How accurately do athletes estimate their limit during strength training?
Athletes do not estimate their limit during strength training very accurately at first, but that judgement improves with experience and training. Those who want more precise control get more from a personally calibrated speed model than from feel alone.
Athletes frequently misjudge their limit during strength training. Simply sensing how many repetitions are left is not very accurate at the start of a training programme. After 8 weeks of regular training that judgement improves, most clearly for the bench press. For squats it remains more difficult.
A more objective approach is to measure lifting speed: the slower you move the bar, the closer you are to your limit. That relationship is stable and easy to measure. With enough rest between sets (5 to 10 minutes) this gives an acceptable prediction. But with short rest periods, and therefore with fatigue, the prediction becomes less reliable. The exception: athletes with more than two years of experience pushing to their absolute limit maintain reasonable accuracy even then.
The choice of calculation method behind the speed model matters. Athletes who determine their personal limit using multiple measurement points and a mathematical fit to their own performance data get more accurate predictions than those who keep it simple with just two measurement points. The type of exercise (upper or lower body), the equipment used, and a rest period of 5 to 10 minutes make no further difference to accuracy.
In practice this means it matters little whether you train to muscle failure or stop with two repetitions still in reserve: muscle growth and strength increase comparably with both approaches. There is a small signal that training to failure produces slightly more muscle growth and has a favourable effect on jump height, but those differences are small. Anyone who wants to use lifting speed as a guide gets the most out of a personally calibrated model and sufficient rest between sets.
This answer draws on two studies: a combined summary of several earlier investigations into lifting speed and repetition prediction, and an 8-week study in trained men and women that compared training to failure with training while stopping two repetitions short of the limit. The evidence for the lifting-speed method is most broadly supported; the evidence that the ability to judge remaining repetitions improves over time is still limited to one study.
How do you adapt strength training smartly as you get older?
Strength training demonstrably works for older adults, and you can adapt it smartly with two sessions per week at 70 to 79 percent of your maximum, two to three sets per exercise, combined with balance exercises and a little extra protein.
Is strength training still safe and worthwhile after the age of 70?
Strength training after 70 is both safe and worthwhile: the evidence for greater muscle strength and reduced fall risk is strong, and even short, light sessions already produce measurable results. When training heavily, it is worth keeping an eye on the potential impact on your blood vessels; blood flow-restricted training with light weights is then a well-supported alternative.