written by Charles I. Staley, B.Sc., MSS
The notion of "training to failure" is perhaps one of the most revered practices in the modern bodybuilder's toolbox. But interestingly, this training method seems unique to bodybuilding. In other iron sports, such as Olympic weightlifting, powerlifting, and throwing, athletes develop enormous levels of muscle mass without training to failure, at least not in the way that most bodybuilders would define it. This observation, coupled with the fact that many elite-level bodybuilders do not embrace this practice, warrants a second look at this concept.
Birth of a Paradigm
Many credit Arthur Jones (the inventor of Nautilus equipment) with developing and popularizing the one set to failure paradigm. Jones argued that bodybuilders should work to the point of momentary failure, using one set per exercise/per session, rather than using multiple sets of multiple exercises. But Jone's commercial success may been potentiated by a long-standing tradition among young trainees (particularly men) who, in the absence of qualified supervision, regularly trained to failure as an intuitive way of obtaining objective feedback about their progress. Whenever an additional rep could be performed with a given weight, the trainee was psychologically reinforced, which further entrenched this habit.
Unfortunately, it also reinforced poor exercise form and the tremendous frustration that set in when, after several months of monotonous training, the inevitable plateau set in. This frustration then paved the way for numerous ill-conceived commercialized training systems that emerged over the past several decades. The result is an endless cycle of unsupervised trainees switching from one miracle method to another, in an endless search for the "perfect program."
Before we criticize Jones or the authors of the many programs available today, it may be necessary to revise our expectations of what a training method should and shouldn't do. Remember that nearly any training method can be effective, at least temporarily, for the following reasons:
1) Beginners will make short-term progress with any training method, provided they aren't injured in the process.
2) Many people train in a very monotonous manner, rarely changing acute exercise variables such as choice of exercise, order of exercise, rest periods, and load (volume and intensity). When such a person changes programs, they will progress, at least temporarily.
http://www.timinvermont.com/fitness/failure.htm