APPLICATION OF STATISTICAL-METHODS FOR DESIGN AND ANALYSIS OF INDIVIDUAL BRIDGE COMPETITIONS WITH A SMALL NUMBER OF PLAYERS

Bridge competitions where the same cards are played at two tables and the results compared are discussed. Statistical models are formulated and good designs considered. In some cases, balanced incomplete block component designs can be used. In others, designs with minimum average pairwise variance for estimated player skill are chosen. All cyclic d esigns are then searched and the best are found to come relatively clo se to the lower limits of average pairwise variances. Results from the competitions are analysed with the least-squares method. Standard bri dge references do not always give acceptable designs and methods of an alysis.