In 3.1, rejecting (R) fraudulent (F) goods resulting in 0 for the seller is a strong assumption. There are all kinds of possible negatives (typically risks of legal fees) for storage costs of fraudulent goods in the game of hot potato. It might be worth your time to look into the literature a bit more.

The paper was aimed at a broader audience, which is why it simplifies some concepts and makes certain assumptions. While there are potential drawbacks to consider, it’s reasonable to assume that a seller of a fraudulent item gains nothing if the item gets rejected. In that scenario, their goal of making money wouldn’t be achieved, and any additional losses wouldn’t significantly impact the protocol. In fact it would deter fraudsters. I'd love to hear more thoughts on this!