Driven by an analogy between a golf tournament and a psychological test, classical methods for the assessment of test reliability are applied to the hole-by-hole scores of players in the 1992 US Masters and US Open tournaments with the aim of measuring the discriminatory ability of individual holes and the reliability of the golf course as a whole. Results were vastly different both from those usually obtained in psychometric testing and that expected from a knowledge of golf. There was often negative correlation between holes within a round and between one round and the others, and this produced negative reliability ratings. If the golf courses studied were questionnaires designed to measure a single attribute such as "golfing ability” they would be discarded. Neither tournament was able to differentiate between the 'ability' of top golfers. Small differences between the abilities of top players are outweighed by high levels of random error.