In most discussions on martial arts training, people’s approaches get categorized into one of three “models.” These models are usually labeled 1) traditional martial arts (TMA), 2) reality-based self-defense (RBSD), and 3) combat athletics, including mixed martial arts (MMA). The exact parameters of each model are pliable to a point, so that a given style or school might identify with more than one of these models and various points. That said, stylists often self-identify with one of these models more strongly than with the others. That self-identification generates many of the debates we see on martial arts training methodology. This tendency, in and of itself, is not the full issue. The other important component that feeds into these debates is the dissonance created by the simple fact that martial arts theory seeks to describe a reality in which most people do not actively engage. Even the RBSD model, which features “reality” right in the name, does not describe a reality that most people experience very often (if at all) in their lifetimes. This truth is what creates so much room for debate on training methodology. An advocate of combat athletics might point out that, if they can make a technique work in the ring against a well-trained and -conditioned opponent, they could also employ it against an attacker in a different scenario. The RBSD exponent might counter that the ring has rules that preclude such mainstays of self-defense as the eye gouge or strikes to the groin. The traditionalist might cite the fact that their style was developed for and employed on the battlefields of one ancient culture or another. I will not try to offer up every variation on and counter to these points in this article. Anyone who has spent enough time on MAP is already familiar with many of these stances. What I will say is that “reality” is not a particularly helpful benchmark in evaluating the success of one model or another. Reality is, at best, a moving target. Furthermore, reality is a target that most of us cannot afford to actually hit in training anyway. What I mean by that last statement is this: The act of personal combat is, by definition, dangerous; life-alteringly dangerous in some cases. For that reason, training has to make concessions. What sparks discord between the various models is twofold: 1) Where do they make their concessions? 2) Do they recognize that they are, indeed, making a concession? Take the groin strike as an example. The RBSD and TMA models might point to the groin strike as both 1) an essential element of actual personal combat and 2) a limitation on the combat athletics model. It would be possible to look and combat athletics and dismiss their relevance to reality based on the fact that they generally do not allow strikes to the groin. On the other hand, while the other models may train these strikes, exponents of them are not actually striking one another in the groin in realistic ways either. They might put on a heavily padded self-defense training suit or practice pulling strikes to the groin. They are not, however, replicating the actual experience of employing a groin strike in reality. So, suddenly, there is room to question how valid a training methodology that is. The other major concern with reality is reproducibility. Even assuming that someone survives an encounter in real life, it can be difficult to discern “best practices” from blind luck. Further, it is impossible to say that the next real encounter will happen as the last one did. There are so many variables that come up in reality. A successful knife defense one day does not mean the person has knife defense “in the bag.” Because reality is such a diverse and broad concept, and because our cumulative experience of reality in personal combat is fairly limited in comparison with our cumulative training experience, I do not find it especially useful to categorize styles or even specific practices into the models described above. Nor do I find it especially helpful to talk about how something would work “in reality.” Personally, I like the idea of “proof of concept.” As a Filipino martial arts (FMA) enthusiast, I spend a lot of time learning to fight with sticks. In reality, I will likely never engage in an actual street fight involving sticks. So I could reason that I am not learning FMA as self-defense. Instead, I am interested in the tradition, culture, and technical aspects of FMA. That, however, creates a new potential for dissonance. If it is true that I am not concerned with reality, then I might be tempted to embrace a lot of training methods, drills, and techniques based on the fact that they have always been done. Speaking personally, however, I would rather try to get proof of concept. I want to know whether a technique or drill does what it says it will do on the tin, irrespective of whether I will realistically ever be called upon to use that technique. If I learn a parry, I want to know that I can employ that parry within certain parameters. “Within certain parameters.” Here is the tricky part. This is where the exercise becomes like a scientific experiment. I want to test a theory (this technique works in this way) I need to identify the equipment and procedures that will allow me to test that theory sufficiently to meet my threshold of proof So, as an example, I want to test the palis-palis (sweeping) parry as a viable defense against forehand attacks to my lead knee. Obviously the knee is a delicate target, and I only have two of them. I am quite fond of both, so I would prefer not to lose either one. From the traditional approach, I might spend a great deal of time with careful technical rehearsal using live sticks but prescripted motions and roles (i.e., “you’re the attacker, I’m the defender, and you’re going to use the forehand to my lead knee on my cue, so I can practice”). I may gradually increase variables such as speed and timing. I may also suggest we vary the location of the attacks, so that I have to correctly anticipate when the knee strike is coming. At some point, however, we are going to come to a safety border. I will have introduced as many variables as I safely can and I will have learned as much as is feasible from that approach. Now, I might turn to the combat athletics model. In combat athletics, I may get the contact level I was missing in line practice, but I will need to make another concession elsewhere. I may have to have my partner target the thigh instead of the knee. It is a close enough target to still extrapolate the lessons learned from one experience to conclusions drawn about the other. I may instead make my concessions with equipment. I may elect to use padded sticks and/or knee pads. Now, the speed can be ratcheted up. The opponent can employ feints, flurrying combinations, and other variables to test my reactions. Real intent to hit the target can be substituted for the earlier intent to help me learn the pattern. Now I have another point of reference through which to triangulate the usefulness of this parry. To some, this advice will seem obvious. That said, I have seen countless exchanges where people have refused to embrace some training nuance that another model takes as obvious. Whether this reaction comes as a result of wanting to belong to some camp or another, I could not say. What I would say is that, in this day and age, we have the equipment and resources available to make testing a viable part of almost any training experience. (Yes, okay, archers. I do not have a simple answer for you guys.) In my view, the real fun of a martial art is the sense of physical problem solving. So I encourage practitioners from every training methodology to 1) identify a question they want answered, 2) decide what their threshold of proof is, 3) find the gear that would make it possible, and 4) play with it.