It seems to me that we mightdo well to consider somethingless than exact optimality;I thinkthe best may be the enemyof the good. I will conclude witha suggestionwhichI hope is constructive.Consider a familyof sequential decision proceduresdepending on a randomized allocation index. The ...