Consensus in Motion: A Case of Dynamic Rationality of Sequential Learning in Probability Aggregation