Crunchy, but in an icky way

I have been reading 13th Age, a roleplaying game built on the d20 open gaming licence.  It is clearly a DnD clone that sits somewhere between 3rd and 4th edition incorporating elements of both.  That isn't all though as it certainly brings lots of new things of its own to the table, but you can't ignore the origin of the thing.  It was pitched to me as a crunchy game with a lot of combat tactics and that seems true but I don't much like the way it does things.

One of the issues 13th Age has is that it relies on rapidly scaling Armour Class to improve character power.  Overall from level 1 to level 10, which is the max, character AC rises something like 13 points.  That means at higher levels weak enemies simply cannot affect you, and moreover that you go from being hittable to being invulnerable in one big chunk.  The problem here is that the die simply isn't big enough.  When people are rolling a d20 to hit that first 2 points of AC might reduce the enemy to hitting you on 13 instead of 11, but that last 2 points of AC means that instead of hitting on 19 they cannot do anything at all.  13th Age tries to deal with this by controlling hit bonuses and AC very tightly.  Items provide +1 to +3 bonuses based on level, stats all start and end at the same values, and everybody gets the same bonuses from level.  The system is so fragile that they box every character into a corner and keep them all the same in order to keep the numbers from breaking.  This means that everyone is on a magic item treadmill, their stats are basically locked up with no flexibility, and they all have the same bonuses.  Boring!

They also imported a lot of things from 4th edition that I really dislike - things like having multiple stats determine the bonus to a particular defence based on which is highest.  That isn't a balance complaint, but just a feeling that raising my Constitution should have a consistent effect, rather than "Well, if I Constitution is my middle stat of Strength, Dexterity, Constitution then it does this thing, but if it is my low or high stat, it does nothing."  That just feels bizarre to me, an example of someone trying to balance things so hard that the system feels wrong.

One thing that they do in this game which I really like is the idea of failing not actually being about failure but rather complication.  If for example you fail a Diplomacy check you still succeed in what you were trying but something weird happens.  The queen might agree to pay you to slay the monster, but she decides to send another group too and give the money to whoever gets the critter first.  I feel like utterly failing should be a real possibility but this does give me an idea for how to handle most failures in a more entertaining fashion.

In Heroes By Trade people roll 1d8 and add their bonus to perform a Skill Check.  My theory is that I could add the rule that if you attempt something and fail you will have a choice:  Either accept failure and move on, or roll an additional 1d8 and add it to the roll.  If that higher check succeeds then you do what you intended but regardless something bad happens.  Basically you are taking risks and being reckless, which is a problem, but might allow a stunning success.  Maybe you pick the lock but you make a ton of noise doing it, or you leap the chasm but your weapon falls off halfway across and is lost.  It feels like this would give the GM a lot of opportunities for hilarity and make it possible to set up really difficult checks that the players can fail at in ways that forward the plot and add interesting twists.

This might even be a good replacement for Fate Points, as I haven't been entirely pleased with them in testing recently.  Using them just doesn't have the visceral impact that I want and they don't feel good in combat particularly - plus stocking up on them makes players feel invincible.  The main thing I really liked in terms of Fate Point use was coming up with crazy ways to succeed at very difficult tasks and taking a complication to achieve something super heroic strikes me as a much better way to approach it.