Minimum Effective Dose for Maximum Strength

Why Simple and Hard are so Effective

by Matt Reynolds | February 13, 2019

training the squat

Traditionally, when the training program you’ve been following plateaus and stops working, you just move to a new program; one that usually looks very little like the old program. The new program is often more complicated than the old program, changing many variables such as intensity, volume, frequency, and exercise selection all at once.

I believe training and programming should be kept as simple as possible, for as long as possible. In Starting Strength’s Novice Linear Progression, the goal is to simply add more weight to the bar every single workout, three times per week, for as long as possible. All good things come to an end, however, and eventually you’ll no longer be able to add weight to the bar every single session. At this point, most trainees move to a new, more complicated program, such as Texas Method, 5/3/1, Block Training, or Daily-Undulating Periodization. They are seduced by these shiny new programs and often ask, “Why not just go ahead and move to a more complicated, advanced program?”

But this is the wrong question. The right question is, “Why would you go to more complicated programming if simple programming will still work?” Among the wide range of programming choices, simple programming leads to more efficient short-term progress and better long-term programming decisions. There’s no reason to make things more complicated if they don’t have to be.

New complicated programs occasionally work and make you stronger. But the problem inevitably becomes one of information. If the new program worked, what variable or variables made the difference? And if it didn’t work, what should you change in your next program to see progress? This is critical because your response to training won’t look exactly like anyone else’s and changing too many variables makes it difficult to tell what changes caused what results. In scientific terminology, your results are not replicable. You can neither repeat and build on your success, nor can you avoid future mistakes, because you do not know which changed variable was responsible. Even if you experience short-term gains, it's better to plan for long-term, even lifetime, success.

What if, rather than moving to a completely new program when an old program stops working, you made a very small change, to maybe one variable of the old program – the Minimum Effective Dose of change – to maximize your return and keep making progress?

The Problem with Programming Theory

Programming can be as much of an art as it is a science. We use our experience combined with what little credible and relevant research can be found to develop optimal plans for "sub-optimal" trainees. The trainees are sub-optimal in that they often have lives outside of training: they have jobs, they have families, they go on vacation, they get sick. Hence, programming in the real world is not a random controlled trial, as there are often too many uncontrolled variables at play, and there is, consequently, no control subject. At best, it’s an inexact science, and at worst, it’s raw trial and error. Therefore, we must develop programming philosophies that are built on guiding biological principles and refined by experience. This means that our programming philosophy is a working theory with its share of holes (as is all programming theory), and anyone who says otherwise about his own programming philosophy isn’t being honest.

As with any theory, we must perform our due diligence utilizing the scientific method: We observe, we brainstorm, we consider our experience, we question and debate. Then we form a hypothesis and start testing it in the lab of the gym using quantitative data to ensure our hypothesis is on the right track. I believe that if we are honest with ourselves about the limitations of both our understanding and our gym-lab environment, we can continually test and improve our theories, making us better programmers, better coaches, and more knowledgeable strength athletes.

At the foundation of all of this lie a few key principles of physiology, wrapped up in the mechanism of biological change known as the Stress/Recovery/Adaptation Cycle.

A Brief Review of the Stress/Recovery/Adaptation Cycle

During the Stress/Recovery/Adaptation cycle, the body is exposed to a stress overload, disrupting homeostasis and resulting in a decrease in performance – during correctly programmed training, this stress is specific to the performance we are training for. The body immediately begins the recovery process, responding to the stress by adapting to it, so that next time it's applied it's not an overload stress. In order to continue making progress toward the desired adaptation, overload stress must be increased incrementally over time.

In addition, the Specific Adaptation to Imposed Demands (or SAID) principle holds that stress must be specific to the desired adaptation. We must squat heavy to improve the mechanisms that allow us to squat more weight. This requires that we consider the basic cellular processes that make up a desired adaptation.

The length of the Stress/Recovery/Adaptation cycle, from stress overload to adaptation, depends in large part on the amount of recovery needed for the adaptation to take place. The time required to recover from a stress depends on the amount and duration of the stress, the body’s previous experience with similar stressors, additional stress imposed on the body during the recovery period, and application of sufficient recovery methods during the period, predominantly adequate nutrition and rest. A host of other genetic and physiological factors also dictate recovery time and cause variance in recovery from one trainee to the next.

Sufficient stress disrupts homeostasis, resulting in fatigue. Fatigue is the actual decrease in performance (in strength training this is a reduction in force production), the net negative effect of the stressors that you are dealing with at any given time. Overload is the specific amount of stress needed to disrupt homeostasis. If managed correctly, the adaptation improves performance so that the next time it’s faced with a similar stress, overload does not occur. The amount of stress needed to constitute overload and drive the adaptation must thereafter increase. In the very short term, you may see improvement from the same stress repeated, but continual adaptation to stress comes from the targeted increase of that stress. Training means we are designing a series of overload events to drive adaptation over time.

The MED Approach

When plateaus occur and the strength adaptation stops, manipulating one variable at a time is the best option because it allows us to more effectively test our hypothesis and derive what works (and what doesn’t) based on quantitative metrics via the scientific method.

More simply, changing one variable at a time gives us good data. Changing multiple variables at once doesn’t. It can’t. The very nature of the scientific method won’t allow it, because we cannot discern which of the variables we manipulated caused the change in response. And because we view programming as a life-long process, and not as an 8, 12, or 16-week template, making small changes provides better data about what works and does not work for the lifter, which makes much more sense if we're concerned about long-term progress.

Variables to Increase and Manipulate Stress

We want to manipulate training variables one at a time to incrementally increase stress and move from simple to complex, comprehensive to individual, and general to specific. Those of us familiar with Starting Strength know how simple the Novice Linear Progression is: Just 3 nearly-identical workouts per week, and 5lbs added to each lift per workout. The SSNLP concept is so comprehensive that it works beautifully for virtually everyone. Finally, if followed correctly, SSNLP makes people generally strong, and becoming generally strong improves all other physical attributes.

SSNLP is simple in that it only manipulates intensity. In the main phase of the program the frequency remains the same (3x/week), volume remains the same (3x5 for most lifts), and only the weight on the bar increases each session (intensity). This works perfectly because our primary goal is to get stronger (increase force production) and this is quantified by weight on the bar.

When it stops working (and it will, eventually), we must begin to manipulate training variables in minimum-effective-dose steps to incrementally increase stress. It becomes necessary to transition to more complex, more individualized, and more specific programming, one small step at a time. The primary variables used to do this are intensity and volume.

Primary Variables: Intensity and Volume

Intensity simply means "how heavy," or literally “how much weight is on the bar” as a percentage of the lifter's 1RM (Example: “80% for 3x5”). While, for an individual lifter, 80% intensity is always greater than 60% intensity, there are several problems with considering intensity only as a percentage of 1RM. First, if percentage-based programming is based on a previous performance at a meet or a test day, often from months in the past, and not the most recent training completed by the lifter, it is based on stale data that is quite likely irrelevant to today's workout.

Second, when intensity is considered only as a percentage of 1RM, it doesn’t take into account the wide variability of stress induced by that same intensity percentage on lifters of various levels of strength and training advancement. For example, a lifter who squats 600lbs can make progress squatting in the 425lb (70%) range for higher rep/volume sets across. But a lifter who squats 175lbs almost certainly cannot drive a strength adaptation squatting 120lbs for any sets across. Despite being lifts of the same “intensity,” they are obviously different stress events. It’s clear that load matters.

Load (the weight on the bar) matters because the goal is increased force production. However, considering intensity only in regard to load is an error as well, because a 275lb squat for one lifter might be an all-out bone-on-bone grinder, and for another lifter might be an easy warm-up weight. The SAID principle reminds us that the stress induced must be specific to the goal – we cannot increase force production with light weights. Therefore, our training weights must be “heavy” in both load and percentage of 1RM in order to get stronger.

Volume simply means "how many," or literally “how many reps did I do?” It is most often measured in terms of work sets x reps (ie. 3 sets of 5 reps = volume of 15). Volume can be increased via increased reps per set, via more sets, or via more frequency on a given lift (or overall).

Novices focus their volume on 3 sets of 5 reps for most of the exercises. This is because 5 reps have proven themselves to make the greatest improvement on both increased force production and hypertrophy. And “3 sets of 5 reps” is the right amount of work for novices to be able to recover from 48-72 hours later, while driving a maximum strength and hypertrophy adaptation. Sets of 1 and 2 wouldn’t drive a hypertrophy adaptation, and sets of 10+ isn’t heavy enough to drive a strength adaptation (and in weak people, most likely wouldn’t provide enough stress to drive much of a hypertrophy adaptation either).

Secondary Variables: Frequency and Exercise Selection

Increased frequency and exercise selection are often cited as two additional variables (along with intensity and volume) for increasing stress specific to the strength adaptation.

Frequency simply means "how often," or literally “how many times did I train (or perform a specific lift) in a given time period?” Because ultimately the purpose of increased frequency is to be able to perform more work in a given time period, it can be considered a function of volume.

Exercise selection, just like programming itself, slowly moves from simple to complex, from comprehensive to individual, and from general to specific. Since everyone needs to get stronger in the beginning, we use three primary criteria for choosing exercises:

  1. Use the most muscle mass
  2. Use the most weight (with perfect form)
  3. Use the greatest effective range-of-motion

Applying these 3 criteria distills down the infinite number of exercises to choose from, down to just 4 primary exercises: the squat, the deadlift, the press, and the bench press. Because these exercises give us the greatest return on our training investment, we stick with these four lifts, and only these four.

As the lifter progresses past the early novice phase of training, we may begin to make small changes to the exercise selection in order to continue driving progress. Advanced lifters may use a significantly larger pool of exercises, but the founding principles of choosing exercises that use the most muscle mass, the most weight, and the greatest ROM still apply. If a strength increase remains the primary goal as the lifter advances, then choosing exercise variants that are similar to the 4 primary exercises is essential. The further an exercise is from the main lift, the less it will carry over to the main lifts, and thus, the less it will contribute to strength improvement.

While the argument can be made for an advanced lifter to use a deficit deadlift or rack pull to drive up the deadlift (a deficit deadlift adding a slightly greater ROM, and a rack pull adding an increased amount of weight on the bar compared to a deadlift), exercises such as a dumbbell deadlift, single-leg Romanian deadlift, trap bar deadlift, or even a power clean will contribute very little to the improvement in force production in the deadlift itself.

There is some discussion in programming circles that a new exercise is more stressful due to its novelty and the fact that efficient motor patterns have not been established for it yet. This is true. But novelty isn’t specific to force production increases, and thus, just because a novel exercise is more stressful in the beginning doesn’t mean it will make you stronger, especially if it looks nothing like one of the 4 main lifts. Some will cite the repeated bout effect, which states that the more we do something the less of an impact it makes on us. This ignores the fact that increasing the intensity or volume of an exercise, even one that you’ve performed thousands of times, will negate the repeated bout effect, because if it's heavier than last time you're not repeating the bout.

Quantitative Vs Qualitative Metrics

Since neither stress nor recovery can be measured in units, we need metrics to measure progress to know if our MED variable change is working – that the stress and recovery is actually bringing about about a strength increase. There are 2 kinds of metrics: quantitative data and qualitative data.

Quantitative data is objective. It’s real. It’s true. Our primary metric of quantitative data is the personal record (the PR). We believe so much in its power that we focus the vast majority of our work around the PR and celebrate it no matter how big or how small.

Secondary forms of quantitative data would include total work volume (sets x reps), intensity (both % of 1RM and the actual load on the bar), and work tonnage (weight x sets x reps). These secondary forms of quantitative data can be important, but they will always pale in comparison to the PR. Calculated 1RM, which uses an equation based on reps at a given weight to give an approximation about the strength of the lifter, is not quantitative data since the 1RM was never performed. Remember that quantitative data is objective, and it allows us to organize and program for training based on our previous training.

Qualitative data is subjective, based on perception, and is most-often self-reported. Qualitative data gives us perceived information about qualities: How hard something was perceived to be, how many reps the lifter perceived they had left in the tank, how close the form was perceived to be compared to the model, perceived fatigue, etc. The problem with Rating of Perceived Exertion (RPE) is that it is, by its very nature, subjective (perceived), self-reported, and therefore qualitative. In practice, it is almost always inaccurate. This doesn’t mean it’s completely worthless, but it cannot be held in the same regard as quantitative data, just as self-reported data does not compare to observed data in any other application of the scientific method.

Qualitative data may be an appropriate communication tool as a descriptor, but it is inaccurate and ineffective when used as a prescriptor, to plan future programming or to draw conclusions about the effectiveness of that programming. In fact, the existence of RPE percentage tables shows that in practice, RPE is often just a proxy for percentage-based programming.

How We Prioritize Stress Variables

Since the primary goal during Starting Strength’s Novice Linear Progression is to get stronger via increased force production, we choose to increase intensity for as long as possible (while keeping volume static), since force production is increased every single session that we are able to add weight to the bar. Furthermore, as a lifter moves out of the novice phase and into intermediate and eventually advanced programming, intensity remains the most important variable since an increase in intensity always results in an increase in force production.

Once intensity cannot be increased at 3 sets of 5 every single session, we begin making small minimum effective dose of complexity changes to continue to drive progress. These MED changes will always result in an increase in stress or an increase in recovery in order to effectively manage the stress-recovery-adaptation cycle.

In part 2, we’ll go through specific examples of how this process is implemented.

Additional Resources

Intermediate and Advanced Training: A Few Ideas

Volume and the Masters Lifter

Practical Programming for Strength Training

The Barbell Prescription

  • Part 2
  • Discuss in Forums

    Starting Strength Weekly Report

    Highlights from the StartingStrength Community. Browse archives.