The Attack and Defense of Weakest-Link Networks

*Brian Roberson, Purdue University, Department of Economics

Abstract
This paper experimentally examines behavior in a two-player game of attack and defense of a weakest-link network of targets, in which the attacker’s objective is to successfully attack at least one target and the defender’s objective is diametrically opposed. Consistent with the theoretical predictions, we find that, as the attacker’s valuation of a successful attack increases, the attacker’s expected resource expenditures increase and the defender’s expenditures decrease. However, both the attacker’s and the defender’s expenditures exceed their respective theoretical predictions. One of the most interesting findings is that in all treatments the attacker always uses a stochastic “guerilla warfare” strategy that consists of randomly attacking a strict subset of targets, while the defender uses a “complete coverage” strategy that allocates a strictly positive level of resources to each target.

Not available

Back to Abstract