To the question of what founds a formation like a sport, or what its conditions of existence are, an obvious answer would be "the rules of the game." But in the history of sport, as with virtually every collective formation, the codification of rules follows the emergence of an unformalized proto-sport exhibiting a wide range of variation. The formal rules of the game capture and contain the variation. They frame the game, retrospectively, describing its form as a set of constant relations between standardized terms. A codification is a framing derivative that arrogates to itself the role of foundation. It might be argued that all foundations are of this nature: ex post facto regulatory framings rather than effective foundings. Once they apply themselves, the rules do effectively frame and regulate the play, taking precedence. Their precedence is retrospective, or fictional, but effective. It has all the reality of a formation of power, of which usurpation might be argued to be the model — usurpation of variation (Parables for the Virtual, p.71).
The history of basketball would suggest otherwise. James Naismith invented the game specifically to find an indoor winter activity as an outlet for aggressive masculine tendencies at the YMCA in Springfield, Massachusetts. In other words, the rules did not form later and apply themselves retrospectively, but were rather present from the outset (and the soccer ball became the first basketball!). Indeed, play in general would not be possible were it not for the loosely formed constraints that condition possibility in the space of ludic movement. Otherwise we would find ourselves squarely in the zone of the virtual.
I also think outside of sport to computer videogames. Julian Kücklich discusses the ruledness of a game environment and how in greater degrees it may also constrain the conditions of possibility. In effect, what he is describing is how linguistic codes can striate a space of play. But the spaces of computer videogames — which are rapidly becoming sites of collective formation not unlike sport — are always already ruled spaces by dint of their linguistic form, or put differently, their programmed nature. There is no proto-game from which Grand Theft Auto emerges to contain variation.
(It should be pointed out that ruledness is not quite the same as the striated-smooth dichotomy proposed by Deleuze and Guattari: a space may be striated in material form without firm rules necessarily being attached to the effect of those gridlines. That said, in the foreword to Baudrillard's In the Shadow of Silent Majorities Lotringer, Kraus and El Kholti state that "Félix Guattari may have answered that it is no longer necessary to maintain a distinction between material and semiotic deterritorializations and that there is no more absolute primacy of one system over another." We will not understand them as synonymous, then, but rather as forms that modulate the conditions of possibility in similar fashion.)
But we are in the realm of sport for this discussion. And so we are more fully in the realm of the moving body, in all its sporting forms. Does this imply that Massumi's broader concerns are invalid? No, I don't think so: the point was still to capture and contain variation, but variation of a different order, the roughhousing of young male physical education instructors, rather than an earlier proto-version of basketball. Through the formalization of rules, variation is captured and contained across space and time: it implies that the sport will be played again on another occasion, in the same place or at a different location, and that there will be some consistency between the two events.
If the rules are ex post facto captures that take precedence, what do they take it from?: from the process from which the game actually emerged, and continues to evolve, to the extent that circumstances arise that force modifications of the rules. The foundational rules follow and apply themselves to forces of variation that are endemic to the game and constitute the real conditions of the game's emergence. The rules formally determine the game but do not condition it. (They are its formal cause, not its efficient cause) (Parables for the Virtual, p.72).
Do not forget that this "proto-sport" continues to exist (in basketball we have called it "pickup") alongside the more formal regulated versions of the "sport" (that is, league basketball). In other words, we do not want to get caught up in a linear "stages of progress" model in which pure play becomes proto-sport, which becomes sport, which is thereafter refined by variations in the style of play that are captured or modulated by new rules. The informal pickup versions continue to exist in parallel and in fact play significant roles in creating the variation that flourishes in (and challenges the formality of) "sport" proper.
Let us not become bound up in the search for origin and instead be more cognizant of the process. For in capturing variation — the motor of sporting emergence — we are capturing the possible of the body athletic. Herein lies the micropolitical moment of sport.