You're assuming the fighter is the cause of the discrepancy. Why not assume that the decision the fighter makes is whether or not to hit someone on the head, and the decision the player makes is whether or not the target might be stunned? That way, we can narrate anything as the cause of the discrepancy. The hammer glanced, or the hammer hit square. The target had his teeth clenched and wasn't disoriented, or the target's jaw snapped shut and he's rattled. The blow was caught short, or the blow struck at optimum distance.
The fighter isn't saying "I'll use my Stunning Hammer Blow", the fighter is saying "I'll hit him on the head". The player is saying "I'll spend my daily use of Stunning Hammer Blow to mean that this hit will stun the target".
All the fighter knows is that sometimes, if you're lucky, smacking someone on the head with a hammer will stun them.
From his point of view, it's no different to a feat that says "On a successful hit, roll 1d6; on a 6, the target is stunned". Sometimes, if the player rolls a 6, the target is stunned; sometimes, if the player doesn't roll a 6, the target isn't stunned. The fighter doesn't get to decide whether or not the target is stunned; sometimes it happens, sometimes it doesn't.
In 4E, sometimes the player decides to use a power that stuns - like rolling a 6. Sometimes, the player doesn't decide to use a power that stuns - like not rolling a 6. From the fighter's perspective, it's exactly the same - sometimes, people get stunned, and sometimes, they don't.
-Hyp.