Techniques are provided for rapidly optimizing control parameters of pacemakers or implantable cardioverter defibrillators. Briefly, the heart is paced using different sets of control parameters during a sequence of consecutive short evaluation periods of equal duration, which each last only about 5-12 seconds. Transient cardiac performance is monitored during each of the short evaluation phases and optimal parameter settings are then estimated based on changes in the transient cardiac performance from one parameter setting to another. By using a series of consecutive short evaluation periods of equal duration, rather than switching between short test periods and longer baseline periods, the overall duration of the test can be reduced as compared to predecessor techniques that require long intervening baseline periods.