimise to - the desired level. I've seen plenty of inept wizard players. I've seen plenty of intept druid players (and been one myself). I've seen MORE inept fighter players. One of the last campaigns I played in crashed because the MUNDANE melee character (me) did too much damage at level 12 or so. Actually, because the GM was lazy, but... his excuse was that I did too much damage. So the individual experiences can be very different. I personally thought the character in question was perfect Tier 3, and NOT in the upper range of that.
So, I think druids are easier to optimize than wizards, and those again are easier than sorcs. Psions seem more complicated yet.
I think one of the goals one could work towards is defining the ease of reaching the area where satisfying play is possible. In a Tier 3 campaign, will it be easier to play a Beguiler than any mundane melee character. It will be easier to play a Barb than a fighter.
So... I have it in the back of my head that something like this was tried before. What happened to that?