Even so often in my Facebook group, the topic of supercompensation of training comes up in one context or another. And I invariably dump on the idea, pointing out that it is neither an accurate or correct model of the training process. Now I want to explain in detail why I believe that. Along with supercompensation, I also want to look at the fitness-fatigue model of training and General Adaptation Syndrome.
A Note about This Article
As I mentioned in the video, a great deal of the ideas I will discuss come from a paper titled The Basics of Training for Muscle Size and Strength: A Brief Review on the Theory. by Buckner et. al. I’d point out two things about this paper.
First, one of the authors (Jeremy Loenneke) puts out work from a lab that seems to have a bit of a HIT bias. I’m not trying to dismiss the paper so much as put some perspective on the author’s direction. I’ve also read much of the literature they cite and agree with more of it than not.
I’d note that Loenneke has some strongly held ideas that I find…questionable. To whit he has argued extensively that the relationship between muscle size and strength is only correlational. That is, he believes that increasing muscle size doesn’t increase strength.
He seems to base this primarily on beginner studies where strength gains are mostly neural and don’t track with size gains. He also seems unaware of the relationship of muscular cross sectional area and force output which is basically linear. A bigger muscle is a stronger muscle. Or perhaps he is simply unaware that the increase in strength is driven by BOTH muscular and neurological adaptations. Regardless of his thought process, I think he’s wrong as hell.
In that vein he’s stated that powerlifters who have a hypertrophy phase are wasting their time. And, well, I find that laughable.
Taken to its “logical” end-extreme his belief would mean that a lightweight and a superheavy weight can lift the same amount. Muscle size doesn’t count and increasing muscle size doesn’t increase strength (potential). Yeah, nah, mate.
Anyhow, that’s irrelevant to this actual article, I just wanted to get it off my chest. As I said, the paper stands on its own in terms of the ideas it presents and the research it cites in support of it. So let’s talk about supercompensation.
What is Supercompensation of Training?
In the context of training, the supercompensation model represents a situation where after training there is a drop in fitness followed by recovery and an “overshoot” where fitness increases above the original baseline. Over time it falls back to normal. I’ve shown this in the graphic below.
So you can see how the baseline of fitness first drops after training. Then there is the recovery period until fitness increases above normal. If training doesn’t occur again, there is an eventual involution back to baseline. Hence the idea is that training first causes a drop in something before an overcompensation above normal occurs.
So think about this in a training context. You’re trying to improve strength so you do a hard workout. When I measure you the next day, you’ll be weaker. Then in a few days you’d recover and hopefully come back stronger. Rinse and repeat until you squat 1000 lbs. Sure.
The above model is part of the ever lingering theory of muscle growth where training breaks down muscle which then rebuilds itself to a slightly higher level. Rinse and repeat until you’re bigger than Ronnie Coleman. The training process, according to the supercompensation model, would look like this.
And, well…it’s wrong. Ok, it’s sort of wrong. By that I mean in a sense it describes what happens with training to a limited degree. But what it represents, in a conceptual sense, is wrong. And that’s because it was based on an irrelevant observation of a singular biological system.
Supercompensation of Glycogen
So far as I can tell, all of this can be traced back to a Russian researcher named Iavorek (that is how I’ve usually seen it spelled) back in the mid 20th century. He was studying liver glycogen in rats. And found that if you depleted liver glycogen, you could supercompensate to a higher level than they started. Of course we know this happens with muscle glycogen as well.
Carb loading has been in use by endurance athletes to supercompensate glycogen for performance since the 1960’s or so. My own Ultimate Diet 2.0 uses a carb-depletion/carb-load approach to generate its magic. The basic idea is that you first deplete muscle glycogen with heavy training and low-carbs. Then you reduce training and eat all the carbs. Boom, muscle glycogen ends up 125% or whatever above where it was normally.
Basically the model relied on first depleting something before the body would supercompensate to a higher level of that something. And that model got applied uncritically to training. Hence to build more muscle we first break it down so that it will overshoot to a higher level. Which leads to the idea of destroying the muscle once per week and then waiting a week to train again among other idiotic ideas.
Two Problems with Supercompensation
The problem being that it only applies to glycogen and not to anything else in the body. Iavorek happened to be studying the one biological system where supercompensation occurs. But it doesn’t happen anywhere else.
When you train for muscle growth you don’t have physically less muscle after training before growing more to a higher level. In its strictest interpretation, supercompensation simply doesn’t occur except for glycogen levels.
Related: Does Muscle Damage Cause Soreness?
Here there is another problem. One of the ways that supercompensation is applied to training is that you train, fatigue, supercompensate, train again, fatigue, supercompensate, etc. And that’s how you increase fitness, with this stair stepping up as I showed in the image above.
But that doesn’t even occur for glycogen. Because once you get that 25% increase above normal, that’s all you get. If you deplete and reload again, you hit 125%. And 125%. And 125%. It doesn’t keep going up and up to 135% and 145% and 155% forever. You can do it once and that’s the maximum. And fitness doesn’t work that way since, at least for some time you continue making upward improvements.
It was only by looking at the wrong system over a single implementation
that supercompensation seemed to be valid.
But it wasn’t, at least not in its strictest sense of depleting some biological substrate that rebuilds to a higher level, improving fitness. It simply doesn’t work that way.
But in a conceptual sense, doesn’t it still sort of describe training? You train, you fatigue, at some point later you’re stronger. It’s sort of a supercompensation curve, right? Sort of.
Except that it’s still not really correct. And the implications it has for how training should be set up are fundamentally wrong. I’ll come back to this, first let me look at a better model.
The Fitness-Fatigue Model of Training
Also called the Dual-Factor Theory of Training, the Fitness-Fatigue (hereafter FF) model of training is considered a much better model of the training process. If nothing else it represents a more nuanced look at what’s actually going on.
The idea of the FF model is that training generates two changes: an increase in fitness and an increase in fatigue. And that it is the difference between those two that determine the performance or preparedness of the athlete. So say you train and generate a 5% increase in fitness and a 10% increase in fatigue. Your performance will be 5% less. Right? Plus 5% fitness minus 10% fatigue is -5%.
How does performance improve? Well the model is predicated on the idea that fatigue dissipates faster than fitness. It’s usually assumed that fitness takes three times as long as fitness to dissipate. And it has to work this way, right? If both fitness and fatigue went away at the same rate, you’d spin your wheels. And, so far as I can tell, these assumptions have turned out to be more or less correct. Or at least close enough for government work.
So you train, generating fitness and fatigue, presumably decreasing performance. Then fatigue dissipates and performance improves, hopefully to a higher level. I’ve shown this below.
This model essentially explains how tapering works. For some period you’ve trained your butt off and generated some amount of fitness and some amount of fatigue. And as you cut volume and maintain/increase intensity, your fitness is maintained or improves as fatigue dissipates. And performance improves. Other aspects of training such as maintenance loads or even heavy/light/medium training and others fall into this model as well.
Supercompensation vs. Fitness-Fatigue
Now, this looks a little bit like supercompensation, doesn’t it? You train, performance goes down, then it comes back up, hopefully to a higher level. And superficially this is true that they are similar as I’ve shown in the graphic below.
But they are fundamentally saying very different things.
Supercompensation is based, fundamentally, on the idea of a depletion of something (muscle, ATP, mitochondria, capillaries?) that then supercompensates to a higher level. And it doesn’t work that way at all.
In contrast, the FF model is more nuanced, including some aspect of fitness and some aspect of fatigue that represent different things and change at different rates.
I’d note that, for the most part, the FF model doesn’t get into any real biological processes. At least not that I’ve seen. That is, fitness doesn’t represent one biological process and fatigue another. It’s just more a way of modeling the training process mathematically.
Post Activation Potential
If the above isn’t clear, an easy example might be Post Activation Potential (PAP). The idea of PAP is that you can use some type of heavy loading to potential performance in a later activity. So you do a 3RM squat or heavy jumps or something and a few minutes later run or do a bike sprint. Performance often improves under those conditions compared to doing it without the potentiation.
But this depends on a few factors. One of those is training status where more highly trained athletes tend to see improvements while less well trained see decreases. There is also a time factor with the duration between the potentiation activity and the performance being critical. If it’s too short, performance is decreased. Too long and there’s no improvement. Just right and performance goes up.
And both of those can be explained with a Fitness Fatigue conceptualization.
In the case of training status, the idea seems to be that advanced athletes get a bigger boost in fitness with less increase in fatigue. Presumably this is due to being more well trained. In contrast, less well trained athletes get more fatigue and less fitness. So the trained athlete gets a performance boost and the less well trained gets a drop.
The rest duration is even more illustrative. So the potentiation activity increases fitness and fatigue to some degree. Even the most highly trained athletes wouldn’t get a boost one minuter after a 3RM squat. But at 3 minutes, or 5 or 7 minutes, fitness is maintained while fatigue dissipates (remember, three times faster). And performance improves. But if you wait too long, any fitness boost also dissipates and you get no improvement.
So even if the supercompensation curve is equally descriptive in this case, it’s not representing what’s actually going on. And even if it did, the implications of supercompensation versus the FF model on real-world training are completely different.
Training Implications of Supercompensation vs. FF
Perhaps the biggest reason to do away with the supercompensation model in favor of the more modern and correct FF theory is due to the implications they each have for real world training. Because they end up leading to completely different conceptualizations of how training should be performed.
Supercompensation is a full-recovery model. The idea that comes out of it is that the optimum time to train is when performance has supercompensated to the higher level. Train too soon and you keep going down and down into fatigue. Train too late and you get involution. The idea is that you wait until fitness (i.e. size, strength, speed) has improved and then train again.
And anybody reading this outside of maybe hardheaded bodybuilders know that it doesn’t work that way. Logically this implies that a sprinter would run a sprint workout and not train again until they were a little bit faster. Or that a strength athlete would train and not train again until they were a bit stronger. Or a bodybuilder wouldn’t train a bodypart again until it was a bit bigger.
And everybody reading this outside of total beginners know that that’s nonsensical.
Sprinters may work for 12 weeks to improve their performance by hundreths of a second. What should they do? Sprint once and wait 12 weeks to supercompensate? A powerlifter or Olympic lifter may train 12-16 weeks to improve a lift by 5 lbs (2.5kg). Growth in bodybuilders is slow as paste. You’d never train if you wait for supercompensation to occur.
Related: How Fast Can I Gain Muscle?
This is the fallacy of HIT, the idea that you should make strength gains at every workout, after supercompensation has occurred, or you’re training too soon. And outside of beginners it doesn’t work. Because if it did you could just train after supercompensation occurred and make linear progress until you lifted the world.
And even in beginners, the rapid improvements are still better modeled by FF. Each workout improves fitness a bit and fatigue a little. But since the training load is so low, fatigue goes away quickly and they can add weight to the bar in 2 days. Since those early adaptations are neural anyhow, a lot of this sort of stops mattering.
But at higher levels, the implications of supercompensation for training are nonsensical while the FF leads to a far different idea. The implication of FF for training is that you are overlapping training sessions on top of one another. They may vary in intensity, volume, etc. but each impacts on fitness and fatigue to one degree or another.
In no case are you waiting for or even expecting full recovery or a measured improvement in fitness. It may happen, athletes who train like this will get adaptations and improvements over time without “waiting” for supercompensation to occur.
It’s just overlapping traces of fitness and fatigue that are changing at different time courses until performance improves. Sometimes it’s during training, sometimes the major improvement is when you finally taper. But the idea is the same either way.
Real World Training Implications of Fitness-Fatigue
So swimmers will bust ass 12 time per week in the pool. Sometimes they are slower, sometimes they are faster. Adaptations occur eventually, then they taper and hopefully set PR’s. Olympic lifters do it too, most of their workouts are not expected to be at a new level of performance.
Hopefully adaptations occur during training that allow them to lift more heavily over time. Bulgarian OL’ing methods are based around this. You go to max all the time and eventually you hit a new PR. Or, in most cases you break completely.
But in most cases like this, you’re not seeing any real long-term decrements in performance. You don’t get slower and slower or weaker and weaker or smaller and smaller until you supercompensate to a higher level with rest.
I suppose the closest thing to this would be double-shock microcycles where you bust an athlete’s ass for 2 straight weeks until performance drops and then let them recover. Basically deliberate overreaching. Those WSBB Circa-Max cycles are similar. The athletes get beaten on for 3 straight weeks until they start to break and then get 3 weeks of recovery to hopefully hit a new level.
Related: What is Overtraining?
But even that is better modeled by Fitness and Fatigue than supercompensation. The heavy training generates a lot of potential fitness but a LOT of fatigue. Performance acutely goes down. Recovery occurs, fatigue dissipates and hopefully fitness improves on competition day.
The General Adaptation Syndrome (GAS) Part 1
Which brings me to the next major topic of this article, the General Adaptation Syndrome or GAS. This is another idea, developed in the early 20th century that was (mis) applied to training for many many decades. It’s used to support a lot of the training ideas I already mentioned along with global concepts about periodization. I still see articles using it to describe things and, well, it’s simply not really correct.
The paper I cited above was looking at, among other topics, that of periodization. Both in the sense of whether or not it has global validity along with the idea of if the GAS can be used to describe or defend it. Before getting into that, let’s look at what the GAS is.
I imagine many readers know the name Hans Selye. He was a researcher (or lab intern, maybe?) back in the early 20th century who, as Robert Sapolsky once put it, wasn’t very good at his job. He was studying rats and apparently couldn’t keep hold of them. He’d chase them around the lab, drop them, etc. And at some point he did some measurements and saw that they were all stressed to hell. High cortisol, all that jazz.
And this would lead him to do research where he subjects the animals to various stressors. Heat, cold, toxic chemicals, etc. As I stated in the video, if Justin Beiber had been around, he could put Baby, Baby, Baby on loop.
And what he found was this: despite the varied stressors, all the rats showed a generalized stress response. Their stress hormones were up and, in extreme situations, their adrenal glands had shrank. And he concluded that varied stressors could generate a generalized stress response.
And this was totally foreign to the savants of the day. The disease model of medicine was based around the idea that a specific disease had a specific set of symptoms and a specific treatment. How could so many things cause a generalized biological response? It made no sense.
But Selye was essentially correct. I say essentially as there can be subtle differences in stress response to different stressors. But big picture there is a generalized response to stress. It’s why you have to consider all the different stressors on the body, exercise, diet, mental stress, etc. a concept called the allostatic load. Due to sex based differences in stress response, I dedicated an entire chapter in The Women’s Book Volume 1 to this topic.
Anyhow, back to Hans Selye.
The General Adaptation Syndrome (GAS) Part 2
Based on this stress response, Selye came up with what he called the General Adaptation Syndrome. This described the stages that a stressed animal went through in response to chronic stress. Those stages were alarm, resistance and exhaustion and I’ve shown this below.
The alarm stage shows the initial response to the stress, where a small drop in stress resistance occurs. Following this, you get resistance, where the body adapts/is able to handle the stressor. Exhaustion occurs when the stress overwhelms the organism and the entire system shuts down.
And there’s little to no doubt that this occurs in animals and humans where there is no doubt that chronic stress is extremely damaging to the body. In that vein, I’d point interested readers to the best book on stress ever written: Why Zebras Don’t get Ulcers. It’s an amazing book, written by Robert Sapolsky. He is the writer I want to be when I grow up but with slightly less crazy hair.
The question is whether or not it was valid to training. At first glance it made sense. Training is the initial stress, causing the alarm stage (and a little dip in something such as strength, power or speed). If the training load isn’t too excessive and the athlete gets sufficient recovery, they enter resistance and can handle the training load. If the training stress is too excessive, or too excessive for too long, they enter exhaustion becoming overtrained.
Related: How Do I Know If I’m Overtrained?
GAS vs. Supercompensation
Superficially, both the GAS and Supercompensation curves look very similar.
It’s not quite identical the way I’ve drawn it but if you kept training too early where you’re never adapting, you end up in the exhaustion phase. Based on that superficial similarity, coaches developed entire training approaches. So you started with high volume, entering the alarm/resistance phase and then reduced volume to avoid going into exhaustion. And hopefully the athlete supercompensated to a higher fitness level, supercompensation being a bad model for all of this to begin with.
The Problem with GAS and Training
So let me address why the GAS isn’t really an appropriate model of training or the training process by quoting from the paper I cited at the beginning of this.
Selye himself referred to the GAS as a “syndrome” that appears when “severe injury is inflicted upon the organism”(72). Indeed, it has been questioned whether “lethal” and “graded” doses of different agents can truly provide insight into the physiological adaptations that would occur with non-toxic levels/doses of resistance exercise (5). For example, in Selye’s original model the exhaustion phase was represented by death of the organism (67).
Read that last sentence again, last 4 words. DEATH OF THE ORGANISM. Being tired or 10 pounds down on the squat wasn’t exhaustion in this model. It was death. Now my powerlifter may tell you I’m trying to kill her but I find that dead athletes don’t taper well.
Their point is that the original description of the GAS does not describe anything about the response to training. It certainly doesn’t support periodization of training. A paper of interest is Kiely’s Periodization Theory: Confronting an Inconvenient Truth which basically dismisses the GAS as a good model for any of this. Others argue that the GAS still provides a useful framework although we still might question if exhaustion in athletes is equivalent to DEATH in animals. I daresay it is not.
Even ignoring that, the observations Selye made in terms of shrinkage of the adrenal gland doesn’t happen in humans so far as I know. Yes, some odd stuff can occur where the stress system shuts down. No, it’s not adrenal fatigue because that’s not a thing. Rather, it’s an adaptation to prevent damage to the system. But the end result is not the same.
Don’t get me wrong, the people who were applying GAS or Supercompensation back in the day were working with the best knowledge of the day. It’s very easy (but unfair) to be critical of something from decades back based on what we know now. At the same time, with the knowledge we have now, outdated models such as supercompensation and the GAS, along with training based on them, need to be thrown out.
We have better models now and we should base our training concepts on them.
Training is Adaptation
To bring this all together, let me present another two quotes from the paper.
Within the field of molecular exercise physiology, it has been suggested that the GAS and its resulting supercompensation hypothesis of adaptation “should no longer be used in an attempt to explain adaptation to exercise” (79). Authors outline several flaws with this hypothesis including:
1) Supercompensation happens with glycogen , but not with most other systems (i.e., mitochondria, capillaries or neurons);
2) the supercompensation hypothesis is a time course and not a mechanism
3) The supercompensation hypothesis implies that recovery is essential for adaptation, yet the heart adapts to exercise despite continuous contraction;
4) There is little actual evidence that the supercompensation is essential for adaptation (79).
Let me look at each. I addressed number one in detail above. Glycogen is the only system that depletes and then supercompensates. No other system works this way. Not skeletal muscle or mitochondria or capillaries or anything else. They never deplete and rebuild. They simply adapt over time.
Number 2 I only partially agree with. Originally, supercompensation was absolutely a mechanism. You depleted something and it then rebuilt to higher levels. Later it became more of a general description of the training time course. It’s still wrong.
Number 3 is a little questionable to me. Their point is that the heart never truly rests as it beats continuously but it still adapts. True. But cardiac muscle is a little bit different than skeletal muscle and it still rests relatively between exercise bouts.
And four is really the big one. Supercompensation in either its literal or descriptive sense simply doesn’t represent what actually happens in biological systems. With training you don’t see decreases or reductions in anything followed by an overshoot. It just doesn’t work that way. Rather:
Alternatively, it has been suggested that the signal transduction hypothesis of adaptation is more appropriate for explaining adaptations. This hypothesis suggests that sensor proteins respond to exercise signals, working through various pathways to regulate gene, protein synthesis/breakdown and other adaptive responses.
Unlike the general adaption and the supercompensation hypothesis of adaptation, this hypothesis explains how adaptations occur, and is supported by what happens in a human model following exercise (79).
This is how the body adapts to training and is actually what has been observed and measured in studies. Repeated training bouts induce gene expression, trigger various adaptational pathways (i.e. mTOR in response to high tension contractions) that lead over time to an increase in some biological system that improves fitness. Capillary number or mitochondrial density increases. They never decrease during training, they simply adapt upwards over time. And the GAS is not a good model of any of it.
Supercompensation is Dead from Exhaustion
The simple fact is that the supercompensation model is inaccurate. It’s certainly wrong in a literal biological sense and it’s barely correct in a descriptive sense. The Fitness-Fatigue/Dual Factor model is far more accurate descriptively but that’s still all it is: a descriptor of how fitness develops. It doesn’t describe biological processes so far as I can tell.
Similarly, the General Adaptation Syndrome is a poor descriptor of what’s going on unless you dilute it to such a degree you might as well throw it out. At no time does training result in death of the athlete (well it shouldn’t). Superficially it makes sense but biologically it does not.
And simply both supercompensation and GAS need to go away
as do training approaches based on them.
Long-term training is adaptation. Specific training triggers specific adaptations via gene expression, protein synthesis (myofibrillar, mitochondrial, sarcoplasmic) and that leads to improvements over time as adaptation occurs. Nothing is depleted, nothing supercompensates in the strict sense of the concept. And the GAS would seem to not describe any of this well at all since it has death as the extreme end point.
As a consequence, modern training systems are (or should) be based around that rather than earlier simpler models that lead us down a path to stupid training conceptions such as train and then wait to train again until you’ve improved fitness. Because that simply doesn’t work.