Generalizing the weights for normalized individual resistances

Abstract

In the following we will first illustrate by an example that weighting the normalized individual resistance levels r_i(s^k’,s^k)/r_i(s^k,s^k’) simply by P_i according to (I.2.9) may yield counterintuitive results. This will inspire us to define generalized weights for the individual resistance levels. Since not all other players necessarily suffer if a player deviates from a given solution candidate these generalized weights take into account whether a deviation of a given player is harmful for his opponents or not. Again we will analyse several examples to illustrate the selection properties of the generalized ESBORA-concept as well as the differences between the preliminary and the generalized concept of resistance avoidance.