Do you agree with this description of character performance?
Within its original context? Yes.
The original context was that, in order to [easily] improve in the class functions of a class, one had to perform the class functions of that class. That idea makes sense.
Clerics who do not remain faithful to their deity are going to require more remedials than clerics who bolster their faith. Cowardly fighters probably didn't learn as much about fighting as bold ones. Magic-users who don't use magic probably don't understand their spells as well. Thieves who refrain from using their thief abilities, likewise, probably haven't learned a whole lot about thieving.
In actual implimentation?
Well, consider that part of the original level concept was that it rewarded players for learning the game, and it makes better sense in implimentation as well. For example, if your magic-user jumps to 8th level without really understanding his spells, the odds are good that he (and his party) will eventually pay a bigger forfeit than longer training times.
The balance in 1e was achieved by the players, not by the system per se. If the players chose to take their 1st level PCs straight to the 4th level of the dungeon, they could do so. If they had mastered the system, they might even do well in quick, hit-and-run bursts.
Training times did more than punish the players; they helped them to guage whether or not they understood their characters' abilities well enough to push into the deep levels.
1e is not 3e, nor is it 4e.
So, in implimentation, I would also say yes.