ok, well again, you said:
Bull rush is the maneuver and the key was choosing that maneuver at the time the orc was in front of the fire pit. It is cool if you didn't mean that. But it is what you said, so I hope you understand I just took you at your word.
Now you seem to want to add an arbitrary game mechanic that lets the character wave a cosmic get out of jail free card and suddenly it is a gimmie that you will succeed.
It's not a gimmie. Is it a gimmie now that I hit my opponent and deal damage?
I like action points and sometimes use them. But action points are pretty precious when I use them, and I frequently don't. And even when used they only give +1d6. But there is some common ground here at least.
I think if Pathfinder would use action points, this might be fine with me. That is a good alternate approach that it's a lot easier to implement then other approaches. And also an easy house rule should I ever run Pathfinder.
But some kind of more potent mechanic is just gimmicky to me, and not fun. It is a freebie and the opposite of heroic or cool. You may as well set aside the dice and declare the PCs "the winners". And that I see it this way is probably part of why I use APs so infrequently and stingily.
No need for hyperbole.
On the other hand, I still take issue with the "try again and again" thing. If it is really that unlikely, then your gnome wizard really needs to stop trying to bullrush orcs.
Why do you hate
Telekinesis?
And it may fail for a normal half-orc barbarian character. There is no glory in victory without the risk of failure. But I don't see anything in PF that makes me anticipate trying "again and again". I think you are building an image that doesn't represent the actual circumstances. The chance of success for a reasonable task is reasonable.
Well, I haven't run the numbers in detail, but it seems the difficulty is typically higher than hitting an opponent. It might also be problematic since it is based on the opponents attack bonus, which does scale very differently depending on level and type of the monster. (Dragons typically have an insane attack bonus, Undead less so.) And it also seems to add two ability modifiers instead of one, so it is bound to even be higher. At what level can you only succeed when you roll a 20 against most foes?
Trying "again and again" will probably simply not happen when it is too difficult.
The combat maneuvers in 3.x targeted touch AC and created an opposed roll.
The "gimmie" problem in 3.x was typically related to two things:
- The touch AC is ridiculously low. A good tactic was to use an iterative attack for that.
- The opposed roll was problematic if you could ensure counter-attempts were impossible to succeed or just didn't happen. 3E Trip negates the counter-trip, IIRC, locked gauntlets made it nearly impossible to counter any disarm attempt.
Of course, against many monsters, their bonuses to the roll were so high that it didn't matter anyway. Quadrupeds, Large or larger, high strength monsters with scores beyond the values of the PCs...
In the end, I am distinguishing between two things:
- How often can you attempt to disarm or trip.
- How easy is it to do it if you attempt to do it.
The simplest approach is just make it a high DC.
What I am saying is that the simplest approach is not the one that leads to the more interesting game.
The frequency of attempts does not need to depend on the difficulty.
A Rogue can only sneak attack when he flanks someone or the opponent has lost his Dex bonus to attack. Sneak Attack is not modeled by simply saying "the DC is 5 points higher". (There is feat that works a little like that, though. It's called Power Attack.

)
Similar restrictions could apply to combat maneuvers.
The actual difficulty of making a sneak attack is the same as hitting someone in normal melee. But you can only achieve it under certain circumstances. Now, flanking someone is relatively easy to accomplish. Combat maneuvers require something more difficult.
But the advantage of sneak attack is that it introduces an interesting combat dynamic. The Rogue is not just standing around and hoping he hits against a higher DC then the Fighter. Instead, he tries to maneuver in a position where he can use his ability and avoids being in situations where he cannot.
There are other ways to create "interesting" combat dynamics. If you required an action point to even try a maneuver, the players challenge would be to figure out the situation where he gets the "most bang for his buck". It might cause some maneuvering on the battlefield if he tries to lure his foes into a situation where bullrush-into-firepit is a possibility.
I get the impression that you see failing to bullrush the orc into the firepit as a break down of the game system. I don't see that in the least.
We are not talking about the breakdown of a game system. Just a mechanic that I think could work a lot better then it seems to do. But we are talking about this mechanic, not the entire game system. No need to generalize it to the entire game system. Combat maneuvers might be more important than a single broken or useless feat, but they are not as big as, say, the spellcasting mechanics for the major classes. Though... Maybe they should. Combat maneuvers have always been the way for me to make playing a Fighter (and other melee classes) interesting. I hated going the boring Weapon Focus/Specialization route. It was probably always the most effective route, but learning Improved Trip, Disarm, Sunder, Bullrush and so on seemed to be a lot more exciting.