Abstract
Taking computer chess as an example an idea is presented how to use the early available principal variations for optimistic pondering under real match conditions, even before the opponent is on move. This way, several processors of a distributed environment can be used to search asynchronously ahead in parallel to achieve a greater thinking-time amount for the game-tree search of a single move. The actual principal variations which can be provided by all typical chess programs during their iterative deepening are the only information needed for this method. The implementation of the idea for a prototypical environment is described and the major results collected so far from this environment are presented.