Stablecoin represents a unique subset of cryptocurrencies designed to offer price stability, achieved either through backing by specific assets or by employing algorithms that adjust their supply in response to market demand. In its landscape, algorithmic stablecoin is one special type that is not backed by any asset, and it stands to revolutionize the way a sovereign fiat operates. As implemented, algorithmic stablecoins are poorly stabilized in most cases; their prices easily deviate from the target or even fall into a catastrophic collapse and are as a result often dismissed as a Ponzi scheme. However, what is the essence of Ponzi? In this paper, we try to clarify such a deceptive concept and reveal how algorithmic stablecoin works from a higher level. We find that Ponzi is basically a financial protocol that pays existing investors with funds collected from new ones. Running a Ponzi, however, does not necessarily imply that any participant is in any sense losing out, as long as the game can be perpetually rolled over. Economists call such realization as a rational Ponzi game. We accordingly propose a rational model in the context of algorithmic stablecoin and draw its holding conditions. We apply the model and use historical data to examine if the major types of algorithmic stablecoins meet the criteria for being a rational Ponzi game.