Personally, I've never had a problem with that sort of metagaming. Look at it from a character perspective: the character knows how to put a little more oomph into their attack. If they're a bit worried they can use a bit of focus and energy to make an attack more potent, but they can't keep it up. The character is experienced enough to know what that extra oomph costs, and experienced enough to know what that extra oomph will give.
In the situation above, they character has also been dealing with the given foe for at least a few rounds. The character knows how hard it is to connect with the enemy, and the character knows how well he's doing in a given round. He probably knows when he's whiffed it real bad even as he's beginning the swing - if a baseball player fouls up his footing as he takes a swing, he's often going to know he missed before the ball crosses the plate. That character also is going to know whether or not that extra focus is worth the effort. If you rolled a three (say, in game, your character stumbles briefly as he tries to make the attack; not enough to put him in danger, but enough to seriously mess up the attack). When you've already failed that much, the character isn't going to exert the effort to try and save the attack, he'll just prepare for the counter. On the other hand, a character is also going to have an idea of when he's close, but not quite close enough - if you managed to get past a blocking weapon and shield, but the sword is still going for a strong bit of armored shoulder, then the extra oomph might let you make a last second twist and get a less-well-armored spot.
A player basing his decisions on actual experience with a given enemy really isn't metagaming. The player and the character both know what options they have, and have an idea of how strong the foe is. Expending resources in line with that is perfectly fair.