I contend that unless a lifter is testing a maximum a ‘buffer’ should always be employed between the training load and recent (repetition) maximum.
As rep number decreases the magnitude of the buffer should increase since load is closer to absolute maximum (i.e. Doubles require a greater buffer than 5’s). Additionally, as lifters develop neural adaptations occur that first permit increased fibre recruitment and rate coding, and later decrease golgi and antagonistic inhibition. THese adaptations increase maximum strength but also increase the central nervous system stress of approaching maximum. The magnitude of the buffer should therefore also increase as the lifter improves.
WHY?
When loads approach maximum there is often a tendency to increase the speed of the eccentric phase and to over utilise cheat methods, e.g. flexing spine during the squat and lifting the ass during a bench press. Greater loads can then be lifted but greater muscle forces have likely not been achieved. By using a buffer, technique is (more) consistent and injury risk is substantially deceased.
The emotional requirement to lift at maximum means central nervous system (CNS) stress is significantly greater than for a sub maximum effort. (Except when a lifter has learned to emotionally detach from his training and accept a performance decrement versus his competition maximum). A maximum effort generally precludes any further quality sets and may also detract from subsequent training sessions in the week. As a result volume may be reduced below the threshold required to generate (or sustain) an adaptation.
There is also research that suggests prolonged training in limit strength exercises (e.g. squat & bench) close to 1RM causes a decrease in speed (Rate of Force Development) that manifests before any decrement in the specific exercise. This is obviously important for power athletes such as sprinters and throwers. The Olympic lifts and their derivatives do not carry this particular risk since because of their high speed component any decrease in RFD will manifest itself immediately.
Lastly, an experienced lifter recognises that day-to-day variations in strength are quite significant. On a given day therefore a 90% effort can, in severe cases, actually represent a 100% effort. Percentages in any program must therefore only be a guide and the athlete should recognise this and have the discipline to decrease the load accordingly. Although rarely available, force plates could provide valuable data regarding bar speed. Perhaps this is how training loads will be determined in the future? (See Shaun Pickering article)
Drawing heavily from available strength literature, I illustrate below an example 4 week cycle (3 sessions per week):
Wk 1
5x3r @ 90% of 3RM (i.e. 10% Buffer)
Wk 2
1,2: 5x3r @ 92% 3RM
3: 3R @ 100% 3RM
Wk 3
6x2r @ 90% of 2RM
Wk 4
1: 6x2r @ 92% 2RM
2: 6x2r @ 95% 2RM
3: 2r @102% 2RM
For simplicity I like to translate percentages into loads, i.e. for a 1RM of 75 to 125k, 10% = 10k.; 1RM of 125 to 175, 10% = 15k etc
Please note, I am not claiming this is a novel idea or that it is particularly complex but am simply putting down in words my own interpretation of available texts. I feel that had I the discipline to adhere to these guidelines I would have progressed significantly further during the last five years. ‘Great athletes don’t always make great coaches’…well here it seems that the opposite is true - great (?!) coaches don’t always make great athletes