The paper studies a property of mathematical objects called tournaments. A tournament is a way of taking a complete network of points (where every pair of points is connected) and assigning a direction to each connection, like recording the winner and loser of every game in a round-robin sports competition. The "eccentricity" of a point in such a network measures how far that point is, at worst, from any other point when you follow the directed connections. The "average eccentricity" is simply the average of these worst-case distances across all points, giving a sense of how spread out or compact the network is overall.
The paper focuses on a special class of tournaments called k-connected tournaments, where k measures how robustly connected the network is. A higher k means you would need to remove more points before the network could fall apart into disconnected pieces. The authors prove precise bounds on how large or small the average eccentricity can be in these networks. Specifically, they show it is always at least 2 and always less than a quantity that shrinks as k grows relative to the size of the network, meaning more robustly connected tournaments tend to have smaller average eccentricity, reflecting a more tightly knit structure.
The authors also pay special attention to the simplest case where k equals 1, meaning the network is just barely connected enough that you can travel between any two points following the directed edges. They derive a clean upper bound on average eccentricity for this case. They verify that both the lower and upper bounds they find are essentially the tightest possible, meaning you can construct real examples that come arbitrarily close to these limits, so no significantly better general bounds could exist.