Suppose that we want to assess the contribution to power of formal position alone. One way to do so would be an empirical study in which we would consider simultaneously all of the various contribution factors, apply some variant of a multiple regression technique, and determine the appropriate coefficients for the position variables. A second way would be an experimental study in which nonposition factors are systematically randomized. A third way would be the one taken by Shapley and Shubik. We can imagine a game involving position variables only (e.g, the formal legislative scheme), and we can assume rationality on the part of the participants and ask for the value of each position under that assumption. Since this value is a direct measure of the system by virtue of this position in the game alone, it is a reasonable measure of the power of that position. Alternatively, we can view the resources themselves as power