In the 1990s Iron Man magazine published a series of articles from Nautilus and MedX inventor Arthur Jones. In one of those articles he wrote the following about the number of sets one should perform, citing several then-recent studies,
“When, in 1970, I introduced the first Nautilus exercise machines, together with the statement that only one set of each exercise was required, or even desirable, several people accused me of making false claims in an attempt to encourage the sale of my exercise machines; which charge, in fact, was utterly false, since my statements were based upon clearly established research results that could not be disputed.
And just what does science have to say on the subject in 1996? As it happens, surprisingly little; but most of what little has been published on that subject clearly supports my statements mentioned above.
Three days ago, on January 19, 1996, Dr. Michael Pollock, of the School of Medicine of the University of Florida, gave me a copy of a study that he had just completed, a study that I was not even aware of until after it had been written up for publication.
This study included careful consideration of the results of several research projects conducted by members of Dr. Pollock’s research staff and also considered every other study that they were able to discover by a careful review of the entire scientific literature: in effect, “everything ever published on the subject in any scientific journal.”
Results? ONE: in 1962, a Ph.D. named Berger, using 177 subjects for a period of 12 weeks, and using the bench press as the exercise being tested, compared the results of one set to the results of both two and three sets. One set increased the average strength of that group by 23.6 percent; two sets increased strength by 24 percent, only four tenths of one percent better than one set; three sets increased strength by 26.3 percent, only 2.7 percent better than one set. Whereupon Berger concluded that one set was as good as two but that three sets were better. Well, in fact, any such slight differences fall well within the differences of random variation, and certainly do not indicate any slightest differences in results.
Secondly, considering the fact that Berger was using healthy but previously-untrained college-age young men as subjects, his overall results were somewhere between pitiful and God awful. The starting strength of his subjects was very low, which means that they had the potential for rapid and large-scale increases in strength, yet failed to produce any such results. In contrast, thirteen years later, in 1975, using military cadets as subjects during a study at the U. S. Military Academy, West Point, we produced an average strength increase of 60 percent in a period of only 6 weeks; so our results were more than twice as good as Berger’s even though we trained our people only half as long as he did his. And, of course, we used only one set of each exercise.
TWO: in 1982, a man named Silvester, using 48 subjects, compared the results of one set of biceps curls to three sets. One set increased strength by 24.6 percent within a period of 8 weeks while three sets increased strength by 26.2 percent, a difference of only 1.6 percent; again, a difference so slight that it is meaningless; or, as they say in the scientific community, “non-specific,” or “no significant difference.
THREE: in 1983, a man named Stowers, using 28 subjects, during a program that continued for 7 weeks, compared one set to three sets in both the squat and the bench press. And found, again, that there was no significant difference.
FOUR: in 1986, a man named Westcott, using 79 subjects for a period of 4 weeks, while comparing one set to two sets, actually produced somewhat better results from one set than he did from two sets; but, again, the difference was not significant since one set increased strength by 11.2 percent while two sets increased strength by 10.8 percent.
FIVE: a later study performed by the same man, Westcott, performed in 1989, using 127 subjects, both men and women, and lasting for 10 weeks, using both dips and chin-ups as exercises, found almost no difference in results from one set, two sets, or three sets.
SIX: in 1993, Dr. Pollock’s group, using 140 subjects, for a period of 12 weeks, using the cervical-extension exercise, compared one set to two sets and again found no meaningful difference.
SEVEN: in 1995, Dr. Jay Graves, using 141 subjects for a period of 12 weeks, with lumbar (lower-back) extensions as the exercise, and comparing one set to two sets, produced quite a bit better results from one set than he did from two.
EIGHT: in 1995, a man named Starkey, using 83 subjects for a period of 14 weeks, with both leg extensions and leg flexions as the exercises being tested, compared one set to three sets; in both cases, extension and flexion, one set proved to be better than three sets.
AND SO IT GOES: in some cases one set was better than either two or three sets, and when multiple sets did seem to be better the difference was so slight that it was meaningless. The American Academy of Sports Medicine has now accepted, as its recommended protocol, “one set to failure, not more than three times weekly;” which, frankly, I still believe is too much for most people, and is required by nobody. I get several calls a week from strangers who tell me about the great results they are producing by only one weekly workout, or even less exercise. In the field of exercise, at least, while it is true that “some exercise” is good, it does not follow that “more exercise” is better; in fact, more is usually worse. Remember: exercise does NOT “produce” results; instead, if properly performed, it “stimulates” results.”
In the two decades since Arthur wrote this, studies have only continued to confirm there is no need to perform more than one set of an exercise:
“Single set training appears to provide similar hypertrophic gains to multiple set training. Frequency of training should be self-selected as there appears no evidence which can support any recommendation.”
“Persons should aim to recruit as many motor units, and thus muscle fibres, as possible by training until momentary muscular failure”
Fisher, James & Steele, James & Smith, Dave. (2013). Evidence-Based Resistance Training Recommendations for Muscular Hypertrophy. Medicina Sportiva. 17. 217-235. 10.5604/17342260.1081302.
“Persons can obtain appreciably the same strength gains by performing only a single set of each exercise 1 x / 2 x week, compared to higher volume workouts. Persons should train when they feel physically and mentally ready to do so. Both physical and mental fatigue have the potential to negatively affect a workout and/or muscular growth and development .No specific periodized routine is unequivocally supported within the literature.”
“Persons should train until momentary muscular failure to actively recruit all of the available motor units and muscle fibres, as opposed to a pre-determined number of repetitions.”
Fisher, James & Steele, James & Bruce-Low, Stewart & Smith, Dave. (2011). Evidence-Based Resistance Training Recommendations. Medicina Sportiva. 15. 147-162. 10.2478/v10036-011-0025-x.
“Overall, it is clear that the great majority of well-controlled, peer-reviewed studies support Jones’ (15,16,18-20) contention that one set per exercise is all that is necessary to stimulate optimal increases in muscle strength and hypertrophy. Though there are exceptions in the research literature, these are few and most suffer from confounding variables and, in some cases, blatant experimenter bias.”
Smith, Dave & Bruce-Low, Stewart. (2009). Strength training methods and the work of Arthur Jones. J Exerc Phys online. 7.
“Several researchers have recently claimed that a series of meta-analyses unequivocally support the superiority of multiple sets for resistance training, and that they have ended the single versus multiple set debate. However, our critical analysis of these meta-analyses revealed numerous mathematical and statistical errors. In addition, their conclusions are illogical, inconsistent, and have no practical application to resistance training.”
“The preponderance of studies suggest that there is no significant difference in strength gains as a result of performing a single set or multiple sets of each exercise.”
Otto RM, Carpinelli RN. A Critical Analysis Of The Single Versus Multiple Set Debate. JEPonline 2006;9(1):32-57.
“There is very little peer reviewed evidence to support a high volume strength training protocol.”
Carpinelli RN. Berger in retrospect: effect of varied weight training programmes on strength. British Journal of Sports Medicine 2002;36:319-324.
“Reviewers, editors and publishers are inclined to reject studies that show no significant difference between specific training protocols such as the effect of single versus multiple sets on strength gains. Statisticians call this the file drawer effect or publication bias; that is, the editors cherry-pick the studies for publication that report a significant difference between protocols. The result of this file drawer effect (studies not published) should be that after a complete search for all the published research on a specific topic, the majority of the published research should be skewed toward studies that reported a statistically significant advantage of one training protocol over another. However, on the specific topic of the effect of single versus multiple sets on strength gains, the majority of published studies – even considering the potential file drawer effect – reported no significant difference between protocols”
Carpinelli, R. N.. “Critical Review of a Meta-Analysis for the Effect of Single and Multiple Sets of Resistance Training on Strength Gains.” Medicina Sportiva 16 (2012): 122-130.
Something else which is important but rarely mentioned is the quality of sets performed by most trainees and demonstrated by some prolific researchers, is terrible compared to how exercises can and should be performed, and this also factors into how many are required for best results.
