The Free Energy Principle and Styles of Muay Thai Fighting: How Muay Khao Shrinks Surprisal and Controls the Fight Space

[Kevin von Duuglas-Ittu]

What follows is intersectional thought between an increasingly robust philosophical and scientific theory of living systems, and the art of fighting in combat sports. If you want to stay on top of the Free Energy Principle theory Maxwell Ramstead on Twitter is a great place to start. For me, personally, in the Free Energy Principle I find a lot of correspondence between itself and the philosophy of Spinoza, and it has always been my instinct that Spinoza has important things to say about some of our most concrete interactions as people. In some sense fighting is one of the most attenuated, physical yet social things that humans can do. This is about looking into the heart of what makes living systems tick, the possible dynamics which give all living things their values and their direction. And it is about thinking on the deeper metaphysical and scientific foundations of fight knowledge, the practical application of techniques and styles that have developed over the last century in Thailand. In bringing the Free Energy Principle to Muay Thai it is much more than a somewhat trite adage of: try to surprise your opponent! It is much more about thinking how surprise governs our choices of space, and the investments we make in maintaining those choices. And it is about recharacterizing the fundamental ways in which fight strategies and their tactics, create success. If organisms are at base in a struggle to control the entropy of information, increasing that entropy, taxing the resources of a fighters, is essential to long term fight success. I don't anticipate that this is for everyone or even most; this is just my passion where Philosophy and Fighting co-exist. As usual to my writing, this is not a finished article, but rather the sharing, mid-stream, the development of thought.

The Free Energy Principle Theory

Editing in the very best introduction to these ideas "An Introduction to Markov Blankets and Information Theory". Read this if you want to full, yet basic perspective.

and from "Brain Entropy During Aging Through a Free Energy Principle Approach" (2021, below:

Quote

 

Free energy principle in the neurocognitive system is linked to BEN in an information theoretical sense, where the brain tries to resist disorder, making assumptions and interpretations on reality, through active inference (Friston et al., 2012a). This makes the human neurocognitive model a hypothesis testing system (Gregory, 1980), in which it tries to guess the best statistical approximation (generative models) of the causes from the sensorial apparatus, through the Bayesian approach, updating the system (Friston, 2010). van Helmholtz (1962) claimed that perception of the world is not direct, but instead depends on unconscious inferences, or expectations that model perceptions, even before we can consciously perceive any specific object. He called the “likelihood principle” a predictive coding approach, also eloquently described by Chris Frith (2007) as “a fantasy that coincides with reality.”

Free energy principle comes from the Helmholtz view and the computational work of Dayan et al. (1995), where the brain tries to make sense of the world, anticipating reality through predictions, active inferences, representations of the environment based on personal experience and expectations, learning from experience, with the goal of limiting the entropy, the uncertainty. Any self-organizing system, including the human neurocognitive model, resists the distributed effects of a natural increase in entropy for its existence, development, and evolution, by trying to minimize free energy (Friston, 2010; Cieri and Esposito, 2019). Our system has evolved with the imperative to decrease the uncertainty within the system and between the system and the outside, with instructions for neuronal messages to interpret the reality and update constantly its own model. The brain has a model of the world and tries to update it using new information from sensory inputs (van Helmholtz, 1962; Gregory, 1980; Ballard et al., 1983; Friston et al., 2006).

 

This rumination will flow from the Free Energy Principle (FEP) which seeks to describe essential characteristics of living systems (being) and perhaps other phenomena as well. If you'd like a great 15 minute introduction to the principle you can check out this video explanation by Karl Friston. The Free Energy Principle generally argues that life forms, human beings included, benefit from reducing what in theory is called "surprisal" ("One can understand surprisal as a measure of how unlikely an observation would be by associating a system's sensory state with an observation or sensory sample"). Surprisal is here maybe best taken as just the surprise of events in the environment which are unpredicted...or unpredictable. This framework of reducing suprisal actually has very interesting broad brush insight into stylistic and skill-set developments in Muay Thai (and all other kinds of fighting arts).

If we take as a principle in fighting the desire to reduce surprisal, there are two basic ways to do so that come at first blush. The first is the reduce the size of the probability space, which is to say limit the number of things that can possibly happen. By shrinking the probability space the number of possible unexpected events becomes reduced. At a very basic level, this is why octopi retreat into coral enclaves to sleep. The probability space becomes reduced (I'm using this term perhaps untechnically, it perhaps could be called the "event space"). This also is why people live lives in more rigidly defined circumstances, whether that involves habits of behavior or the habits of mind which condition them. Reducing the probability space can control surprisal.

Reducing the probability space though has its drawbacks. You may not be able to control that space. This leads to the much more fruitful - though expensive - method of controlling surprisal, which is complexifying the prediction mechanism (the brain, nervous system, etc) so that it can predict events in larger probability spaces. More varied things can happen, and the mind (& body) is able through its development to read those patterns and become less able to be surprised. The richness of embodied knowledge over a probability space involves a mastery over it. And this process of complexification actually feeds on surprisal, because surprise causes it to grow in terms of complexification. The more unaccountable things it can account for, the greater its knowledge over a space.

The Free Energy Principle, Surprisal and Fighting

This leads to the tale of Sylvie's early year experiences of fighting in Thailand. She fought at an intense, never before documented, historic rate having over 33 fights a year for several years, but she was facing a serious disadvantage in these years. She was fighting more experienced opponents, with better eyes, who were much more familiar with the scoring aesthetics of Thailand. These Thai opponents had a very significant advantage in terms of surprisal in the fight space. Sylvie's stylistic solution to this was her discovery of the Muay Khao fight style tradition (a close-pressed, stalking fighting style that capitalized on clinch and knee fighting). She was aided in this in the discovery that there was in Thailand's history a traditional opposition between Muay Femeu (artistic, technical, often counter fighting) and Muay Khao fighting. This is where things become interesting.

Muay Khao and Probability

Faced with a disadvantage of fighting in a larger probability fight space, and not having the eyes and experience to read those probabilities visually, Muay Khao fighting tactics seek to shrink the probability space. In fighting close pressed, closing down angles, limiting distances the number of possibles to account for became more limited. It leveled the playing field more. But it is more than this. It also changed the dynamics of fighting and perception skills at play. In a more "padwork distance" fight space (fighters standing at basically the distance where training pad work is done) visual acuity reigns, but in more close range, clinch-oriented fighting tactile sensing can become paramount. By developing other sensory pathways (other than just eyes), and specific techniques to use them (locks, trips, turns, leverages) this new, (seemingly) smaller probability space then became expanded. Expanded in a different sense: what is possible in terms of historical skill development. The entire art of Muay Khao fighting actually is about expanding the possibilities of what can happen, at close range, exposing your opponent to increased surprisal. What began as a strategic shrinking of the probability space lead to the complexification of the art within that space, through both an inordinate number of fights (now well over 260), but also through the study of the art of Muay Khao fighting itself, and the richness of its specialized tactical knowledge. 

An example of a recent development of this tactical knowledge, anecdotally, is Sylvie's discovery and development of clinch throws from an outside position. Usually the preferred position in clinch is with an inside frame (Sylvie talks about the frame here). This where you are able to control the body of your opponent most directly, framing them. But through lots of sparring with larger partners Sylvie has spent a lot of time over the last few years in an outside (inferior) position. From this position, though technically inferior in important ways, one gains a dynamic advantage of some leveraged turns, trips and throws, introducing more centrifugal rotations to otherwise very erect, linear opponents (standing erect or stiffening can be a principle of strong clinch counter-control). Along with inside position locking and overhook tactics which have proven successful in countless fights, she now has an increased vocabulary of rotational force...which increases the surprisal for opponents. You can see a cut up of this dynamic here:

 

It is a complexifcation of a reduced probability space, introducing an additional axis of dynamic movement, and its attendant vocabulary.

There is of course quite a bit of complexity that is involved in the very act of shrinking probabilities spaces into fitness landscapes you've focused on understanding. These can involve hidden skills that require the control over space. Fighting is often a battle over space, and seeking to impose your probability space upon your opponent's preferred probability space. Getting from one space to another makes up a great deal of the art of fighting itself. But, one's growth as a fighter involves constant exposure to surprisal itself (hence the importance of live sparring & freestyle clinch training), so that your embodied prediction mechanisms will be stimulated to increase depth of knowledge over the fight environment, regardless of probability space in its variety. This ultimately involves increased visual acuity, and a soaked-in attendance to pattern recognition, what we've called "growing eyes". We've discussed it before, but the time we've spent with the legend Karuhat has uncovered the various ways in which he was able to read very large probabilities spaces through the detection of weight transfer, and body position alone. It has seemed to us that Karuhat, as he reads an open spaced opponent is often through a felt calculation of changes of probability that arise from how the opponent is standing, where their weight is distributed, and the shape of their bodily disposition (limb positions, head position, etc). He is able to reduce the complexity of what is probable, shrink "surprise", through a genius-level understanding of the fighting art. It comes off as mind-reading. Exposure to new and different probability spaces, controlled exposure to surprisal is that grows the complexity and knowledge of the fighter.

If interested we sketch out a rudiment of Karuhat's spatial reading in this article: The Secrets of Karuhat’s Style – Four Internal Games From Southpaw as discovered in this study.

 

 

[Kevin von Duuglas-Ittu]

Cliffnotes summation:

Muay Khao fighting style:

1. Shrinks the probability space (closes physical distance, limits weapons).

2. Changes the dominant affect register (from visual to tactile/kinesthetic).

3. Through its ars technica increases the probability space (learned/discovered dynamics of movement & control).

[Kevin von Duuglas-Ittu]

Ah yes. I forgot to get a bit into another way in which fighters seek to overcome surprisal, which is neither the shrinking of the probability space, nor the complexified growth of predictive awareness. This is the use of memorized combination patterns of attack. This is often a bite-down approach to the probability space wherein one inures oneself to surprise itself. It does not matter if something unexpected happens. One just tunnel-visions and locks into a very rehearsed and trusted somatic pattern of attack, designed to (hopefully) produce favorable outcomes regardless of opponent/environment. This is locking oneself into a closed form.

An interesting historical occasion of this in the history of Muay Thai was when Namkabuan at the end of his career fought Ramon Dekkers who founded much of his fighting style on memorized combination fighting, Namkabuan's commentary on the fight here.

There are also more restricted versions of this kind of bite-down approach to surprisal, for instance the insensate use of some kinds of guards, in which surprisal is just weathered through.

[Kevin von Duuglas-Ittu]

Quote

We understand this to imply that an agent is able to minimize free energy, and therefore surprise, by actively sampling and changing the hidden causes of its environment. This means that biological systems have expectations and make inferences about the causal regularities and make-up of the environment in which they are situated [30]. In short, given enough time, agents will come to be the authors of the external states (i.e. environments) that reciprocate with predictable, uncertainty resolving sensory feedback of exactly the right sort to sustain cycles of self-evidencing.

This is a fighter's capacity to impose regularity upon another fighter, making another fighter predictable. source: on Markov Blankets

[Kevin von Duuglas-Ittu]

There is another suggestive productive branch to the Free Energy Principle application to Fight Theory in John Boyd's military thinking about combat and his model of the OODA loop. If combat is a form of information warfare, seeking to stress the opponent with information overload, John Boyd's fighter pilot derived theory may provide rich resource for thinking about where information entropy can occur. wikipedia on the OODA Loop.