This was originally an email reply to Sebastian's post on 'Starting Strength' by Mark Rippletoe. Sebastian invited me to post it up on the Community Site.

It was essentially a primer to the "auto-regulatory" style of weights training that is used by advanced lifters. 

Despite the previous statement, auto-regulation can be used by anyone with a moderate level of experience in the gym (>6 months) to great effect.

So a summary: 

• The body's capabilities vary from day to day. What constitutes a good training-stimulus on one day can wreck you on a day where you're feeling crappy.
• When it comes to weights training, we have both objective and subjective measures of "difficulty" that can be used to determine how much work to do.
• This leads to more effective workouts that deliver a better training stimulus in a shorter amount of time
• The best intra-set method of measuring "difficulty" is the Rating of Perceived Exertion (RPE). This is a subjective measure of how hard a set was. An RPE of 10 means that you couldn't do another repetition on a lift. An RPE of 9 means that you could have done one more rep (but didn't).
• Always workout to an RPE of 9 -- always leave a rep in the tank. This prevents burnout, and allows one to recover quickly to do another bout of "RPE 9 work". 
• Examples of auto-regulation based systems include Myo-Reps, and Reactive Training Systems. Such systems train you to "listen to your body" and give it only the required stimulus for the day.
• Auto-regulation likely applies to other endeavours as well, like building a software project, or running a marketing campaign, or just work in general. This is a point for discussion, and I'd like to hear what people think.

---

*Post Start*

Your recent post on strength training Rippletoe style caught my eye though. I've been in the gym for about 8 years now (ever since I was 13), and I guess I could say I've tried and failed with a whole bunch of workout routines to know what works.

Anyway let's keep with the theme of your site, and discuss strategy. Anyone starting out with strength training can and should stick to a cookie-cutter style program of mostly compound lifts. Strength gains at this stage are largely neurological, and hypertrophy will naturally follow.

Once the beginner phase is over though (6 months of proper training), then we start to transition into intermediate programs which are geared towards a particular objective. 

---

Most intermediate programs are still cookie-cutter style programs. You follow the program to a 'T', and if it didn't work, then tough luck. 

What the new-generation strength coaches are finding out is that every person's capabilities, needs, and objectives are different. Moreover, the body is a dynamic system, and experiences good days where every movement feels light, horrendous days where putting on your socks almost makes you cramp, and everything in between. 

Enter the world of auto-regulated strength training programs.

---

By "auto-regulated", what we mean is that we use a system that adapts to the body's capabilities at that particular moment, delivering just the right amount of stimulus for that person at that time.

One example of an auto-regulated training is Myo Reps, invented by Norwegian strength coach Borge Fagerli. 

It will require reading the linked article to fully understand the concept of Myo Reps, but essentially, you begin with an "activation set", which primes the muscle and exhausts certain muscle fiber types. 

You then follow up that activation set with multiple sets of only a few reps (myo reps), resting for only a short period between sets. Because the rest period is so short, the muscle doesn't fully recover in between those sets, and the repeated bouts of "myo reps" sets actually mimic the lifting of a weight that is heavier than what is used. 

The benefit is that you give the muscle a tremendous growth stimulus, in a very time-efficient manner (you only needed to do one slightly longer set as opposed to 3 or 5 sets).

More importantly, working out this way allows one to "auto-regulate" the amount of work done based on how you feel on that day.

Let's say you plan to do a set of 10 reps +3x on the chest press, meaning 10 reps for the activation set, and then 3-rep sets until exhaustion. You start the exercise, and by rep 7, you know that you're feeling down today, by rep 9, you know you're spent, so that's it for the activation set. You rest for 10 seconds, and then proceed with the first set of 3 ... it's tough, but not impossible. The next set of 3 ... really tough, no point in pushing further. 

You record that you did (9 + 3x2) reps on the chest press with X weight. 

The next time you come into the gym, everything feels much better, and using the same weight, 10 reps fly by, so you continue, and the activation set ends up at 13 reps. The Myo Rep sets feel light as well, and you eventually end up doing 5 added myo-rep sets of 3, finishing at (13 + 3x5) reps for the day. 

Here's the big man himself showing us a myo reps set -- https://www.youtube.com/watch?v=utmC52t6id4 

I'm sure you get where we're going right now. We shoot for a particular range, and then auto-regulate our proceeding workload based on what our body tells us.

To make this system work, one needs to learning how to auto-regulate volume and intensity on-the-fly. This is why it is not a technique recommended to beginners.

However, if you know what to look out for, attaining this ability is largely a matter of experience (and resisting the urge to train yourself into the ground). Most people will get really good within 3-6 months if they "listen to their body". 

---

"Level of difficulty" of a set actually has a proper definition -- the "Rating of Perceived Exertion" (RPE). 

This is a rating from the scale of 1 - 10, with 10 being absolute failure (you cannot lift another rep), and has been demonstrated in both research and practice to be very effective at regulating oneself during workouts. 

The most important RPEs are 8 and 9, two reps and one rep before failure respectively. It seems that going all the way to failure leads to an exponentially longer timer for recovery to take place, but stopping just before failure completely attenuates that effect. ie: going balls-to-walls may require 3 days of recovery, but stopping one set shy of that could have you in the gym the next day. 

On this topic, Mike Tuscherer is arguably one of the leaders. He runs Reactive Trainings Systems, and while his programs are specifically tailored toward powerlifting, any training can gain from the knowledge of that training sytem.

He himself is quite the powerlifting, and puts up some big numbers -- https://www.youtube.com/watch?v=nbQEk7GLAbk. That's a video of competition however. What we're interested in as strategist is how to get there, ie: the preparation and training. 

We see that in training, he's always very deliberate and controlled -- https://www.youtube.com/watch?v=2TiubINxYcw . That's an example of a set of 4 on the deadlift with 745 lbs at an RPE of 9 (meaning that he could have managed another rep without failure). An extended commentary on a training session can be found here -- https://www.youtube.com/watch?v=sY0gQJRanfg .

There is a method behind all of this, and I'm afraid this post is already too long to elaborate. He actually makes use of a whole bunch of metrics, some subjective, and some objectively quantifiable. Examples include bar speed, total volume tolerance, rest-times, etc ...

---

Essentially, the whole premise of autoregulation is that you establish objective ranges that strategically help you converge towards your particular goal, and then rely on both objective and subjective measures of work output to guide the micro-level decisions.

I'm sure you can already see the parallels to excellence in other endeavours. Working up to "RPE 9" on a programming project can be thought of as just the right amount of "stimulus" before breaking down into mess of unproductive behaviour (like trolling Hacker News). By cutting yourself short of that, you stem the bad impulse, and manage to recover quickly for another bout of "RPE 9" work. 

The question now is how to determine quickly and efficiently the metrics for auto-regulating such work. And can we possibly generalise this into a meta-system?