Dealing cards in poker games

This paper proposes a new protocol for shuffling and dealing cards, that is designed specifically for games of mental poker. Our protocol takes advantage of two features of poker games that are overlooked by generic card-shuffling protocols: 1) cards in poker games are dealt in rounds, with betting in-between, rather than all at once and 2) the total number of cards dealt in a game of poker is small (it depends on the number of players but is typically less than half the deck). With these observations in mind, we propose a protocol that spreads the computational cost of dealing cards more evenly across rounds. Compared to protocols that shuffle the whole deck upfront, our approach offers a dramatic decrease in latency and overall computational cost. Our protocol is fair, private and robust. It is ideally suited for resource-constrained devices such as PDAs.