Divergent Performance Outcomes Following Resistance Training Using Repetition Maximums Or Relative Intensity

September 12, 2018




Many people who resistance train for strength, physique, or sport performance purposes train to momentary muscular failure, or very close to it, thinking that this practice is necessary to maximize results. However, a number of studies suggest otherwise. This recent study from Carroll et al. provides further evidence that training to failure is suboptimal for improving power (e.g., jumping ability) or strength (e.g., peak force during a strength test).


The purpose of this study was to compare repetition maximum (RM) to relative intensity using sets and repetitions (RISR) resistance training (RT) on measures of training load, vertical jump, and force production in well-trained lifters.


Fifteen well-trained (isometric peak force= 4403.61+664.69 N, mean+SD) males underwent RT 3 days per week for 10 weeks in either an RM group (n=8) or RISR group (n=7). Weeks 8-10 consisted of a tapering period for both groups.

The RM group achieved a relative maximum each day while the RISR group trained based on percentages.

Testing at five time-points included unweighted (<1kg) and 20kg squat
jumps (SJ), counter-movement jumps (CMJ), and isometric mid-thigh pulls (IMTP).

Mixed design ANOVAs and effect size using Hedge’s g were used to assess within and between-group alterations.


Moderate between-group effect sizes were observed for all SJ and CMJ
conditions supporting the RISR group (g=0.76-1.07).

A small between-group effect size supported RISR for allometrically-scaled isometric peak force (g=0.20).

Large and moderate between-group effect sizes supported RISR for rate of force development from 0-50ms (g=1.25) and 0-100ms (g=0.89).

Weekly volume load displacement was not different between groups (p>0.05), however training strain was statistically greater in the RM group (p<0.05).


Overall, this study demonstrated that RISR training yielded greater improvements in vertical jump, rate of force development, and maximal strength compared to RM training, which may partly be explained by differences in the imposed training stress and the use of failure/non-failure training in a well-trained population.


Repetition maximum (RM) training has been posited to be a superior training strategy based on daily fluctuations in strength levels. However, multiple lines of evidence suggest that alternating heavy and light days and/or avoiding training to failure tends to result in more favorable performance outcomes. Notwithstanding, this is still debated in various circles of training for realizing certain adaptations (e.g., hypertrophy). In this study, RM training consistently to failure was compared to a relative intensity style of training (RISR). Here’s a great write-up derived directly from the manuscript unpacking what this meant:

..RISR training used mostly submaximal intensities (i.e. percentages of set-and-rep maximums), heavy-and-light training days within each week, and down-sets (where appropriate). The maximums for each set and repetition combination were: 100% was very heavy, 90-95% was heavy, 85-90% was moderately heavy, 80-85% was moderate, 75-80% was moderately light, 70-75% was light, and 65-70% was very light. Heavy and light training days consisted of a specific intensity reduction from Day 1 to Day 3 in the RISR group: 10% for strength-endurance and overreach, 15% for maximum strength, and 20% for speed-strength (Table 1). Loads were adjusted weekly based on estimated set-rep bests within each set-rep combination (3×10, 3×5, 5×5, 3×3, 3×2). Unlike RISR training, the RM training group used maximal loads within each training session and RM zone prescription (3×8-12, 3×4-6, 5×4-6, 3×2-4, 3×1-3). The goal of the RM zone prescription was that each subject would reach muscular failure on the final set of the exercise, indicating a maximum had been achieved. If the failed set resulted in repetitions fewer than were prescribed, the load was subsequently reduced by a minimum of 2.5%. However, if the repetitions achieved surpassed the prescription, the load was increased by a minimum of 2.5%. All other
factors not pertaining to the loading strategy (i.e. training times, rest intervals, training volumes, etc.) were controlled between groups to the best of our ability. Rest periods between RT sets were 3-5 minutes for both groups. Throughout the intervention, subjects were instructed to refrain from excess physical activity outside of training and on rest days. Subjects were also instructed to maintain their typical dietary habits throughout the intervention and to abstain from taking stimulants prior to any testing or training sessions.

Although an ideal approach is yet to be determined I personally discourage training to failure, in general. Overall, the evidence from my perspective points to no clear additional benefit of training to failure and in some cases better results have been shown for a variety of fitness characteristics by not training to failure. This is particularly the case for abilities that require high power outputs or high velocities of movement; which makes sense physiologically. For hypertrophy purposes, I feel it’s a bit more permissible but should be reserved for weeks prior to a deload and generally confined to single joint exercises. Practically, the costs outweigh the benefits in most cases. The disproportionate amount of fatigue generated and the increased risk of injury dissuade me from regularly programming using RM or failure training.

To be clear, this does NOT mean training should be easy. This does not mean that one should never train or test for a max. It does not mean heavy weight should be avoided. In fact, for best outcomes, this approach demands overload, attention to detail, consistent technique, intent to produce maximal forces with submaximal loads, weeks of very difficult but strategic training, and other typically misunderstood facets of this style of programming which is quite the arduous process. Training to failure is more simple. So, I think people gravitate toward it. It gives the perception of hard work and the perceived attainment of an effective stimulus for the session but at a cost. The two primary caveats are: a) disproportionate fatigue, and b) increased risk of musculoskeletal injury. In the words of Einstein: “Everything should be made as simple as possible, but not simpler.” Failing is simple enough and can be intelligently argued for, but is too simple in my view. Reaching the point of momentary muscular failure is quite literally training to fail. The more this happens, the more likely it is to continue to happen. Don’t get me wrong, peppering in some single joint failure sets during a functional overreach or training very close to failure (e.g., 0.5 rep shy of fail) on certain moves is likely ok in select contexts, but probably isn’t any better and may actually be worse than an alternative approach. Sometimes, a great training decision is to train just shy of failure. But, this should be strategic and systematic in my view for best results.

Considering the points above, Carroll et al. reported significantly higher “training strain” in the group training to failure by using a combination of subject’s rating of perceived exertion (RPE) and session duration. This indicates the RM group was experiencing greater levels of fatigue. Both unloaded and loaded squat and countermovement jump tests tended to be higher in the group using the relative intensity programming method. A greater, positive effect on strength (measured by the isometric mid thigh pull test) was also observed in the group using the relative intensity method. This well thought-out and executed study is another empirical example of the overly simplistic method of training to failure being suboptimal for select performance improvements.