This paper utilizes a high frequency dataset on taxi rides in New York City to investigate how emotions due to sporting event outcomes affect passengers' tipping behavior. I formulate and empirically test a reference-dependent preferences framework of tipping behavior. The results indicate that the tipping amounts are driven by deviations from expectations much more so than wins and losses, with the most salient effects found under unexpected close wins. However, there is no support for loss aversion. The findings suggest that loss averse behavior may be subdued in the presence of social norms while surprises can result in freedom on the upside of tipping.